Republican Delegates after NV & VI

Trump gets all 4 delegates from the Virgin Islands.

Haley got 26% of the vote there, but the Virgin Islands are Winner-Take-All, so that didn't matter.

Trump gets all 26 delegates from Nevada.

Haley got 30% in the primary vote there, coming in 2nd to "None of these Candidates", but that didn't matter since in Nevada, the primaries were not how delegates were allocated.

Instead, they were allocated by caucus, where the only other candidate besides Trump was Ryan Binkley, who got 0.71% of the vote. Never heard of him? Yeah, I hadn't either. Here is his website.

Anyway, that means Trump got 100% of the 30 delegates that were available tonight, which starts moving his "% of remaining delegates needed" number away from the 50% line, while Haley (and the others)  zoom upward.

Here are the most important charts and graphs:

Click on any of the above for the rest of the charts, and for the Democratic side too.

Next up: the Republicans in South Carolina on the 24th (winner-take-all), followed by the Democrats in Michigan on the 27th (proportional).

157.6 days until the Republican National Convention

192.6 days until the Democratic National Convention

Democratic Delegates after Nevada

As expected, Biden gets all 36 delegates.

There is no serious competition of course, so while I'll keep updating, we are just going to see a Biden march to clinching the nomination.

Next up will be the Republicans in Nevada and the Virgin Islands on Thursday, which is expected to be equally boring from the other side.

In any case, here are the graph and table:

159.7 days until the Republican National Convention

194.7 days until the Democratic National Convention

Democratic Delegates after South Carolina

Well, this is boring.

Biden gets all the delegates from South Carolina.

This was also completely expected of course. The only place where the token opposition of Phillips and Williamson were expected to be able to get a non-trivial amount of support was in New Hampshire, where Phillips got 19.66% of the final vote, and Williamson got 4.05%, just ahead of write-in votes for Republican Nikki Haley at 3.84%. But of course no delegates were awarded based on that vote. New Hampshire will undoubtedly eventually get to send delegates to the convention, but how they will be allocated is yet to be seen.

Here in South Carolina, as of a few hours after poll closing, the partial count has Biden with 96.22%, Williamson with 2.08%, and Phillips with 1.71%. To get any delegates, Phillips or Williamson would have had to do MUCH better than that, either state wide, or at least in one of South Carolina's 7 congressional districts. Either way, they were not even close.

So Biden gets all 55 delegates from South Carolina, and starts on what will likely be an uninterrupted journey toward clinching the nomination on March 19th, the first date where it will be mathematically possible.

We'll track the updates as they happen from now until then, but unless something very unexpected happens, there won't be any drama.

Here is the main "% of remaining delegates needed to win" chart:

And the tabular summary of where things are:

Next up, Democrats in Nevada on Tuesday night.

It will probably be just as boring. But we will be here to confirm!

162.8 days until the Republican National Convention.

197.8 days until the Democratic National Convention.

Delegates After New Hampshire

On the Democratic side, New Hampshire has been penalized and only has 10 delegates, which theoretically won't be allocated by the results of today's primary, but maybe eventually will be, just not directly. In any case, we go by The Green Papers, and at least so far, they haven't estimated any delegates there. So lets stick to the Republicans, because that is where the action is anyway.

So what happened there? Well, Trump won New Hampshire as expected, although perhaps Haley made it a little closer than Trump would have liked.

Here is the full state breakdown so far:

Which makes the overview look like this:

I explained last time that the "% of Remaining Delegates Needed to Win" column there is the most important to watch, so here is the chart of that:

Ramaswamy and DeSantis are of course racing upward out of contention since they have dropped out.

With her showing tonight, Haley managed to not have her situation deteriorate TOO much. She went from needing 50.52% of the remaining delegates to win, to needing 50.57% of the remaining delegates to win. So she didn't improve her delegate position, it continued to deteriorate, but not by all that much.

