Finish up the calculations to use all four elections from 2008 to 2020 as the baseline "how well did Election Graphs averages do compared to the final results" data I use to try to generate odds from the polling averages and do a blog post about that.
So I guess it is time to do that.
OK, actually, it is way past time for that. I had hoped to have all this done by midterms, but I ended up spending most of the time I would have spent on that helping to do things like put out campaign signs for my wife's campaign. She won. Other things took up my time too. Anyway. I was delayed. But lets get this done…
Let's start with a simple scatterplot showing every state (and the DC and ME/NE congressional districts) from 2008 when I started doing this through to the 2020 results. Each data point will have the FINAL Election Graphs average on the X axis, and the ACTUAL election results on the Y axis:
As you would hope, these are nicely correlated at this scale. If EG's averages were always exactly right, every point would be along the black diagonal line. Of course, polls don't work like that, and even poll averages don't work like that. There is a vertical spread due to the inherent randomness of polling. A pretty wide spread actually.
But also if you look carefully, you can see that on the right side of the graph there are more points above the line, and on the left there are more points below the line. This means that there is also some bias here. Specifically a bias where the Election Graphs polling average tends to UNDERSTATE the magnitude of the winner's margin.
Let's do a transformation on the graph to try to look into the patterns a bit more deeply though:
That just looks like a messy colored blob initially. But what did I do here? I just transformed things by subtracting out the diagonal. Instead of looking at the actual election results vs the Election Graphs final margin, I look at the Delta… how far off the election results were from the Election Graphs final margin, vs the Election Graphs final margin. So, for instance, if on the top graph we had a point where the EG average was a 10% Republican lead, but the actual result was that the Republican won by 15%, that would show up as x=10%, y=5%.
OK, but can we say anything at all about this blob? Is this just two things that show no useful relationship at all?
Well, in this post from 2019 I actually looked at this before. So let me just quote a bit from there:
Before going further, let's talk a bit about what this chart shows, and how to interpret it. Here are some shapes this distribution could have taken:
Pattern A would indicate the errors did not favor either Republicans or Democrats, and the amount of error we should expect did not change depending on who was leading in the poll average or how much.
Pattern B would show that Republicans consistently beat the poll averages… so the poll averages showed Democrats doing better than they really were, and the error didn't change substantially based on who was ahead or by how much.
Pattern C would show the opposite, that Democrats consistently beat the poll averages, or the poll averages were biased toward the Republicans. The error once again didn't depend on who was ahead or by how much.
Pattern D shows no systematic bias in the poll averages toward either Republicans or Democrats, but the polls were better (more likely to be close to the actual result) in the close races, and more likely to be wildly off the mark in races that weren't close anyway.
Pattern E would show that when Democrats were leading in the polls, Republicans did better than expected, and when Republicans were leading in the polls, Democrats did better than expected. In other words, whoever was leading, the race was CLOSER than the polls would have you believe.
Finally, Pattern F would show that when the polls show the Democrats ahead, they are actually even further ahead than the polls indicate, and when the Republicans are ahead, they are also further ahead than the polls indicate. In other words, whoever is leading, the race is NOT AS CLOSE as the polls would indicate.
In all of these cases the WIDTH of the band the points fall in also matters. If you have a really wide band, the impact of the shape may be less, because the variance overwhelms it. But as long as the band isn't TOO wide the shape matters.
Now, back in 2019, at this point I jumped directly into looking at the pattern based on the combination of all the data from 2008 to 2016. Rather than doing the same thing now, but just adding in 2020, I think it is actually instructive to take a bit of a detour to look at each of the four election cycles separately.
First lets look at just the 2008 data points:
OK, the circles are just the individual 2008 data points, but what is the rest, what have I done here?
I've constructed "envelopes" using windowed averages and standard deviations.
Specifically, I am looking at windows with a 5% radius. At every value for the polling average at 0.1% increments, I look 5% in either direction (so a window 10% wide) and find all the data points within that window, then if there are at least five points, I calculate the average and standard deviation of those points. The 5% is of course just an arbitrary round number, as is the 5 data point minimum.
The bold line in the center is the mean, the next lines out are 1 standard deviation from the mean (about 68.3% of the data points should be inside these lines), and the next lines are 2 standard deviations from the mean (which should contain about 95.4% of the data points).
So the final chart here shows both the general trend in how far off final election results were from the Election Graphs average given where the average was, but also just how variable those results are.
For 2008, we can now see pretty clearly that this looks like "Pattern F".
When Election Graphs showed a tie, on average the actual election results were a Republican win by 0.2%. That's pretty close to a tie too. So not a lot of bias in one direction or another at the center. There is a BIG window though. If the EG average was a tie, the 95% confidence interval goes all the way from Democrats winning by 6.8% to Republicans winning by 7.3%.
But there is also definitely a trend where in cases where when Republicans were ahead in the EG average, they actually tended to win by MORE than the EG average, and if the Democrats were ahead, THEY tended to win by more than the EG average.
OK, you get the idea, so lets look at the rest of the election cycles Election Graphs has covered.
I've made the scales the same on all of these to make it easier to compare the cycles. While these all have the same high level diagonal pattern, the detailed shape of the curves is very different cycle to cycle, both in terms of the central average, and pattern of the variances.
Lets concentrate on those central curves and put them all on one chart…
You can immediately see the two elections Obama was in (the red and goldish ones) clump together, and the two that Trump was in (the light and dark greenish ones) clump together too.
They all have the general diagonal shape (Pattern F), but it stands out how, at least near the "Election Graphs average near zero" area which by definition are the states that matter in close races, how much more the poll averages underestimated Trump in 2016 and 2020, compared to a more neutral 2008, and an underestimation of Obama in 2012. Maybe these all have the Pattern F shape, but the slopes are different, and even more importantly, they are shifted vertically. They don't all just go through the origin.
And of these four years, 2020 is the only race yet where the average curve is above the zero line for its full length. For 2008 and 2012 if the Democrats were ahead in the average, on average they would do even better in the election than the polls indicated. In 2016, that was only true if the Democratic poll advantage was more than a 12.8% margin. But that didn't happen at ANY part of the range in 2020. At every range for the Election Graphs average, on average the Republicans did better than the average indicated, and the Democrats did worse. (Note that is not true for every single data point, just for the trend line.)
Does this mean anything for 2024? Quite probably not. As they say "past performance is not a guarantee of future results". It is unclear what all the reasons are for the shape of these error curves, and if pollsters are actively working on "correcting" things in the next cycle. Were the two Obama curves more "normal" and there is something unusual that just makes it harder to poll races with Trump in them? Or were the Obama races unusual too, and "normal" is somewhere in between? Or was there a systematic error that pollsters have a handle on now, and 2024 will just be a nice flat line? Or none of the above?
It would be nice if I had this kind of data from before 2008, but since that is when I started doing this, I don't. Also of course, you could argue that the world of 2024 is so different to 2004 that it wouldn't really be meaningful to look at that anyway. For other things on Election Graphs I use five election cycles (20 years) as a baseline, so I might still use older data if I had it. At least 2004. In general, I'd love to see how these kinds of curves have varied over even longer time periods. Oh well.
As it is though, since we don't know which way the errors will go in 2024, the best option I have available is to create this same sort of envelope using all four available election cycles.
This is what that looks like:
Doing this you lose the specific distinctiveness of the four presidential election cycles going into this. Instead, you essentially fold in some information about just how much polling accuracy has varied cycle to cycle.
Now, the "core" of Election Graphs has always been the dead simple method of taking the polling average at face value and classifying any states where the margin is less than 5% as states that could go either way, and presenting that range of possible results.
But starting in 2020 I experimented with producing probabilistic results too. Those probabilities were based on the 20082016 version of the chart above, and for 2024 I'll be using this 20082020 version.
The key is that for every value of the Election Graphs average, we have numbers for the mean actual election result (the EG average plus the delta), as well as an associated standard deviation. From this, we can construct a chart showing for each value of the Election Graphs average, the chances of a Democratic win and the chances of a Republican win based on the historical data. (As usual, since Election Graphs operates off the margin and not raw support numbers, if a 3rd party is ever in contention, this method falls apart.)
So what does that look like?
OK, that's nice at the same scale as all the other charts, but lets zoom into the critical central part here.
OK, given my methodology, this is a bit bumpy. I should probably smooth it out a bit or do a logistic regression or something. But given the levels of uncertainty we are talking about, I probably won't bother. I like a little bumpiness.
Anyway, this is the critical graph. And yes, this does smell a little of "unskewing" the polls. Don't worry though, ElectionGraphs will continue to show the straight up averages.
But what we see here is that over the last four election cycles, there HAS been a tendency to underestimate Republicans. Yes, as discussed above, all four cycles look different, and 2008 and 2012 look distinctly different than 2016 and 2020. But taken as a whole, the Election Graphs poll averages have underestimated Republicans.
So lets look specifically at a few data points on this chart.
Now, technically speaking because the way Election Graphs calculates averages (see FAQ, which I'm just realizing I need to update as well) there can't be exact ties in the Election Graphs average. But if there was an exact tie according to the curve above based on historical data, the Republican would actually have a 62.4% chance of winning, and the Democrat only a 37.6% chance. In order to have a better than 50% chance of winning, the Democrat would need to lead by at least 1.3%.
Looking at this another way, for the Democrat to have a better than 95% chance of winning, they need to be leading by 7.1% or more. For the Republican to have a better than 95% chance of winning, they only have to lead by 3.8%.
Again, this is based on comparing the Final Election Graphs averages to actual election results from 2008 to 2020. There is a chance 2024 looks nothing like the last four elections. Polling may be better. Polling may be worse. Polling may underestimate the Democrats this time rather than underestimating the Republicans. We just don't know. So looking at the four cycles of data I have so far is the best I can do…
Anyway, that is the curve I will be using to make my "probabilistic" views for 2024. This will not impact the traditional "categorization" views at all. They will remain as they always have been, classifying any state with a margin less than 5% as a "weak" state that could go either way. No "unskewing" there.
But I will allow the probabilistic views to take into account that based on this historical data, the Democrats have to be further ahead to have an even shot, etc.
Also important to note, is that these are the odds on a PER STATE basis (and CD for ME and NE). You can't apply these numbers to national polls. It just doesn't work that way. For that, I'll once again be doing Monte Carlo simulations using the state poll averages and the odds on the chart above.
One thing none of the above takes into account though is trying to estimate how correlated errors are between states. If errors were just completely uncorrelated, then when you run your simulation, you just roll the dice for each state. Distributions end up a bit narrower. But if things are completely correlated, so that if one state underestimates the Republican, they all do, then you essentially roll once to see how far off ALL the states are, and you end up with a much wider distribution with higher odds for the tails.
The reality is somewhere in between. I haven't had a good method for modeling "somewhere in between". So in the 2020 cycle, after a brief time mentioning only the fully independent version (which was a mistake, for which I was indirectly called stupid by Nate Silver), I ended up just showing both extremes. This wasn't entirely satisfactory either though. But at the moment I don't have a better idea.
If you are a stats and modeling person who wants to help me properly model the right degree of correlation between errors here, please get in touch. I'd love to learn more and do better.
That applies to everything else in the analysis above as well. I am well aware I am doing a few things that may not be exactly the right way to do things. I don't think anything is outright "bad", but I recognize there may be better ways of accomplishing what I am trying to accomplish, and maybe I'm wrong and some things ARE just bad. If so, I'd love to hear about it and learn… as long as you can give me that feedback nicely and gently, rather than being mean about it. Thanks.
OK. I guess that is it for the "preliminary math stuff".
Next up is standing up the actual 2024 page with only the previous actual election data (no 2024 polls yet) to define a "starting point". And updating the FAQ and things like that.
I'll post again once that is done, and before I start feeding it actual state polls for 2024 (of which there have already been quite a few).
]]>I had intended to post some sort of detailed 2020 post mortem like I did for the 2016 cycle. This site did pretty well again, in the same ballpark as some of the big guys. Roughly. I didn't ever sit down and look through all that in detail. So that post never happened. Frankly I was tired and exhausted with the election stuff, and aside from a few very minor things I would occasionally pick up and look at, I immediately fell into the mode of letting Election Graphs lie dormant until it was time to start thinking about 2024.
Welp, we are now only 78 days until the midterm election, which is traditionally the starting gun for the next Presidential election, and we have at least one candidate making noise that they intend to make their run official even before then. So I think it is probably time.
There is no Election 2024 content live on this site yet. I've generally tried to start updating the site in earnest right around the midterms, although sometimes it has ended up being a few months later.
I have a to do list which I am starting to work through though, so here are some of the highlights:
I don't want to hold up anything for that last one. I have a few ideas in my head, but nothing solid, the ideas may or may not work in real life, and time is at a premium, so I may or may not have time to do them anyway. But one thing at a time.
And the one thing right now is to post this, put my stake in the ground as to what I am doing next, and get to it…
Welcome to the start of Election 2024 coverage here on Election Graphs!
]]>So we'll just stick to the best and worst cases for the candidates based on which states have been called, and assuming that only the "Weak" states might go the "wrong" way. Of course there can be surprises in the other categories too. We'll adjust appropriately if that happens.
Newer updates will be right under this introduction, scroll down for the older updates. Times listed are UTC. 0:00 UTC in November is 7 PM Eastern, 4 PM Pacific.
Refresh the page periodically to see updates.
20210107 09:12 UTC
There were no faithless electors when the electoral college voted in December.
Despite a very eventful day in January when the electoral votes were presented to Congress, there were no changes to the totals.
So the final result is as expected since November 13th:
Biden 306 to Trump 232 – Biden by 74 electoral votes
At some point I'll start doing some analysis of the final state by state election results vs the final Election Graphs polling averages. I had hoped to actually do that by now. Oops.
I guess it is still fine as long as I do it before it is time to launch the 2024 version of this site, right?
Thanks again everybody for following us through the 2020 cycle.
20201113 19:15 UTC
NBC just called Georgia for Biden. This is the last state that hadn't been called by anybody officially, although it has been clear for days that it was going that way.
So, final table:
Biden  Trump  Margin  

Trump Best  306  232  Biden by 74 
Expected  306  232  Biden by 74 
Biden Best  306  232  Biden by 74 
This leaves us with four jurisdictions where the final results differed from the Election Graphs calculations:
All four of these places that went "the wrong way" were clearly in the zone where we gave significant odds to those states going the other way though, so none of them qualify as a huge surprise. So I am actually pretty satisfied with these results.
In addition, the "Biden by 74" EV was clearly within all of the "envelopes" this site produced, even if you go with the 1σ (68.27%) ranges for the two probabilistic views.
So I don't feel bad about these results at all. Yes, there is a lot of handwringing about how far polls were off, but with ranges and probabilities that were generated by looking at how far off polls have been in the past, you end up with views that give you a level of confidence given the historical accuracy, you get a sense of just how variable the results might be given by the polling we have. Bottom line is there is a lot of uncertainty. But you can measure that uncertainty.
Once all the counts in all the states are final and certified, I will do a more detailed look at the state by state polling errors and what patterns we see there. And of course we'll track if there are any faithless electors this time around. Look for posts on both of those topics in December most likely.
But for the moment, we have calls in every state, so we'll close out this post.
Thanks everybody for following us for the 2020 cycle!
20201112 03:47 UTC
Lots of other places are now calling Arizona for Biden. But we moved it into the Biden column way back when Fox called it. So nothing changes here. We're only waiting on calls for Georgia, which is also expected to be Biden at this point absent some huge surprise.
20201111 15:49 UTC
A couple of hours ago several outlets called Alaska for Trump. This does not change the matrix. The only state without a call is now Georgia, where Biden currently leads.
20201110 18:50 UTC
DDHQ calls North Carolina for Trump. This is one where our averages had Biden slightly ahead, so it is the third place where Election Graphs had the wrong winner. Given the 0.6% Biden margin in the state though, we still gave Trump a 47.9% chance of winning the state, so once again, this was essentially a coin flip, so we don't feel too bad about the miss.
All three misses so far were "Weak Biden" states which ended up going Barely Trump though. If current counts hold, Georgia will be wrong in the other direction.
The new matrix:
Biden  Trump  Margin  

Trump Best  290  248  Biden by 42 
Expected  290  248  Biden by 42 
Biden Best  306  232  Biden by 74 
Only Georgia and Alaska left uncalled.
Note that because the Election Graphs poll average had Trump leading by 5.9%, which put it in our "Strong Trump" category, it is not included in the swing above. Based on current trends, that doesn't look likely to introduce a surprise.
So the results above depend only on the call in Georgia, where Biden is currently leading. But nobody has officially called it yet.
20201107 16:43 UTC
And Fox calls Nevada for Biden. So the matrix tightens further:
Biden  Trump  Margin  

Trump Best  290  248  Biden by 42 
Expected  305  233  Biden by 72 
Biden Best  321  217  Biden by 104 
We're only talking about what Biden's margin is now. We've known Biden was the winner since 14:00 UTC yesterday. (Arguably even before that.) Some places are just slow officially making that conclusion.
20201107 16:34 UTC
Looks like all the major news outlets simultaneously called Pennsylvania, and the whole election, when Biden's lead went over 0.5% a few moments ago. But we were past that point yesterday. Still no additional change to the matrix for us. We'd already moved PA to the Biden column.
Still waiting on media calls for Nevada, North Carolina, Georgia, and Alaska.
20201106 14:30 UTC
Our "Expected" case in the table below is based on who led each state based on preelection polling. If all the remaining states just go to the candidate who currently leads the vote count, which seems very reasonable based on current trends, you get Biden 306 to Trump 232, or Biden by 74 EV.
20201106 14:00 UTC
DecisionDeskHQ just called PA for Biden as soon as Biden took the lead in that state. And that is that.
Biden  Trump  Margin  

Trump Best  284  254  Biden by 30 
Expected  305  233  Biden by 72 
Biden Best  321  217  Biden by 104 
Biden has won the Presidency. Only question now is by how much.
(And even if Arizona is uncalled, the Trump's best cases would still be Biden by 8, so it is still just a question of how big Biden's win is.)
20201106 09:22 UTC
Biden just took the lead in Georgia, and it seems unlikely that will change given the trends. Pennsylvania is expected to flip to Biden in the next few hours too. But no official calls yet. So not changing the matrix yet.
Also worth noting that while Fox/AP/Bloomberg all called Arizona for Biden, and thus we adjusted out counts here to reflect that, all the other news outlets are holding back, because Trump is closing the gap as more ballots are counted. We'll leave that call reflected here unless those outlets actually retract their calls.
20201104 21:25 UTC
CNN calling Michigan for Biden. Takes that off the table for Trump.
Biden  Trump  Margin  

Trump Best  264  274  Trump by 10 
Expected  305  233  Biden by 72 
Biden Best  321  217  Biden by 104 
20201104 18:35 UTC
MECD2 gets called for Trump by AP. That was also “Weak Biden” with a Biden lead of 2.7%. Odds wise we said Trump had a 28.6% chance of winning MECD2. So along with Florida, this is the second "miss" relative to our categorizations, but both were states where the odds of a flip were high enough so it isn't a surprise. That's why these are "Weak".
New matrix:
Biden  Trump  Margin  

Trump Best  248  290  Trump by 42 
Expected  305  233  Biden by 72 
Biden Best  321  217  Biden by 104 
20201104 17:13 UTC
Wisconsin election authorities say all votes are counted, and Biden is ahead. Of course, this was "Strong Biden" by the poll average, so it wasn't even supposed to be in contention, and it is actually close enough it could be in recount territory. For the moment, the matrix doesn't change though.
20201104 06:14 UTC
DDHQ calls NECD2 for Biden. It was Strong Biden, so expected.
20201104 05:29 UTC
Montana called for Trump, no surprise.
20201104 05:22 UTC
Fox/AP call Iowa for Trump. Takes it off as a possible Biden pickup:
Biden  Trump  Margin  

Trump Best  248  290  Trump by 42 
Expected  306  232  Biden by 74 
Biden Best  322  216  Biden by 106 
20201104 05:15 UTC
Minnesota goes to Biden as expected.
20201104 05:05 UTC
Hawaii goes where you expect.
20201104 04:55 UTC
Fox calls Texas for Trump. So that's off the Biden pickup list.
Biden  Trump  Margin  