Meanwhile, Trump improved from needing 50.02% of the remaining delegates to win, to only needing 49.98% of the remaining delegates, which is his first time under 50%. But just BARELY. His line has yet to start diving down toward 0%.

But it probably will.

The next contests on the calendar that actually allocate delegates on the Republican side will be the Nevada Caucus on February 8th (not the Nevada primary a few days earlier, which doesn't matter) and Caucus in the Virgin Islands. Haley registered for the Primary that doesn't matter in Nevada, but not for the Caucus, so Trump will likely win all 26 delegates there.

Then the next real competition with both Trump and Haley will be in South Carolina. Recent polls there have had Trump far ahead, and it is a winner take all state.

The only real hope on an actually interesting delegate race that isn't just a coronation for Trump is if somehow Haley's finish in New Hampshire was "close enough" to Trump that it shakes up Republicans in South Carolina (and beyond) and they start abandoning Trump in droves. Which seems really really unlikely.

We'll have the Democrats in South Carolina and Nevada before that though. So we'll see you again for that. Of course that will almost certainly just be walking to a coronation for Biden on that side of the fence.

So… yawn!

Wake me up if something interesting shakes things up.

173.6 days until the Republican National Convention

208.6 days until the Democratic National Convention



Trump Wins Iowa Caucus (as expected)

So far, all the posts here on Election Graphs in the 2024 cycle have been about looking at state level polls for the general election in November.

That ends today, as the second part of Election Graphs opens up, as the delegate race begins on the Republican side.

The numbers may still shift a bit as all the T's are crossed and all the I's are dotted for the final counts, but the preliminary delegate counts at The Green Papers haven't shifted in a couple of hours, so that's good enough for now, lets start looking at the new charts and graphs.

Many places will have tables showing the delegate totals. Here is ours:

The first column shows how many delegates each candidate has right now, the second shows that as a percentage of the delegates so far, the third shows the percentage of the remaining delegates each candidate needs to win, and the last column is how much better than their performance so far each candidate needs to do to win.

The delegate race page shows graphs for all of these, but the main one, the one that matters the most, is the "% of remaining delegates needed" graph:

OK, we've only had the one delegate estimate from Iowa so far, so this is just a bunch of straight lines right now… kinda boring. It will get more interesting as we get results from more states.

Well, how interesting will depend on if we get any surprises in the race, but it will stop just being straight lines in any case.

Let me explain what you are looking at.

First, we specifically show graphs showing how those numbers evolve over time, not just "now". The data table is nice and all, but to get a sense of how things are going, it really helps to see how things have been changing. Is someone catching up? Or falling behind?

Second, while we do have a page that shows things in a date-based way, the primary view has the % of delegates that have been allocated on the x-axis rather than the date. This gives a much better sense of how far though the process we are. After Iowa, the Republicans have allocated 1.65% of their delegates. We are still a very long way from the end. On some primary days, only a small percentage of delegates will be allocated, but on the other extreme, 35.99% of the delegates will be allocated on Super Tuesday alone.

Third, the y-axis is of course the % of the remaining delegates that each candidate needs to win.

To me, the other graphs are nice, but this is the most important metric of all.

Why is this the most important thing? Because the absolute numbers can be misleading, because as more and more delegates get allocated, there is less and less opportunity for whoever is trailing the leader to catch up, and the leader has to do less and less in order to coast to the victory.

Just like any race, if you are behind, the less track you have left before the finish line, the faster you have to go to in order to catch up to the winner before they get to the finish line. It does you no good to catch up to them after they have already won.

The "% of remaining delegates needed to win" graph shows that more clearly than anything else.

Even though only a very small number of delegates have been allocated so far, you can already see that in the chart above. Everybody starts out needing just barely over 50% of the delegates to win.

But now with just a few delegates estimated, Ramaswamy, who actually already dropped out due to his 4th place finish here, would now have needed 50.73% of the remaining delegates to win. Haley needs 50.52%, DeSantis needs 50.48%.