Trump Best  248  290  Trump by 42 
Expected  306  232  Biden by 74 
Biden Best  321  210  Biden by 118 
20201104 04:46 UTC
Fox calls Ohio for Trump. Takes it off the possible pick up list for Biden.
Biden  Trump  Margin  

Trump Best  248  290  Trump by 42 
Expected  306  232  Biden by 74 
Biden Best  366  172  Biden by 194 
20201104 04:23 UTC
Fox calls Arizona for Biden. That was a "Weak Biden", so not a flip, but takes it off the table for Trump's best case. New matrix:
Biden  Trump  Margin  

Trump Best  248  290  Trump by 42 
Expected  306  232  Biden by 74 
Biden Best  384  154  Biden by 230 
20201104 04:08 UTC
Idaho goes the way you would thing.
20201104 04:02 UTC
Washington, Oregon, California for Biden.
20201104 03:52 UTC
NBC calls Nebraska, EXCEPT the 2nd.
20201104 03:35 UTC
Louisiana, Utah, and Kansas called the way you would expect.
New Hampshire too.
20201104 02:30 UTC
NBC calls North Dakota for Trump.
20201104 02:10 UTC
And AP calls Wyoming for Trump.
And New Mexico and Colorado for Biden
NBC calls South Dakota for Trump.
20201104 02:05 UTC
Biden gets New York. Shocker.
20201104 01:40 UTC
Arkansas for Trump. No surprise.
20201104 01:30 UTC:
Missouri for Trump, no surprise.
MECD1 and MEAll to Biden. No surprise.
20201104 01:20 UTC
DDHQ calls Florida for Trump. That had been "Weak Biden". First wrong state of the day.
New chart:
Biden  Trump  Margin  

Trump Best  237  301  Trump by 64 
Expected  306  232  Biden by 74 
Biden Best  384  154  Biden by 230 
20201104 01:15 UTC
Massachusetts and Maryland and Delaware and DC and New Jersey to Biden. No surprises. Illinois and Connecticut too. Rhode Island.
Trump gets Alabama, Oklahoma, Mississippi, and Tennessee
Still no states that were actually in question.
20201104 01:00 UTC
AP says Trump in South Carolina. No surprise.
20201104 00:15 UTC
Fox calls Virginia and Vermont for Biden. Also no surprises.
AP calls Kentucky for Trump. No surprise.
DHQ calls West Virginia for Trump. No surprise.
I'll only put in a new matrix if it changes.
20201104 00:05 UTC
Indiana for Trump. It was Strong Trump, so no surprise and no change to the matrix.
Biden  Trump  Margin  

Trump Best  237  301  Trump by 64 
Expected  335  203  Biden by 132 
Biden Best  413  125  Biden by 288 
20201103 23:00 UTC
No states called yet.
Biden  Trump  Margin  

Trump Best  237  301  Trump by 64 
Expected  335  203  Biden by 132 
Biden Best  413  125  Biden by 288 
Election Graphs has logged its last poll on Twitter for the 2020 season. In just a couple of hours, polls will start closing, and we will start getting actual election results.
So where did Election Graphs end up? What is our final prediction?
If you just want the bottom line, without any of the pretty charts or discussion, it is this:
Biden continues to be heavily favored to win. He is still in a stronger position than Clinton was four years ago. However, our numbers show a significant lastminute strengthening by Trump which makes Biden's lead much more tenuous than it was a week ago. If we had the same size polling error as 2016 in Trump's direction, he would win. (The same size error in the other direction would get to a Biden landslide… there are LOTS of close states.)
Having said that, the nature of some of the recent polls raises a real question about how much of this final move is "real" rather than just a lot of partisan and lesserknown pollsters flooding the final averages with polls that are favorable to Trump.
OK. If that is all you need, goodbye! See you for the 2024 cycle!
If you want more detail though, keep reading.
(There will also be a post tracking results as they come in tonight, and eventually, later this year once all the results are known and final, an analysis of how we did, so if those interest you, check back in later!)
The last blog update here was yesterday on November 2nd. There were hundreds of new polls (or at least it seemed that way) on that last day. Here are the changes in our metrics just in the last 24 hours.
Model  Metric  2 Nov  3 Nov  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +10 Biden +112 Biden +228 
Trump +6 Biden +108 Biden +232 
Trump +16 Trump +4 Biden +4 
Trump Win Tie Biden Win 
1.2% 0.1% 98.6% 
2.5% 0.3% 97.2% 
+1.3% +0.2% 1.4% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +84 Biden +102 Biden +314 
Trump +86 Biden +132 Biden +294 
Trump +2 Biden +30 Trump +20 
Trump Win Tie Biden Win 
18.8% 0.0% 81.2% 
31.9% 0.0% 68.1% 
+13.1% FLAT 13.1% 

Categories  Trump Best Expected Biden Best 
Trump +64 Biden +102 Biden +288 
Trump +64 Biden +132 Biden +288 
FLAT Biden +30 FLAT 
Tipping Point  Biden +3.5%  Biden +2.3%  Trump +1.2% 
On the one hand, North Carolina moved back to the Biden side of the centerline. Which strengthens Biden by 30 EV in both the Uniform Swing Median and the Expected Case Median. But in the meantime, the huge wave of final polls reduced Biden's margin in the critical "Weak Biden" states.
And yes, it is the same thing I mentioned last time when I said:
There have been dozens of new polls since the update on the 31st. While there were exceptions, the overwhelming theme was polls showing a better picture for Trump and a worse one for Biden than we had seen any time recently. So almost all of the metrics above move in Trump's direction.
As with a couple of the other recent spikes toward Trump in the last few weeks, you can look at the specific polls in the average, and start wondering if this is just a bunch of rightleaning pollsters flooding the zone with polls in the last few days. While there were plenty of results from large mainstream pollsters released in these last few days, there were quite a few partisan polls released too, and some of them definitely helped move the averages toward Trump. I talked about this phenomenon on October 20th, October 24th, and October 31st… so I won't delve into that indepth again.
…
Will this deteriorate further in the limited time left? Or bounce back? Or neither? If I had to guess, I'd say reversion to the mean. There have been no major news events in the last week that would seem likely to drive a tightening. So I expect this "tightening" is actually just an artifact. But that is just an educated guess. We will all know soon enough.
It did not revert to the mean. Biden's situation deteriorated a lot more. It ended up being significant in the final numbers.
So I guess I do have to delve into it again. Bottom line, in the last few days before Election Day, every pollster in existence seems to have wanted to make sure they got a shot at Election 2020, and a huge volume of polls was released. Many of these were from pollsters that have clear partisan leanings, and whose results were much more positive to Trump than what we had been seeing. Others were just lesserknown pollsters.
Election Graphs uses a "Last X Polls" model, specifically so that as the election approaches and there is more polling, we look at shorter and shorter timeframes and are more responsive to changes. We also use the middate of polls to determine just which polls are part of the "Last X" rather than the enddate. This means that in terms of the "final" averages we favor the very lastminute polls put out with very short times in the field.
The result is that many of the large wellknown polling companies with good reputations that put out their final polls late last week or even over the weekend got crowded out of the critical state averages by the large volume of random polls, which often just covered one or two days in the field.
Those pollsters tended to be much more bullish on Trump than the usual suspects who had been polling this race regularly over the last year.
So the averages in many states jerked a few percentage points toward Trump, just in the last few days.
If there was a big event in the news that was bad for Biden, then it would be easy to say that was the cause. The closest thing to that has been the dripdrip of information about Biden's son, but for the most part that has not seemed to actually get much traction. But maybe?
In the absence of a clear reason for movement, while there is most definitely a clear difference in the nature of the polling that comprises the average, it seems reasonable to think that maybe the movement is just an artifact of the polling, and doesn't represent a real change.
So perhaps the picture the site had of the race a week or two ago is actually a better representation of what is happening?
Maybe.
But four years ago we had a similar move in the last few days before the election. And that time it also seemed to be driven by a surge of polls from outlets that hadn't been doing a lot of polling earlier in the cycle. And we ended up closer to the final result because we reflected the change caused by those polls. That last burst of polls ended up being closer to reality than the more established pollsters they displaced.
Of course, that could well have just been luck too.
We set the rules for how we define the averages, what we would include and would not include, etc over two years ago, based heavily on what we did in 2008, 2012, and 2016. We're not going to change anything on the last day.
So it is what it is.
We'll find out in a few hours if the short time frames and the burst of lower quality polls ended up causing Election Graphs to dramatically overestimate Trump and underestimate Biden. If we are way off, maybe we'll change something for 2024. Or maybe the movement seen in these last 48 hours actually better represents what is going on.
I don't know. We will see.
So I can give the caveats above, but the numbers are what they are, and so we'll do the rest of this discussion taking them at face value.
Let's look at all the graphs.
First of all, the comparison with 2016:
After a brief spike toward a stronger position for Biden, the collapse we have been talking about happened. The final tipping point was only a 2.3% Biden lead. Biden had seen worse tipping points, but not since early June.
Even at this level though, Biden is stronger than the 1.6% that Clinton ended with. So he is still in a better position.
However, the difference between the Election Graphs final tipping point and the actual election results in 2016 was 2.36%. So the same size error in the correct direction would make Trump the winner.
In terms of the expected case, where each candidate wins every state where they lead the Election Graphs average, Biden is in the low end of his recent range, but not any lower, and still significantly ahead of where Clinton was.
OK. Let's look at the range of margins predicted by our three models:
In all three of these cases, the center lines have moved toward Trump, but are either still in their normal ranges, or are just barely out it. What has changed though is that in all cases the upper end of the envelopes, representing Trump's best cases has stretched out significantly further into the Trump wins zone.
This makes sense because most of what we saw was not states actually flipping from Biden to Trump in the averages, just Biden's lead in states being significantly diminished. So the straightup scenario that happens if the averages are all correct doesn't move much. But the sensitivity to the averages being wrong and therefore states flipping to Trump based on those errors increases significantly.
As a sanity check, let's compare our three centerlines with what other sites are saying at the moment:
So I'm in the zone. Most of the big folks have Biden doing better than I do, but there are a few places that have him doing worse.
And the Election Graphs "expected case" exactly matches what the Upshot says would be the result if the polling error in 2020 was the same magnitude and direction as the error in 2016. Which is interesting.
I note of course that nobody has Trump winning in their "expected case".
OK, with that done, let's look at odds.
While still small, the red zone in the Independent States odds view is now significantly larger, with Trump's chances now at 2.5%.
Trump's odds in the uniform swing model are massive now though, at 31.9%.
Fundamentally, since this imagines the extreme case where all the states move in a completely synchronous way, the only thing that matters is the odds in the tipping point state. And as of the final situation, the tipping point is Pennsylvania, and our average has Biden leading Pennsylvania by only 2.3%.
And in our analysis of our results from 2008 to 2016, we discovered that when we have a Democrat leading by 2.3%, that translates into a 68.1% chance of the Democrat winning, and a 31.9% chance of the Republican winning.
So Trump's chance of winning Pennsylvania is 31.9%, and if all the states are locked together, that means his odds of winning the whole election would be 31.9% too.
Since both of these models are extremes, and the truth is somewhere in between, the official Election Graphs statement on Trump's chances at this point is "between 2.5% and 31.9%". Election Graphs doesn't actually model where the right spot is within this range, but the middle is as good a spot to look at as any. That would be a 17.2% chance of a Trump win at the moment.
Once again, let's compare with the other folks:
My two extreme models are unsurprisingly near the extremes. "The middle" puts me lower than most of the "big" outlets. So maybe they think which there is some correlation between the states, their models still think they are more independent than not.
In any case, Biden is a favorite in all of these views. But in some, the chances of an upset are much more than others.
Not including the three from me, the median is a 9.3% chance of a Trump win. Including all three of mine, the median is 10.0%.
So as usual I make the statement that people tend to be bad at interpreting odds. 10% is not 0%. 10% happens all the time. Biden is favored, but a Trump win is still very possible.
Looking at the tipping point without the 2016 comparison, the main thing to notice is the huge volatility at the end. Swinging first in favor of Biden, then against him. This is an indication that perhaps for 2024, I should look into ways to make this a little LESS sensitive to short term changes in the last weeks.
OK. Time for the map and spectrum:
This view makes the volatility of the race clear. There are 10 states (and Maine CD2) with margins less than 5%.
The polling error in 2016 was 2.36%. There are 8 states closer to the centerline than that, including the tippingpoint state.
If we get a 2016 size error favoring Trump, we get a very narrow win for Trump, squeaking past the post with an 18 EV margin.
If we get a 2016 size error favoring Biden, Biden wins by 288 EV, the largest winning margin since 1988.
More likely than either extreme of course is that we just get a solid but not extraordinary Biden win.
And now the trends in all the close states:
It is election night. The first results will be coming out within a couple of hours. We may or may not end up knowing who wins tonight, but we'll still learn a lot. And soon enough, we will indeed have a winner.
I hope you have enjoyed Election Graphs and found it useful this year. Aside from seeing how things turn out and how we did, we are done.
It was fun. Thanks everyone!
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>The last blog update here was on October 31st. It is now November 2nd. We only have about a day and a half to go until we start getting actual results from Election 2020. I expect to do one more blog update before polls close on Election Day.
Model  Metric  31 Oct  2 Nov  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +52 Biden +150 Biden +256 
Biden +10 Biden +112 Biden +228 
Trump +42 Trump +38 Trump +28 
Trump Win Tie Biden Win 
0.1% 0.0% 99.9% 
1.2% 0.1% 98.6% 
+1.1% +0.1% 1.3% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +52 Biden +176 Biden +294 
Trump +84 Biden +102 Biden +314 
Trump +32 Trump +74 Biden +20 
Trump Win Tie Biden Win 
14.6% 0.0% 85.4% 
18.8% 0.0% 81.2% 
+4.2% FLAT 4.2% 