Meanwhile Trump, after getting exactly half of the delegates from Iowa, only needs 50.02% of the remaining delegates to win.

As we continue through the primaries and caucuses, the losers will have their numbers move toward 100% faster and faster, while the winner will start dropping down toward 0%. This represents those who are behind needing to get a larger and larger percentage of the delegates to catch up, while the person in the lead needs less and less.

Later on in the race, it is very possible that a candidate could win a state, but not by enough to be on a pace to catch up, so even though they won, they still end up falling further behind.

In the epic 2008 race between Obama and Clinton, this happened to Clinton over and over in the second half of the race. She was winning states, but not by enough to be on a pace to catch up  to Obama.

As each losing candidate hits 100% and then goes beyond it, they are eliminated.

When the winning candidate gets to 0%, they have clinched the nomination.

OK, that sounds complicated. It really isn't. It will become clearer as the race progresses.

But if you really like raw counts better, it is there too:

On there is it just about who gets to 1,215 first.

In any case, next up New Hampshire, including delegates for the Democrats too, despite the penalties the Democrats put on New Hampshire for not moving their primary later in the year.

It is very very possible, probably very likely actually, that there will be no real drama in either party's delegate race this year. Even if Haley puts on a good show in New Hampshire next week.

Regardless, we'll track it all here, and see how it goes.

181.6 days until the Republican National Convention

216.6 days until the Democratic National Convention

Off we go.

295 Days Out: Peak Trump?

Here we go. It has been another 50 days, so time for another blog post update on the 2024 race. I'll do these more often when we get really close to the end, but for now 50 days still seems good.

Remember, here on the blog I try to do periodic narrative updates, but I update the live status on as quickly as I can when new state level polls come up, so the graphs and charts there will always be more up to date than these blog posts. Also follow Election Graphs on Mastodon for updates every single time a poll is added, plus daily status summaries.

As usual, I'll start with a summary.

  • Bottom line: There may be lots of time left for things to play out, but Trump is doing great right now, and Biden is… not.

OK, a little bit more:

  • The last 50 days of polls have continued to move in Trump's direction. Biden's best position of the last year was in June. The next three months were not great, but things were mostly steady. Then the bottom fell out around the start of October, and things have only gotten worse since.
  • Some of the graphs have started to turn around and move toward Biden again in the last month or so, but due to the winter holidays polling has actually been pretty sparse, and so it is too early to tell if this is the start of a new sustained trend, or just a temporary bump.
  • The tipping point right now is Trump ahead by 4.0% in Michigan. That still represents a close race. 4.0% is an amount that can disappear in less than a week depending on news events, or if polls are wrong in the opposite direction this year compared to 2020 or 2016, that lead could be a mirage.
  • Polls this far out ARE NOT predictive of the final result. However, we can see where things stand today and compare to previous cycles. Trump continues to massively outperform his state level polling at the same time in the 2016 and 2020 cycles.
  • In 2020, Biden won Georgia and Arizona. At the moment, those two states don't even look competitive, with polling showing strong leads for Trump in both.
  • In 2020, Biden also won Pennsylvania, Michigan, Wisconsin, and Nevada. Trump is currently showing narrow leads in all four. These states have been battlegrounds before, and clearly will be again this time. To get things moving back in his direction, Biden needs to start here.
  • Some Democrats are acting as if Trump has a weak hand and any current good polling for him is just an aberration and will disappear as the election approaches and Trump's legal issues continue to unspool. Or they point to special elections and abortion referenda where they have over performed relative to expectations. Or to the large number of undecided voters that they think will end up breaking for Biden in the end. Perhaps they will be right on all that. But right now Trump is ahead, and it is foolish to underestimate him.