Categories  Trump Best Expected Biden Best 
Trump +20 Biden +176 Biden +288 
Trump +64 Biden +102 Biden +288 
Trump +44 Trump +74 FLAT 
Tipping Point  Biden +4.3%  Biden +3.5%  Trump +0.8% 
There have been dozens of new polls since the update on the 31st. While there were exceptions, the overwhelming theme was polls showing a better picture for Trump and a worse one for Biden than we had seen any time recently. So almost all of the metrics above move in Trump's direction.
As with a couple of the other recent spikes toward Trump in the last few weeks, you can look at the specific polls in the average, and start wondering if this is just a bunch of rightleaning pollsters flooding the zone with polls in the last few days. While there were plenty of results from large mainstream pollsters released in these last few days, there were quite a few partisan polls released too, and some of them definitely helped move the averages toward Trump. I talked about this phenomenon on October 20th, October 24th, and October 31st… so I won't delve into that indepth again.
But there is one critical difference. Judging by what happened in the last couple of cycles, there might be a handful of straggler polls released Tuesday morning, but we really only have one full day left for polls to be released. And many of the major pollsters have already released their "final polls" for the critical states. That means that any outliers introduced in the last few days, or in the remaining time we have left, will very likely still be part of the final averages this site produces.
There is no longer much time for an outlier poll to be "washed out" by additional polling. We are nearly at the end.
So while there may be some changes tomorrow, let's take seriously what we have today, and not try to make excuses. Four years ago on this site, there was a similar move toward Trump in the final days, and I somewhat dismissed it by pointing out the influence of partisan pollsters and what looked like possible outliers in the latebreaking polls. And of course, in retrospect, that move was real.
Is this move real? We'll know once the actual votes get counted.
So what do we have today if we take our averages seriously?
So let's look at all the graphs.
First of all, the comparison with 2016:
In the tipping point, which represents the degree polls need to be wrong and/or change before the end in order to flip the winner, Biden now holds a 3.5% lead. This is the worst level Biden has been at since midJune, but it still beats the 1.6% Clinton was at four years ago by a significant amount.
Will this deteriorate further in the limited time left? Or bounce back? Or neither? If I had to guess, I'd say reversion to the mean. There have been no major news events in the last week that would seem likely to drive a tightening. So I expect this "tightening" is actually just an artifact. But that is just an educated guess. We will all know soon enough.
In terms of the expected case, where each candidate wins every state where they lead the Election Graphs average, Biden has a slightly lower margin than he has seen before in the last three months. Georgia, North Carolina, Texas, Ohio, and Iowa, all of which have spent time on the Biden side of the centerline, are all now in "Weak Trump" territory in the Election Graphs averages at the same time.
But Biden is still ahead by 102 electoral votes in this view. By this time four years ago, Clinton was hanging on with only an 8 electoral vote lead.
OK. Let's look at the range of margins predicted by our three models:
All three of these still have their centerlines with the narrowest Biden lead they have seen in the last three months. These levels were last seen in June. So not unprecedented, but still breaking out of the recent zone.
In terms of odds from the probabilistic views, the Independent States chart no longer just looks like a blue square. So here it is:
See that little bit of red at the bottom right? For the first time in a long time, Trump's chance of winning is visible in this view, as it rises to 1.2%. Prior to this bump, the highest this has been in the last three months was 0.2%.
Those odds assume that there is no correlation between polling errors in different states though. So time to look at the other extreme.
In this view, Trump is now at an 18.8% chance of winning.
As with all of the other views, this is the best Trump's numbers have been in the last three months.
Since both of these models are extremes, and the truth is somewhere in between, the official Election Graphs statement on Trump's chances at this point is "between 1.2% and 18.8%". The middle of that range is 10.1%.
While Election Graphs doesn't actually model where the right spot is within this range, the middle is as good a spot to look at as any. So I'll just call it a 10.1% chance of a Trump win at the moment. That is up from 7.4% when we looked at this two days ago.
As 2016 should have taught everyone, a 10% chance of winning is not the same thing as a 0% chance of winning. Things that happen 1 in 10 times happen many many times every day.
Having said that, Biden is still a heavy favorite. We're not in a coin toss scenario, and certainly not in a situation where Trump is favored.
Looking at the tipping point without the 2016 comparison, there is one additional thing I would like to point out besides the fact that the tipping point is at a 3.5% Biden lead, which is outside of the range it has been in for the last three months.
Specifically, look a few days earlier. The many polls that came since the last update included a lot of polls that were very favorable to Biden that were in the field a few days earlier than the ones that drove the current numbers toward Trump. So I now show a peak on October 29th to a 7.4% tipping point lead for Biden based on that polling! That was also outside the normal range for the last three months. In fact, it was the best tipping point number Biden has EVER seen!
Because polls are coming in fast and furious, and Election Graphs is intentionally very sensitive to short term changes as polling velocity increases, we are seeing LOTS of volatility as outliers come and go from the averages.
Which is another reason to think that the "truth" here is probably a fairly stable Biden +5% tipping point lead, and the ups and downs here are just polling noise. Maybe in future cycles, I should consider making things a little LESS sensitive to short term changes. Maybe a 10 poll average instead of a 5 poll average. That would smooth things out a little bit.
For now though, when you see a noisy graph like this, it almost always means that the true "signal" is not actually an underlying reality rapidly moving up and down, but rather you are just seeing measurement artifacts, and you want to consider the overarching trend, not the transient jiggles.
OK. Time for the map and spectrum:
As with all the other views, we have shown, this spectrum shows a lot closer race than we have seen since June. There are more states on the Trump side of the centerline, and the "Weak Biden" states are weaker than they were.
Subject to any changes due to additional polls coming in over the next day or so, what does that mean in different polling error scenarios?
As we discussed last time, the Election Graphs tipping point was off by 0.89% in 2012, 2.36% in 2016, and 3.45% in 2008.
You need a larger error than any of those three election cycles to get a straightup Trump victory.
But a 2008 level error would have Trump winning Maine CD2, Arizona, and Florida in addition to the states he leads, and leave Biden's win dependant on a lead of less than 0.1% in Pennsylvania. With all of the talk of legal disputes over what ballots are counted, a 2008 level polling error in this direction would certainly put that scenario into play.
On the other hand, even a 2016 level polling error in the other direction would result in Biden winning Georgia, North Carolina, Texas, Ohio, and Iowa. That would be Biden 413, Trump 125 in the final count, or a 288 electoral vote margin, which would be the largest win since Bush beat Dukakis in 1988.
The most likely result is of course in between: A Biden win, but not by an overwhelming margin.
And now the trends in all the close states:
And that is that.
1.5 days until the first results start coming in on election night.
If you are eligible to vote in the US and have not yet done so, make your plan and get it done.
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>The last blog update here was early UTC on October 28th (with numbers as of the end of the 27th UTC). Just over three days ago. Here are the highlevel changes since that last post:
Model  Metric  27 Oct  31 Oct  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +38 Biden +142 Biden +256 
Biden +52 Biden +150 Biden +256 
Biden +14 Biden +8 FLAT 
Trump Win Tie Biden Win 
0.1% 0.0% 99.8% 
0.1% 0.0% 99.9% 
FLAT FLAT +0.1% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +52 Biden +144 Biden +300 
Trump +52 Biden +176 Biden +294 
FLAT Biden +32 Trump +6 
Trump Win Tie Biden Win 
11.8% 0.0% 88.2% 
14.6% 0.0% 85.4% 
+2.8% FLAT 2.8% 