With that out of the way, anybody who only cares about the summary should leave. For those of you who are left, I'll dump a lot of charts and graphs…

Here is what the spectrum of close states (poll average margin 10% or less) looked like when I did the last blog post on 2023-11-24:

And here we are on 2024-01-15:

Categorizing the movement of all of these states:

Movement toward Trump:

  • Maine-CD2 (1 EV): Biden by 1.6% -> Trump by 2.4% (Trump+4.0%)
  • Maine-All (2 EV): Biden by 11.1% -> Biden by 7.8% (Trump+3.3%)
  • Michigan (15 EV): Trump by 1.1% -> Trump by 4.0% (Trump+2.9%)
  • North Carolina (16 EV): Trump by 4.0% -> Trump by 6.2% (Trump+2.2%)
  • Virginia (13 EV): Biden by 6.0% -> Biden by 4.2% (Trump+1.8%)
  • Arizona (11 EV): Trump by 4.4% -> Trump by 5.8% (Trump+1.4%)
  • Florida (30 EV): Trump by 7.3% -> Trump by 8.7% (Trump+1.4%)
  • Minnesota (10 EV): Biden by 5.3% -> Biden by 4.3% (Trump+1.0%)
  • Nevada (6 EV): Trump by 2.8% -> Trump by 3.8% (Trump+1.0%)
  • Georgia (16 EV): Trump by 5.7% -> Trump by 6.7% (Trump+1.0%)
  • Iowa (6 EV): Trump by 9.2% -> Trump by 10.0% (Trump+0.8%)
  • New Hampshire (4 EV): Biden by 8.6% -> Biden by 8.0% (Trump+0.6%)
  • Colorado (10 EV): Biden by 7.8% -> Biden by 7.7% (Trump+0.1%)

No change:

  • New Mexico (5 EV): Biden by 9.3%
  • Nebraska-CD2 (1 EV): Trump by 4.7%
  • Ohio (17 EV): Trump by 9.2%

Movement toward Biden:

  • Wisconsin (10 EV): Trump by 0.1% -> Trump by 0.0% (Biden+0.1%)
  • Texas (40 EV): Trump by 5.8% -> Trump by 5.4% (Biden+0.4%)
  • Pennsylvania (19 EV): Trump by 1.2% -> Trump by 0.6% (Biden+0.6%)

Just like last time the moves were overwhelmingly toward Trump.

Wisconsin and Pennsylvania, both important states given how close they are, did go the other way by a small amount, but the impact of this was overwhelmed by so many other states moving against Biden.

The only thing that actually flipped from Biden to Trump since 50 days ago was Maine-CD2. This was due to the very first poll for this congressional district this cycle, showing  a 14.0% Trump lead, displacing the 2004 election results in the average. In 2004 Kerry beat Bush by 5.8% there.

But in 2016 and 2020 Trump won ME-CD2 by healthy margins, so this is essentially just the new poll in the average confirming that the move toward the red in this district is likely to continue in 2024, and now that it has moved into that zone, it should not be surprising if ME-CD2 not only stays red, but gets redder once there are a few more 2024 polls in the mix.

But that's just one electoral vote. Lets take a deeper look at the four states currently in the "Weak Trump" category, in order by how many Electoral votes they have:

First up, Pennsylvania. This is of course one of the states that bucked the trend and seems to be moving back toward Biden. From June through the middle of December, with a few exceptions, the movement was usually toward Trump, with some individual polls showing Trump with margins as high as 8.9%. But starting in mid-December, things started to go the other way, with 2 out of 3 polls since then actually showing Biden in the lead, and the third with only a 1% Trump lead.

You never know what the next poll will bring of course, but for the moment at least, Pennsylvania seems to be trending back toward Biden.

Repeating something I've mentioned before though, the "neither one" category, which includes both people who say they will vote for other candidates, and folks who say they are truly undecided, is huge compared to the margin between Biden and Trump. Trump leads by 0.6%. The average has 16.6% saying "neither" in Pennsylvania at the moment.

That is massive, and indicates that the situation is potentially very volatile, as both 3P support and "undecided" tend to collapse as elections approach. As an example, the "neither" category in Pennsylvania on Election Graphs on election eve in 2020 was only 3.7%.

Next up Michigan. Michigan has been moving dramatically toward Trump since July. The average has gone from Biden by 2.8% to Trump by 4.0% today. A 6.8% swing in six months.

And it looked worse a few days ago. The most recent poll in Michigan, from Target Insyght, in the field from January 4th to 10th, shows Biden ahead by 4%. But this is the first poll showing a Biden lead since early November. It is way out of line with all the other recent Michigan polls. It honestly looks like an outlier.

Immediately before this poll was added, the average in Michigan had Trump up by 5.8%… which was redder than Texas currently is in the EG average.

Michigan with Trump ahead by more than 5% is very very bad news for Biden. With Michigan at those levels, the EG "categorization view" which takes the poll averages at face value, and just says any state with a margin under 5% could go either way, shows Biden still losing even if he won ALL of those close states.

That new poll puts Michigan back as "Weak Trump" rather than "Strong Trump" though. Because while I may speculate here on the blog that maybe it is an outlier, for the averages on the site, I don't judge, and just include everything.

Maybe that new poll is actually just the start of a new trend, and it is really the earlier polls showing Trump significantly ahead in Michigan that were the outliers, in which case maybe things are starting to move toward Biden again. Who knows? But if not, Michigan is looking very challenging for Biden right now.

If Biden campaign people aren't deep diving to figure out how to tackle Michigan right now, then they are committing political malpractice. They need Michigan. (And all of the states we are talking about today actually, but Michigan is the biggest problem at the moment.)

Oh, and once again we have a massive 17.8% of people in the "neither" category. They are probably the people that Biden needs to target first.

The current EG average shows Trump by 0.0%. This is of course rounded. The unrounded average has Trump leading by 0.04%. But that level of precision is false precision. This average is really a tie.

Looking at the actual individual poll results included in the average, they range from Biden up by 5% to Trump up by 6%. The scatter on the chart is all over the place. There is no obvious trend, and we have about as many polls with Biden leading as with Trump leading. The average has been close to the tie line since the beginning of November.

I really want more polls in Wisconsin. For the moment, all I can say is that it looks like Wisconsin is close, the variation on the polls there is pretty wide, and polling there is slow enough that we can't really judge short term trends.

And of course big "neither" numbers in Wisconsin too.

Biden hasn't led a poll in Nevada since October, but looking at the polls since then, it looks like they might be trending back toward Biden after a spike toward Trump, but with Trump still firmly in the lead. Well, once again with a "neither" category much larger than the margin. But still a lead on the margin.

Nevada is only 6 electoral votes, but at the moment if you gave Biden all the states where he leads the average, plus Pennsylvania, Michigan, and Wisconsin…  you end up at a 269/269 tie, which would send the election into the House of Representatives where the NEW House that just won election would vote by state delegation, not by individual representatives… a voting structure that would probably mean a Trump win in the end, even if the Democrats win back the House.

So if Biden wants a clean win, Nevada needs to get pulled back over the line too.

OK, so now lets zoom back out to the national view, and bring in all three of the different ways EG has of looking at the national picture:

Let's look first at the "odds ranges" between the Independent States and Uniform Swing views. These are the odds of a Biden win:

  • 2023-03-25: Between 17.8% and 25.8%
  • 2023-04-24: Between 24.9% and 32.3%
  • 2023-05-21: Between 34.9% and 35.9%
  • 2023-06-25: Between 43.9% and 46.0%
  • 2023-08-13: Between 42.9% and 44.2%
  • 2023-10-01: Between 35.7% and 44.2%
  • 2023-11-26: Between 3.0% and 23.1%
  • 2024-01-15: Between 1.2% and 4.4%

This is by far the worst performance we've seen in these blog posts. The two ends of these ranges represent the difference between assuming that the polling errors are going to be the same in every state, and assuming that the polling error in each state is completely independent of the errors in other states. The truth is going to be somewhere in between.