Categories  Trump Best Expected Biden Best 
Trump +20 Biden +176 Biden +294 
Trump +20 Biden +176 Biden +288 
FLAT FLAT Trump +6 
Tipping Point  Biden +4.4%  Biden +4.3%  Trump +0.1% 
I'm updating more often since we are so close to the end, but the bottom line is again that things are fairly steady. If you look at the numbers above, yes, they move around a bit. But we still have nothing that looks like a sustained move as opposed to just poll averages meandering up and down as specific polls come in and out of the average.
Now, if I'd been making this post a bit sooner, it might have looked a lot like my "The Race Tightens? Or Not?" post from October 20th. This time the tipping point moved from Biden+4.4% on the 27th, down to Biden+3.1% on the 29th. I would be asking if this was the beginning of the race starting to tighten.
But as with that post, I would have expressed skepticism because just as it was back on October 20th, this spike was driven by a rash of polls from pollsters that tend to put out outliers. As new polls came out on the 30th, the tipping point bounced back to Biden+4.3%, almost back to where it had been before. It was just a transient spike.
In some cases the outliers that drive this are extreme. In one dramatic instance, Trafalgar came out with a poll showing Trump ahead by 2% in Michigan at a time when the other polls in the Election Graphs average ranged from Biden+7.7% to Biden+13%. The Trafalgar number was not just at the Trump end of the range we had been seeing, it was on a whole different planet.
Whenever I see polls like that come in, I start wondering if I shouldn't have switched Election Graphs from using averages to using medians years ago. Or maybe that I need some sort of weighting by historical pollster accuracy, or to do something more complex to deal with and adjust for outliers, or just include more polls to wash these outliers out a bit more.
But one of the main ideas of Election Graphs from the beginning has been to include everything and keep to relatively simple averages, and see how far that gets you. It does make us bounce around when we get outliers though, and makes us VERY sensitive to just which pollsters have been in the field most recently.
If there is a REAL move, then after you get a spike led by a particular poll, additional polls will confirm the move, and you'll see a sustained change that lasts even after that first poll rolls off of the average. When it is just a movement driven by an outlier though, things will move back to where they were once a few new polls come in.
The problem here at the very end of the race is that at some point there will be no new polls. The last pollsters in the field will dominate the average. And they may or may not be the ones with the most accurate results. Last time around in 2016 Election Graphs did pretty well. Maybe we were lucky. We'll see how it goes this time.
For the moment though, when you look at the graphs, it seems that after a brief spike where it looked closer for a moment, we're back in the same familiar territory for this race as we have been since June, showing Biden as a heavy favorite, but with the possibility of a Trump upset not completely out of the picture.
So let's look at all the graphs.
First 2020 vs 2016 comparisons since everybody on both sides keeps saying "But what about 2016?" when people talk about Biden's lead.
In the tipping point, which represents the degree polls need to be wrong and/or change before the end in order to flip the winner, Biden now holds a 4.3% lead. Clinton was at 1.6% by this point in 2016.
Biden is closer to the 4% part of his "normal range" than the 6% end, but we are still in very familiar territory. This race is pretty much in the same place it has been for months, Biden remains stronger than Clinton was, and we do not see any evidence of a collapse.
In terms of the expected case, where each candidate wins every state where they lead the Election Graphs average, Biden is still pretty much in the center of his normal zone, and significantly ahead of where Clinton was.
OK. Time for the three envelopes we have covering this year:
Instead of saying it three times, I'll say this once, since it is the same for all three:
There has been no substantive change to the race in months. All of these charts just show little bounces up and down as polls move in and out of the averages.
Fundamentally, the race looks the same now as it did at the beginning of August or even midJune.
In terms of odds from the probabilistic views, the Independent states chart still looks like a big blue square. Biden's chances in that model have been pegged near 100% for many months.
So looking at Uniform Swing…
In this view, Trump is now at a 14.6% chance of winning. That is higher than his chances have usually been, but not out of the range we have seen.
Since both of these models are extremes, the official Election Graphs statement on Trump's chances at this point is "between 0.1% and 14.6%".
We don't officially specify a spot within that range since we haven't modeled how much correlation between states to expect, but I guess the midpoint is as good a value to look at as any other. The middle of that range is 7.4%. That intuitively feels reasonable. Not total correlation, not total independence, but right in between. And frankly, that seems to be in the same general ballpark as all the big players who are doing election modeling. So call it 7.4%.
At 7.4%, Trump has about a 1 in 14 chance of winning. That would clearly be an upset, but is not so far out of the realm of possibility that people should be thinking about it as if a Biden win was a done deal. Of course, because of 2016, almost nobody is.
By contrast, Trump was at about 14% in 2016 if you looked at the median of all the sites giving odds. That would be about a 1 in 7 chance of winning.
So yes, Biden's polling lead is not so great you can say that there is a negligible chance of a Trump upset, but the chances of that kind of an upset are around half what they were in 2016.
This is just zooming in on the Tipping Point line without the 2016 comparison. Like all the other charts, this shows some bouncing around, but fundamentally a static race. The few times it looked like we might have been breaking into new territory in one direction or the other, it has proven to only be a short term spike once more data came in.
Of course, there is no guarantee that something won't change in the last few days, but there simply isn't much time left for a significant change.
Looking at this spectrum, one thing that you can look at more clearly than some of the other views is imagining polling errors of the sizes we have seen in the last three elections.
As measured by the difference between the final Election Graphs tipping point, and the actual tipping point based on votes cast in the elections, you had these errors:
Assuming uniform swing, and that the polls are overestimating Biden:
So in order to win based on "polls were wrong" (rather than a lastminute move in the polls in the next few days), we would need a bigger polling error than we have seen in the last three cycles.
On the other hand, also assuming uniform swing, but this time assuming the polls are overestimating Trump:
And now the trends in all the close states:
That is all for today.
3.5 days until the first results start coming in on election night.
We are so close. Almost there now.
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>The last blog update here was three days ago on October 24th. Here are the highlevel changes since that last post:
Model  Metric  24 Oct  27 Oct  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +52 Biden +162 Biden +254 
Biden +38 Biden +142 Biden +256 
Trump +14 Trump +20 Biden +2 
Trump Win Tie Biden Win 
0.1% 0.0% 99.9% 
0.1% 0.0% 99.8% 
FLAT FLAT 0.1% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +52 Biden +144 Biden +294 
Trump +52 Biden +144 Biden +300 
FLAT FLAT Biden +6 
Trump Win Tie Biden Win 
9.1% 0.0% 90.9% 
11.8% 0.0% 88.2% 
+2.7% FLAT 2.7% 