Both probabilistic views do base their probabilities on the differences between the final EG poll averages and the actual vote totals in 2008 through 2020. For details of how all this works see my post from January 2023. The bottom line though, is the probabilistic views assume that in the closest states, the Election Graphs poll averages are underestimating the Republican by a bit, because that is how it was averaging out all the results from 2008 to 2020. If the poll error is in the other direction this time, things will of course look a bit better for the Democrat.

I should make a graph that explicitly shows the range between these two probabilistic views, but in the mean time, here are the two charts on the probabilities:

Still looking at the probabilistic view, but looking at likely Electoral College outcomes instead of the odds:

The independent states view is the most dynamic in terms of how sensitive it is to every single new poll added, so when looking for trends, it is the one to look at. Looking carefully at this chart, there is an inflection point around December 12th.

Trump is still way ahead of course, but there starts to be SOME movement toward Biden after that date. Trump had "a" peak on December 12th. Only time will tell if this was "Peak Trump" overall.

The uniform swing view is a lot less clear, but you can see a little movement toward Biden in recent weeks if you squint.

Taking the probabilities out of it, and just looking at the Classification view, you get this:

The only sign of movement toward Biden on this view so far is that the spike representing Michigan being out of reach because Trump's margin was greater than 5% was really short lived.

For the categorization view though, there is another way of looking at things, namely through the "tipping point", which is the margin in the state that puts the winner over the edge:

  • 2023-03-25: Trump by 0.7% in Wisconsin
  • 2023-04-24: Trump by 0.5% in North Carolina
  • 2023-05-21: Trump by 0.2% in Wisconsin
  • 2023-06-25: Biden by 0.9% in Georgia
  • 2023-08-13: Biden by 0.7% in Michigan
  • 2023-10-02: Biden by 0.7% in Pennsylvania
  • 2023-11-26: Trump by 1.2% in Pennsylvania
  • 2024-01-15: Trump by 4.0% in Michigan

We do have a graph of this one though!

We see the spike over 5% for the short time Trump's margin rose up to 5.6% in Michigan, moving the tipping point very briefly to 5.4% in TEXAS.

That was stunning. Texas as the tipping point state? But it didn't last long, and the tipping point is back down to a 4.0% Trump lead in Michigan.

That is still substantial though. Since I have been doing this sort of election analysis, the biggest delta between the final EG tipping point and the actual tipping point based on the election results was 3.45% in 2008. That year the polls underestimated Obama.

We have many months to go, but if we still had a Trump by 4.0% tipping point on Election Day, the overall polling error would need to be underestimating Biden this time, and be bigger than the errors we had in any of the previous four presidential elections in order to result in a Biden win.

The tipping point can also move very quickly. In 2016, Clinton's lead in the tripping point lost almost 6% in just the last couple of weeks before the election. There is the variability inherent in polling in general, plus news events really can move opinion. Leads can evaporate almost overnight under the right circumstances.

This is one of many reasons why polling this far out is not predictive, it only tells you about how things are right now, and where candidates need to put in effort to change things.

Here is the tipping point comparison with the last two election cycles:

For almost this entire election cycle Trump has been doing better than he was doing in either the 2016 or 2020 cycles. Since October, he has been doing SUBSTANTIALLY better than these previous cycles.

You can also see this looking at the center lines for the electoral college margin view:

All of the above assumes Biden vs Trump, because that currently seems like the almost certain matchup, absent something really unexpected coming up.

Election Graphs will be covering the delegate race starting with the Republican caucus results in Iowa, which will start to come in less than 24 hours from when I publish this blog post.

Expect a post about Iowa delegates once things settle down a bit in the election coverage and it looks like we have stable delegate estimates.