Categories  Trump Best Expected Biden Best 
Trump +12 Biden +220 Biden +294 
Trump +20 Biden +176 Biden +294 
Trump +8 Trump +44 FLAT 
Tipping Point  Biden +4.9%  Biden +4.4%  Trump +0.5% 
So, as we are only 7 days out now, I'm trying to update much more frequently. But honestly, not much has changed since the post three days ago.
Some states have bounced around. Biden had taken a tiny lead in Texas last time. Texas has now moved back to a small Trump lead. Meanwhile, Georgia had bumped into a small Trump lead but is now back to a small Biden lead. Iowa crossed the center line too, moving from Trump to Biden. Florida moved from Weak Biden to Weak Trump, and then back to Weak Biden again since last time as well.
All of these are states that are really tossups getting jostled around by whatever individual polls come out at any given moment. There is still no evidence of any sustained trend.
It looks like we are back once again to the part of the normal range that is most favorable to Trump, but we are still in the same range we have been since June. That translates to the reasonable possibilities ranging from a narrow Trump win if there is a substantial polling error or lastminute move, to a Biden landslide if the error or move goes in the other direction.
That entire range is reasonably possible, but the most likely outcome continues to look like a Biden win of a magnitude somewhere between Obama's 2012 win and Obama's 2008 win.
So with that out of the way, let's just look at all of the graphs.
This time, we'll start with 2020 vs 2016 comparisons since that continues to come up again and again in the discussions of this election cycle.
In the tipping point, which represents the degree polls need to be wrong and/or change before the end in order to flip the winner, Biden now holds a 4.4% lead. Clinton had dropped to 1.3% by this point in 2016.
Biden is closer to the 4% end of his 4% to 6% range than he is to the 6% end, but he is still in the same zone he has been in for a long time. By this time Clinton had collapsed and there was a really close race happening.
To be fair though, 7 days out from the 2016 election we didn't know this yet. Her tipping point number at the 7day mark as we look back at it today is influenced by polling that was in the field at this point, but not actually released until the last few days before the election. The first Election Graphs post noting the final Trump surge in 2016 came at the 4day mark.
So while we haven't seen any signs of Biden collapse yet, and Biden's pattern has been much less volatile than Clinton, this is just another reminder that Clinton's final collapse didn't show up here until the very very end.
In terms of the expected case, where each candidate wins every state where they lead the Election Graphs average, Biden is pretty much in the center of the normal zone. Within the last two weeks, the expected case has ranged from Biden +86 to Biden+240. That is a huge range!
But with Texas (38 EV), Florida (29 EV), Georgia (16 EV), and Iowa (6 EV) all close enough that they are flipping back and forth across the center line semiregularly, having volatility of this sort should be expected. With the exception of one brief moment though, Biden has consistently been ahead of where Clinton was four years ago, and usually by quite a bit.
OK. Forget 2016 now.
Time for the three envelopes we have covering this year:
Our median of the probabilistic view assuming all states are completely independent shows us near the top end of the normal range, but still in the range.
The 3σ envelope (99.73% of all outcomes within the bubble) just barely stretches to a Trump win by 4 EV, while the other end of this bubble would be Biden winning by the largest margin since Bush beat Dukakis in 1988.
The darker bubbles are of course more likely than those extreme scenarios. The 1σ (68.27%) range goes from Biden by 84 to Biden by 206. Comparing that to previous elections, that is basically from a narrow win similar in scale to Trump's win in 2016 to a Clinton 1992 level win on the other.
Converting this to odds, this view has Biden with a 99.8% chance of winning. This is down a little from before, but still so close to 100% that our chart of this just looks like a blue box. So we once again skip that chart.
As we have discussed before though, the independent states view results in the narrowest possible distribution of possibilities, since a candidate outperforming the polls in one state tends to be balanced by underperforming in another.
The uniform swing view instead locks all the states so they move together. This results in a much wider range of possibilities, since now if a candidate overperforms the polls, we assume they do so in EVERY state.
With this view, the 3σ (99.73%) range goes all the way from Trump beating his 2016 numbers with a 114 EV win, to Biden having a landslide even bigger than Bush in 1988 (but not quite as big as Reagan in 1984).
This view also shows a race with very little movement though.
Looking at the odds corresponding to the uniform swing view, Trump is now at an 11.8% chance of winning. Again, not breaking new ground.
Both of these views are extremes though. The official Election Graphs view of Trump's chances at this point is "between 0.1% and 11.8%". The middle of that range is just about 6%, but we don't specify a spot within the range since we haven't modeled how much correlation between states to expect.
And of course the good old categorization view we have used on Election Graphs since 2008. The lack of movement in this race is even more obvious in this view.
Biden's best case (where he wins every single state with a margin under 5%) barely moves at all.
Trump's best case does wander a little bit as some blue states dip over and under the 5% Biden lead line, but there is no trend here. At the moment though, if Trump wins every one of the close states, he squeaks out a 20 EV win.
Similarly, the centerline just meanders up and down as the very close states move back and forth between "Weak Biden" and "Weak Trump" as new polls bat them back and forth.
But overall, we just don't have any substantial change in the overall picture or trend in months.
Without the 2016 comparison line, we're more zoomed in on the tipping point here, so it looks like this is swinging dramatically, but that is just the scale. Aside from very brief moments, this has been between 4% and 6% for months.
OK, now the map and spectrum of the states:
And now the trends in all the close states:
And that is that.
7.0 days until the first results start coming in on election night.
One week. That's all.
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>I'm trying to do these blog updates more often now that we are in the last two weeks. The last blog update here was on October 20th. Here are the highlevel changes since that last post:
Model  Metric  20 Oct  24 Oct  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +32 Biden +140 Biden +254 
Biden +52 Biden +162 Biden +266 
Biden +20 Biden +22 Biden +12 
Trump Win Tie Biden Win 
0.3% 0.0% 99.7% 
0.1% 0.0% 99.9% 
0.2% FLAT +0.2% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +52 Biden +164 Biden +294 
Trump +52 Biden +144 Biden +294 
FLAT Trump +20 FLAT 
Trump Win Tie Biden Win 
23.3% 0.0% 76.7% 
9.1% 0.0% 90.9% 
14.2% FLAT +14.2% 

Categories  Trump Best Expected Biden Best 
Biden +40 Biden +164 Biden +294 
Trump +12 Biden +220 Biden +294 
Trump +52 Biden +56 FLAT 
Tipping Point  Biden +3.1%  Biden +4.9%  Biden +1.8% 
Last time we talked a lot about how the big move in Trump's direction we saw then could very well just be a transient thing because there was a rush of Pennsylvania polls from small pollsters that were favorable to Trump, which might not be sustained once the bigger pollsters released new results for Pennsylvania.
That is exactly what happened. In the days since that blog post, there were quite a few additional polls released in Pennsylvania. Here is what Pennsylvania looks like now:
A brief foray into "Weak Biden", and then right back to "Strong Biden".
These results not only moved the tipping point back toward Biden, but they also knocked the top off the peak we saw last time, so instead of the tipping point getting down to a 3.1% Biden lead, now Biden's worst tipping point was only 3.6%.
The tipping point is now at Biden by 4.9%, right smack in the middle of that 4% to 6% range it has normally been in ever since midJune.
Not only that, but the same is true of almost all of the main metrics. They are all in their "normal ranges".
Let's take a quick look:
The Tipping Point, the Independent States View, the Uniform Swing View, the Categorization View, and the odds generated from the two probabilistic views all show the race back in the "normal zone". (I didn't bother with showing the second odds chart since it is just a blue rectangle.)
So bottom line, a few days of bad polls in Pennsylvania moved things in Trump's direction. But with more polling, it doesn't seem like a "real" move representing a fundamental change.
Instead, it once again looks like this race has basically been frozen since June. The moves up and down we see are mostly just random results of which pollsters released results recently, and normal sampling errors and such. Very little, if any, real movement.
Roughly speaking, ever since June, this has looked like a race where Biden leads by between 4% and 6% in the tippingpoint state, with a potential winning margin somewhere between 100 and 200 electoral votes.
We still have 11 days left of course. There is still a little bit of time left for that to change.
And also, there is still the possibility of a systematic polling error. Between our two probabilistic models based on how far off Election Graphs state averages were in 2008, 2012, and 2016, the Election Graphs estimate for Trump's odds of winning if the election was today currently stand at "somewhere between 0.1% and 9.1%".
Where in that range depends on how closely linked the results in the different states are, which we don't estimate. But everyone needs to continue to remember that 10%, or 5%, or even 1% aren't the same as 0%, and there are still paths to a win for Trump. This isn't over yet.
The map and the center of the spectrum of states look like this now:
To win, Trump needs to run the table and win every single close state, or pull in some not so close states.
Let's do the comparison to 2016 now:
In terms of the Electoral College in the categorization view, Biden has been in a better spot than Clinton since we hit 76 days out, most of the time being more than just a little bit ahead of her pace.
But Biden's tipping point has been comparable to where Clinton's was at the same point four years ago for the last few weeks. Biden has not been consistently ahead of Clinton on this metric.
As of this blog post, Biden is once again ahead of where Clinton was though. Eleven days out Biden is at 4.9%, Clinton was at 3.2%. And it was mostly downhill for Clinton after this point.
Clinton's final drop was precipitated by the Comey letter. So far nothing comparable has hit Biden, and as I write this, nearly 53 million votes have already been cast. So yes, something could still change, but time is running out fast.
OK. We didn't do it earlier in the post, so here is a quick look at the current graphs for all of the states and CDs where the Election Graphs margin is currently under 5%. I won't comment on all of them individually. Some of them have switched categories in the last few days, some have not. But all of these should be considered as active possibilities for both candidates at this point.
And that is it for today. More soon.
10.9 days until the first results start coming in on election night.
We are almost at the end!
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>The last blog update here was 10 days ago on October 10th. Here are the highlevel changes since that last post:
Model  Metric  10 Oct  20 Oct  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +78 Biden +174 Biden +276 
Biden +32 Biden +140 Biden +254 
Trump +42 Trump +34 Trump +22 
Trump Win Tie Biden Win 
0.0% 0.0% 100.0% 
0.3% 0.0% 99.7% 
+0.3% FLAT 0.3% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +12 Biden +176 Biden +326 
Trump +52 Biden +164 Biden +294 
Trump +40 Trump +12 Trump +32 
Trump Win Tie Biden Win 
5.2% 0.0% 94.8% 
23.3% 0.0% 76.7% 
+18.1% FLAT 18.1% 

Categories  Trump Best Expected Biden Best 
Biden +20 Biden +212 Biden +288 
Trump +40 Biden +164 Biden +294 
Trump +60 Trump +48 Biden +6 
Tipping Point  Biden +6.2%  Biden +3.1%  Trump +3.1% 
We'll hit the main story for this update before going through the cavalcade of all of the charts:
Last time we said "not too fast" to notions that Trump's numbers were collapsing. Instead, we suggested that Trump was just at the low end of the normal range we had been in since June, and it would be unsurprising to see some reversion to the mean.
And that is exactly what we have seen. Things have moved back toward Trump on almost all metrics we track. In fact, in the critical tipping point metric, we've moved out of the 4% to 6% Biden lead band that we have been in since June, as Biden falls to 3.1%.
In the three election cycles we have tracked this, the largest difference between the final tipping point and the actual tipping point in the election was not in 2016. It was in 2008. Nobody cared, because the actual results were a bigger Obama win than predicted by the polls, rather than having a different winner than the polls predicted. But in 2008 the final tipping point was off by 3.45%. That is a bigger error than the 3.1% that currently separates Biden's tipping point from a Trump win.
So we are in the zone where simple polling error could make the difference to who is leading, even without further "movement".
As a consequence, the chances of a Trump win in the Uniform Swing view (the most optimistic for him) have jumped up to 23.3%.
Wow. This is a big change.
Reader Jonathan T emailed to ask if I had any thoughts on the possible causes of this big change. So let's talk about that.
Rather than look for a specific "cause" though, it is worth discussing if this change is even "real".
The tipping point change is driven by one state. Pennsylvania.
As polls that were very favorable to Biden from early October drop off the average, they are being replaced by new polls that show a much narrower race.
Now, what are the actual polls currently in my Pennsylvania average?
Why do I bring up the specific pollsters? I never bring up the specific pollsters. I just throw them into the average.
Well, I bring it up because right now we have no bigname highquality pollsters in the mix. In fact, we have Trafalgar, which is widely panned as intentionally constructing their polls to find "hidden" conservative voters, and therefore push results to the right, show up twice! And we have others that people have criticized as being lower quality for one reason or another.
538's Nate Silver tweeted this earlier today:
Folks, Biden's lead didn't shrink from 7.3 points to 3.6 points in PA in a week (as per RCP) at the same time it was steady or slightly growing nationally. This is why you need poll averages that take a longer time horizon and/or adjust for house effects.
RCP's averages are extremely subject to who happens to have polled the state recently, which is often the spammier, lowerquality pollsters, and that's been especially true recently with livecaller polls not having been terribly active in the states over the past 2 weeks.
I love many things about RCP, but if you have an average and 1/3 of it consists of Trafalgar and InsiderAdvantage and 0% of it consists of livecaller polls, it's not going to be a very reliable average.
He is talking about the RCP Pennsylvania average, not Election Graphs. We're too small for 538 to notice. But all of the same things are true for us.
Both Election Graphs and RCP are straight numerical averages without weighting for historical pollster quality or correcting for historical pollster bias. And we both decide which polls to include in ways that result in looking at narrower time windows as the election approaches.
These are valid criticisms. This may be a temporary transient spike caused by a series of polls from lowquality pollsters which will immediately move back in the other direction as soon as the bigger more respected pollsters put out some new numbers.
If I had to bet right now, I'd actually bet on that. We moved from the high end of Biden's range in PA to the low end of his range, and I would expect to see it revert back to the middle since we have been in a pretty steady range for months, and this seems to be an aberration, especially since there are no big news events, and as Nate Silver points out, we haven't seen a similar movement in the national polls.
But…
Let's do a quick look at where a bunch of websites ended up right before Election Day in 2016. This is from a postmortem I did of the 2016 performance of Election Graphs. At the time I logged the following as the final electoral college predictions from a bunch of sites:
Hmmm. Who got closest to the actual results? Election Graphs and RCP.
And specifically, WHY did that happen? My hypotheses are:
I could be wrong, I have not done an in depth analysis, but at first blush, those seem to be the common elements.
Now, as I said, I would still bet on reversion to the mean here, and that we will see Pennsylvania bounce back toward a greater than 5% Biden lead over the next week or so as new polls come in.
But in 2016, right before the end, I doubted the results of my own average because it was moving in a way that most of the big sites were not in that last week and because there were other sites specifically calling out Trafalgar and others as garbage noise that maybe should just be excluded from the averages because they were clearly biased and wrong. But it turned out those polls were closer to what actually happened than some of the others.
So we're not doing that this time. We throw in all the polls, and we see what happens, and yes, near the election we have a very short time frame, so what polls have been in the field lately does make a big difference. But we are where we are.
At the moment Election Graphs shows a significant tightening in Pennsylvania. And because Pennsylvania is the tippingpoint state, and there is somewhat of a gap between the states that are closer than Pennsylvania and the states where Biden has a more solid lead, that means that as Pennsylvania moves, so does the national race, at least for the moment.
Don't be surprised if this moves back in the opposite direction tomorrow though. And don't be surprised if the highquality polls confirm this movement and it stays tight either. I view that as less likely, but certainly not impossible. I'm not going to preemptively say to ignore this tightening as clearly not real though. In 2016, it was an indicator of actual tightening at the end of the race.
Or maybe Election Graphs and RCP were just lucky in 2016. That might also be the case. This is VERY POSSIBLE!
Anyway, that is the big story of the week.
But we still have to review the rest of the main charts! So here we go!
First up, states that moved in or out of our "Weak Biden" and "Weak Trump" categories:
OK, we already talked about Pennsylvania, but here it is again. It moved from Strong Biden to Weak Biden since the last update, and as the current tipping point state drives a lot of the national picture too. But we discussed all that, so… moving on…
Last time Ohio had moved into the blue zone by the thinnest of margins. Now it returns to "Weak Trump" where it has usually been. Either way though, Ohio is extremely close.
Since last time, Georgia popped over to the Trump side of the centerline, but it didn't last long, and Georgia is back to being just barely blue. Just like Ohio though, the truth is that Georgia is extremely close and could easily go either way.
Since the last update, there has been significant weakening in Wisconsin, with it just barely moving into the "Weak Biden" category with the last poll. But if you look at the specific polls in the Wisconsin average, you see EXACTLY the same sort of issues we discussed with Pennsylvania.
If the tightening in Pennsylvania turns out to be a mirage based on having a bunch of lowquality pollsters releasing results recently, then most likely it will also be a mirage in Wisconsin. If Pennsylvania turns out to be real on the other hand, then Wisconsin probably will be too. (Thus showing why assuming completely independent states is not realistic, and we need to balance that by also showing uniform swing.)
But like Pennsylvania, if I had to bet right now, I'd say that Wisconsin has been in a 4% to 8% range for most of the last few months, and we are just at the low end of that range for Biden right now, and we'll probably revert back to the middle of that range with a few more polls.
But we won't really know until we indeed get a few more polls.
Like Ohio, last time Iowa had popped over to the blue side of the line, but this time it reverts to being just barely on the red side like it usually has been. But like Ohio and Georgia, the bottom line is Iowa is too close to call.
And finally, Trump's lead drops below 5% in Alaska, bringing it into range as a possible pickup in a Biden landslide scenario.
OK, now the rest of the close electoral votes we haven't already discussed, without additional commentary:
OK. With those out of the way, let's look at our three "envelopes" with the ranges of possibility in the categorization view, and the two extreme probabilistic views:
Unlike the tipping point, where it looks like a breakout from the normal "since June" range, all three of these views show us basically consistent with where this race has been since July.
In all three cases, Trump's high end does look a bit better than it did a few weeks ago and is near the high end of Trump's range. And in all three the centerline is closer to the worst of Biden's range. But none of these are breaking out from the normal range.
Right now that "breakout" is only showing up in the tipping point. Which means that while the range of possible options is about the same, it is more precarious, because Biden's lead in the tipping point state is a lot less than it had been.
Subject again to all the caveats we discussed earlier in the post of course. I won't repeat them here.
Bottom line, carefully watch the next polls in Wisconsin, Pennsylvania, and North Carolina… the three states currently closest to the tipping point… to see if the tipping point reverses itself right away, or if it starts to look like the new closer tipping point might be real and sustained.
In the meantime, here is the current spectrum of states where the margin in the averages are under 10%:
The RCP average of national polls currently shows an 8.6% Biden lead. Comparing to the 3.1% tipping point, this implies that the structure of the electoral college is currently giving Trump a 5.5% head start… if this tipping point holds up. This "head start" is up from 3.4% in the last update.
The increase in this head start is due to the margin in the tipping point of Pennsylvania dropping considerably while the national margin, while it has tightened a bit, does not show the same kind of movement. Which again, is one reason to suspect maybe the Pennsylvania movement will prove to be an illusion. More polls will resolve that question soon enough.
Finally, time for the 2016 comparison:
In the expected case, where each candidate wins every state where they lead the average, Biden continues to outperform Clinton at the same point in time four years ago, although not by as much as he was a month ago, and there was a short moment where he moved above her curve. But for the most part, Biden has been outpacing Clinton on this metric.
Meanwhile, the same can't be said at the moment for the tipping point.
While the general trend of Clinton weakening started at about the 30day point, at about this moment, two weeks out, Clinton had a bunch of good polling and had one final peak before her final collapse. Meanwhile, as we have discussed, Biden's tipping point is now the smallest it has been since June 10th.
If this is just an aberration, and it gets reversed or erased by new highquality polls that come out in the next few days, then the picture will be back to what it has been for months. Namely, Biden is heavily favored, but a Trump win is not impossible.
If however new polls sustain this change, then it would indicate that Trump's chances of winning have increased substantially, and there is a real race happening in these last two weeks.
Watch the next few days of polling, and we should know which scenario we are in.
Right now, with our current averages, and the two extreme probabilistic scenarios, Election Graphs gives Trump's odds of winning as somewhere between 0.3% and 23.3% depending on how correlated the polling errors in each state end up being.
That's a big range of course, and that is "if the election was today" when we have two weeks left. But at the very least, it means to take the chances of a Trump comeback and win seriously.
We shall see.
Finally, the map as it stands right now:
I locked the poll updates on Monday evening US time to make this update. As I finish writing the post it is now Tuesday afternoon. As usual for this point in the cycle, there have already been a bunch of new polls released during that time. So I'll be getting back to data entry shortly.
I've taken the remaining time from now until the election off from the day job to better be able to keep up with the deluge of new polls. That also probably means you'll get more than just one update here on the blog before election day.
So keep checking in for updates, both here on the blog and on the main 2020 Electoral College page.
But first, the usual closing with the countdown:
14.1 days until the first results start coming in for Election 2020.
We are in the home stretch. Almost there now!
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>The last blog update here was 11 days ago on September 29th. Here are the highlevel changes since that last post:
Model  Metric  29 Sep  10 Oct  𝚫 
Probabilities (Indep States) 
Trump 2σ Median Biden 2σ 
Biden +46 Biden +144 Biden +256 
Biden +78 Biden +174 Biden +276 
Biden +32 Biden +30 Biden +20 
Trump Win Tie Biden Win 
0.1% 0.0% 99.9% 
0.0% 0.0% 100.0% 
0.1% FLAT +0.1% 

Probabilities (Uniform Swing) 
Trump 2σ Median Biden 2σ 
Trump +84 Biden +168 Biden +312 
Trump +12 Biden +176 Biden +326 
Biden +72 Biden +8 Biden +14 
Trump Win Tie Biden Win 
7.5% 0.0% 92.5% 
5.2% 0.0% 94.8% 
2.3% FLAT +2.3% 