We will know before too much longer if the Biden vs Trump assumption about Election 2024 ends up holding true.

In the mean time, a quick look at other possibilities for the Election 2024 matchup, using the Independent States Probabilistic View median electoral college result since that is the most interesting to look at:

There is enough polling on Biden vs Trump and Biden vs DeSantis to feel like the lines above are pretty good. Trump is considerably stronger than DeSantis when pitted against Biden.

Biden vs Haley is getting there, but should still be taken with a significant dose of salt, because state by state polling for that combination is still pretty sparse, so a lot of the EG state poll averages are still based significantly on election results from previous cycles. But it looks like she PROBABLY does better against Biden than DeSantis, but not as well as Trump does.

All three of these Republican candidates are leading Biden though, the question is by how much, and how secure that lead is.

The other three combinations on the chart above have barely been polled, and should be ignored. Harris, Pence, and Youngkin aren't even running at this point anyway.

Finally, closing the post with the current Biden vs Trump map:

295.8 days left until the polls start to close on Election Day 2024.

Less than a day until the Iowa caucuses on the Republican side.

Buckle up everybody. Things are starting for real now!

345 Days Out: Trump Stronger Than Ever?

I had thought about doing another blog post update at the one year mark, but never got around to it. So I guess for now I am sticking to approximately every 50 days. We are now just under the 350 day mark. And it is a long holiday weekend, so I have some time. So here goes.

These only represent snapshots where I look back at how things have changed recently. The live current situation is updated whenever new polls come out, as quickly as I can manage given other commitments. That live status is now the main page of instead of this blog, since it is always more current.

Anyway, the TL;DR of where things are right now:

  • The last 50 days of polls have been great for Trump and horrific for Biden. Specifically starting in early October there have been poll after poll after poll showing Trump leading in critical states.
  • With this change, the situation has moved from "A toss up race with Trump slightly favored" 50 days ago to "If the election was today, Trump would be the heavy favorite."
  • Although down a little bit from his peak at the end of October, Trump is very near his all time best polling compared to his Democratic rivals, not just in this cycle, but also far outperforming his own numbers in both 2016 and 2020.
  • That fact should NOT be underestimated. Because of that, Trump is most definitely favored right now. Democrats acting like everything is fine are deluding themselves. Polls right now are NOT predictive of the final outcome. But they do indicate where things are now, and where things are now is that Biden has lots of work to do.

There are several reasons for Biden and his partisans not to despair yet though:

  • The election is not today, there are nearly 350 days of developments yet to play out, including several Trump trials. These may or may not end up damaging Trump's poll numbers, but they represent a major source of uncertainty.
  • Historically incumbent candidates have been weakest about a year out before their elections, and have often "come back" and finished strongly, using the advantages of incumbency to benefit their campaigns.
  • The "heavy favorite" narrative I quoted above is in large part due to the fact that this site assumes the pattern of polls underestimating the Republican that held in 2 of the last 4 elections (2016 and 2020) holds again. The picture starts looking very different if the polls underestimate the Democrat like they did in 2012 (which would perhaps even indicate a Biden lead), or even if they are about right as they were in 2008 (where Trump still leads, but it seems much closer).
  • While Trump currently leads in all the critical battleground states in the EG averages, his leads are quite small, with the current tipping point being Pennsylvania, where Trump only leads by 1.2%, a margin which could literally disappear with a single good poll for Biden.
  • The percentage of respondents saying "neither", including both 3rd party support and people who insist they are undecided, is huge. This number swamps the margins between Biden and Trump in the swing states. Both 3P and undecided numbers tend to collapse as the election approaches, leaving a lot of room for persuasion.

Reminder that those who want to see a list of all the polls can just look here, and if you want updates on each and every poll as it comes out, as well as daily summaries of the status for the best polled matchup (currently Biden vs Trump), follow Election Graphs on Mastodon.

OK, now for actual numbers and graphs and such for those who want them.