Categories  Trump Best Expected Biden Best 
Biden +8 Biden +168 Biden +294 
Biden +20 Biden +212 Biden +288 
Biden +12 Biden +44 Trump +6 
Tipping Point  Biden +5.4%  Biden +6.2%  Biden +0.8% 
To summarize before we get to review all the current graphs:
These 11 days have been very good for Biden in the polling. He hasn't moved past his alltime best marks, but he is once again near the top of the range we have observed over the last year. The flip side of course is that Trump is near the bottom of his range.
The key metrics right now show that if the election was today (it is not) Trump's chance of winning is no more than 5.2%, probably less, and Biden's tipping point is up to a 6.2% lead, which is the best it has been since midAugust, and which is better than Clinton saw any time in the last 100 days before the 2016 election.
But I wouldn't call this a Trump collapse quite yet. He has returned to the low end of the range we have been seeing for months. We are not breaking new ground, seeing things worse for Trump than we have ever seen. At least not yet.
From here it is quite possible that we get a reversion to the mean, with Trump improving a bit, rather than seeing further Trump deterioration. Does it look bad for Trump? Yes. Certainly. But not in an unprecedented way, and he could still make up ground before the end.
OK, now for the parade of graphs…
As usual, we'll look at the critical states, then the nationwide electoral college trend charts, the win odds graphs, the spectrum of states, then finally close it up with the comparison to 2016 and a map. Here we go.
This time there were 9 states that moved in or out of the "Weak Biden" or "Weak Trump" categories since the last blog post:
Biden's lead in Florida very briefly popped over 5% into the "Strong Biden" zone, but then dropped back down again into "Weak Biden". So no net change to the category. Biden remains strong in Florida, but not so strong that it is completely out of Trump's reach.
Biden's current 4.2% lead maps to an 84.9% chance of winning the state given the previous accuracy of Election Graphs averages.
Ohio is pretty much as close to a tie as possible. Since the last update, it moved briefly from the Biden side to the Trump side, and then back again. No net change.
Because of the order of the polls, the brief trip across the line isn't visible on the chart anymore, but it did happen. Either way though, Ohio is too close to call.
Looking at the specific margin, Ohio is actually in an oddball zone. Biden leads by 0.1% at the moment, but our win odds are based on the results in 2008 through 2016, and looking at those years, a Democratic lead this small actually means the Republican is still slightly favored. We currently have Trump at a 53.2% chance of winning Ohio based on this poll average.
At the time of the last blog post, Georgia had moved to be just barely "Weak Trump". It is now "Weak Biden" again, where it has been most of the time since July. Biden now leads by 2.0% in the average, which we have as a 67.9% chance of winning.
Only one is left visible on the graph, but since last time Michigan went from Stong Biden to Weak Biden, then Strong then Weak, then finally Strong again before Biden's lead increased up to 8.6% as we write this, firmly within the Strong Biden zone. This corresponds to a 99.1% chance of a Biden win. If nothing changes before election day of course.
Wisconsin is another back and forth with no net category change. It dipped into "Weak Biden", then moved back into "Strong Biden" where it has been most of the time since July. Biden currently has a 6.2% lead, or about a 94.8% chance of winning.
Since the last update, Trump's lead in South Carolina very briefly dipped below 5%, then went back up again. Once again, this is a chance that was actually erased by subsequent polls. Right now, Trump leads by 5.3%, which means about a 97.3% chance of winning.
Biden's lead in Nevada grows to 5.3% or a 91.7% win chance, pulling it out of the "Weak Biden" zone.
For the second time in the last few months, Iowa slips to the blue side of the line. Barely. The poll average has Biden at a 0.4% lead, which would be a 50.8% chance of winning the state. This is clearly a tossup, but also given the history, it would not be surprising at all to see Iowa slip back into the red with the next polls that come in. The "usual" situation for Iowa seems to be just barely Trump, and it is way too early to make any determination that there has been a lasting change.
Last time we had Montana move into the "Weak Trump" category. A series of poll updates covering older time periods, including a correction to one outlier data point, completely erased that change. And by the time of this post, Trump's lead was actually up to 10.1%, putting the state not just into "Strong Trump" but all the way into "Solid Trump". Trump's chance of winning the state is now 99.9%.
So yeah, the Democrats probably won't be flipping Montana this time around.
Beyond the states that changed categories, here are the rest of the states that are currently in the "Weak" categories, but did not shift their classifications at all this time around. With no additional commentary.
With all the state views out of the way, time for the three different models for the range of electoral college possibilities.
The centerline of this chart is at a 212 EV Biden win, the best it has been since midSeptember, but overall, this shows the steadiness of the race. The ups and downs we are seeing are a small handful of very close states going back and forth across the zero line as the polls jitter.
Looking at the best cases… where one candidate or the other wins ALL of the close states… Biden's bestcase very rarely moves at all. Trump's bestcase moves a lot more. This is a factor of there being a lot more states close to a 5% Biden lead than there are states close to a 5% Trump lead. But even there, the Best cases have just bounced in a range since July. There is no clear trend.
Next up, the probabilistic view using "uniform swing" where we force all of the states to move up and down together. This gives both the maximum chances of an upset and the maximum chances of the current leader getting a landslide.
And we see the same thing. Yes, improvement for Biden lately, but well within what has been the normal range of possibilities lately.
The only place where you can claim a breakout is the top of Trump's 2σ (95.45%) range, which is down to a 12 EV Trump win. That is indeed the worst we have seen for Trump on that line in this election cycle.
This range does still include Trump winning, but barely. This version of the model currently gives Trump a 5.2% chance of winning.
And then the "Independent States" view. This does a simulation assuming what happens in each state is completely independant of what happens in all the others. So if the polls underestimate Trump's support in Ohio, it means nothing to what may happen in Pennsylvania, etc. This results in the tightest range of possible results.
On this view as well, Trump has declined the last few weeks, but we aren't breaking new ground.
We are however in the territory where even the 3σ (99.73%) range does not include the possibility of Trump winning. In fact, the standard display on the 2020 Electoral College page now rounds Biden's win odds to 100%. Looking at the unrounded simulation results, we actually have 99.996% Biden wins, 0.0027% Trump wins, and 0.0013% ties.
Both the "Uniform Swing" and "Independent States" represent the extremes of how correlated or not state results are to each other. The truth is somewhere in between, but where in between is not predicted by these models.
So the Election Graphs statement on odds right now is simply that Trump's odds of winning are between 0.0027% and 5.2%.
So let's look at the odds charts.
So, uh, yeah. We probably won't show this one much unless Trump starts spiking and there is something to see. In the Independent States view, Trump has been so close to zero for so long that you can't even see any red on this chart.
On uniform swing, there is at least something to see. But again, we are bouncing around in a welltrod range. We're currently at a 5.2% Trump win probability in this view. The range on this chart (currently showing the race since July 12th) is from a 4.6% to a 16.9% chance of a Trump win.
Now, to be clear, there is a big difference between a 4.6% chance of winning and a 16.9% chance of winning. But that seems to be the range we have been living in for Trump's maximum odds of winning. We are still in that range, although near the low end.
And now the tipping point. This represents how much the national popular vote would need to move… assuming a uniform swing of states… in order to flip the overall result. Once again, we see us near Biden's best, but not quite setting records. Right now the 6.2% is the best Biden has seen since August 11th. But he was close to this level as recently as September 19th.
Unless there is a breakout, this is once again a picture of a steady race. Yes, Biden is near the top of his range, but basically, things have been flat for months. There is no overarching trend.
And there is the center of our spectrum of states. There are a LOT of close states right now. Trump winning all of them isn't enough. He has to also pull in Nevada and Wisconsin. Or some of the states where he is behind by even more.
The RCP average of national polls currently shows a 9.6% Biden lead. Comparing to the 6.2% tipping point, this implies that the structure of the electoral college is currently giving Trump a 3.4% head start. This is up from 0.7% in the last update.
The increase in this delta since last time would indicate that a lot of Biden's gain in the national popular vote is coming from states that don't matter. Biden winning by a bit more in New York, or losing by a bit less in West Virginia is simply irrelevant given how we actually choose presidents.
Finally, let's do the 2016 comparison.
Reader Ali D recently asked in the post comments if I could do 2016 comparisons based on the probabilistic views in addition to what I have been doing here. He wasn't sure the expected case was a good comparison since it could change so much based on close states crossing the centerline.
The short answer is no. The probabilistic views are all new for 2020, so I have no 2016 line to compare to. Sorry!
Going into more detail, the estimates of how close to reality the poll averages end up that drive the probabilistic analysis are based on looking at the 2008 through 2016 results, so even though I once did it as a oneoff, applying that analysis to 2016 would not be valid since you are essentially using the 2016 results to predict 2016.
It theoretically would be possible for me to generate separate state win odds using only 2008 and 2012 data using the same method I used, then retroactively generate a full probabilistic trend for 2016, but even doing this with only 3 election cycles of data seemed a little iffy, doing it with two really would be. Plus, realistically, I'm not going to have time to do anything like that before election day.
So once again, sorry!
Ali D is also correct that JUST looking at the expected case doesn't give you a good sense of how close things really are since you don't know if the states making up the winning margin have the winner ahead by 0.1% or by 1% or by 10%.
This is why we look at BOTH the Expected Case and the Tipping Point to understand the comparison… I guess we could expand this to also compare the best cases for both candidates, but things are already complicated enough… so we'll stick to just these two.
Here we go…
In the expected case, Biden has been doing better than Clinton ever since we hit the 76 day out mark. We are now past Clinton's last surge, and into the part of the race where her lead had started to slip.
So far, Biden's lead is not slipping. It is increasing.
But let's look at the tipping point comparison:
In this metric, between 37 and 29 days out, when Clinton was at a high point and Biden was at the low end of his range, Biden was actually doing worse than Clinton was at the same point. But that Clinton peak is over, Biden is gaining, and so Biden once again is stronger than Clinton was at this point.
We certainly can't say it is impossible for this picture to change before the election. But for Trump to actually be FAVORED to win, Biden needs to start collapsing soon, and he needs a bigger and faster collapse than Clinton suffered in 2016.
Unlike Clinton, who regularly fell to levels where the race was tight and she was favored but Trump clearly had a very good shot, Biden has been consistently strong all the way along. So to let Trump take the lead, he would have to do significantly worse than we have ever seen, not just a little bit worse.
Or course Trump doesn't have to be favored in order to win. With existing polling, the Uniform Swing view gives Trump a 5.2% chance of an upset. And 5.2% is NOT ZERO.
Embedding the most recent XKCD because it is directly relevant:
Of course Election Graphs is actually saying that the chances of a Trump win are LESS than 5.2% if the election was today. Even so, it is STILL NOT ZERO.
Plus there is still time for things to change. Over 9 million people have already voted. Voting is well underway. But as I write this we have a little over three weeks until the the end of voting and most people haven't voted yet. While it seems increasingly likely that it will be hard for Trump to actually get to a position where he is the favorite on election day, he may very well improve his odds in that time.
There still could be an October Surprise that makes a difference!
We shall see.
Finally, the map as it stands right now:
Everything above represents the situation as of when I started writing this blog post. Looks like at least two new polls have been published while I have been writing. So time to close this out and start updating the site again…
But first, the usual closing with the countdown:
24.1 days until the first results start coming in for Election 2020.
Tune in for the inevitable ups and downs that happen in those days…
This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a pollbased estimate of the Electoral College. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for current interactive versions of the chart, along with additional details.
Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of updates. For those interested in individual poll updates, follow @ElecCollPolls on Twitter for all the polls as I add them. If you find the information in these posts informative or useful, please consider visiting the donation page.
]]>