This is how the spectrum of close states looked as of the last blog post on 2023-10-02:

And here is what it looks like as I write this on 2023-11-24:

There is clearly a lot more red than there was. Here is the movement since last time:

Places where the poll average moved toward Trump:

  • Arizona (11 EV): Trump by 0.3% -> Trump by 4.4% (Trump+4.1%)
  • Florida (30 EV): Trump by 3.3% -> Trump by 7.3% (Trump+4.0%)
  • Wisconsin (10 EV): Biden by 3.8% -> Trump by 0.1% (Trump+3.9%)
  • Michigan (15 EV): Biden by 2.0% -> Trump by 1.1% (Trump+3.1%)
  • Nevada (6 EV): Biden by 0.3% -> Trump by 2.8% (Trump+3.1%)
  • Georgia (16 EV): Trump by 2.7% -> Trump by 5.7% (Trump+3.0%)
  • North Carolina (16 EV): Trump by 1.3% -> Trump by 4.0% (Trump+2.7%)
  • Pennsylvania (19 EV): Biden by 0.7% -> Trump by 1.2% (Trump+1.9%)
  • Minnesota (10 EV): Biden by 6.9% -> Biden by 5.3% (Trump+1.6%)
  • Virginia (13 EV): Biden by 7.0% -> Biden by 6.0% (Trump+1.0%)
  • Texas (40 EV): Trump by 5.0% -> Trump by 5.8% (Trump+0.8%)
  • Iowa (6 EV): Trump by 8.6% -> Trump by 9.2% (Trump+0.6%)

Places with no change:

  • New Mexico (5 EV): Biden by 9.3%
  • New Hampshire (4 EV): Biden by 8.6%
  • Maine CD2 (1 EV): Biden by 1.6%
  • Nebraska CD2 (1 EV): Trump by 4.7%

Places where the poll average moved toward Biden:

  • Colorado (10 EV): Biden by 7.0% -> Biden by 7.8% (Biden+0.8%)
  • Ohio (17 EV): Trump by 10.6% -> Trump by 9.2% (Biden+1.4%)

The overwhelming trend is obviously toward Trump in this timeframe. The small movements in Colorado and Ohio in the other direction don't matter much in the face of all the movement elsewhere, especially since neither of these two states are particularly competitive.

No states flipped from the Trump side to the Biden side. And while there are lots of states that moved toward Trump without flipping sides, lets take a closer look at the four states (highlighted in red above) where the EG average actually flipped from having Biden ahead to having Trump ahead:

Starting with Pennsylvania as it is both the tipping point state and of these four the one with the most electoral votes, so a doubly critical state.

In the past in these posts, I've just shown the graph, but this time I'm including the status block that shows the averages for each candidate as well. Plus the "neither" number, which I added to the display a few weeks ago within the "<>" brackets.

As I mentioned in the TL;DR, this shows a big part of the story that otherwise is not visible. Yes, Trump is up by 1.2% in my average, but the average also shows full 14.4% refusing to support either. This just puts that 1.2% in context that is necessary to understand the potential volatility of that margin number.

Having said that, let's look at the trend. Biden's lead in Pennsylvania had peaked in June, and was slipping away a bit since, but the big jump came at the beginning of October. Although it looks like this jump coincides pretty closely with the start of the Israel-Hamas war, a few of the polls that started this move toward trump actually were in the field right before that. But additional polls later in the month confirmed the trend.

In the end this moves Pennsylvania from a small Biden lead to a small Trump lead. But given where all the other states lie, Pennsylvania is one of the critical states. For both candidates, most of the likely paths to victory include Pennsylvania.

The story looks similar in Michigan, except the upward trend seems less abrupt, looking like more of a continuation of a trend that started in June rather than something abrupt and new in October. The "neither" group is even bigger here than in Pennsylvania.

All of these seem similar to a degree. All have a movement toward Trump in October, all have a huge number in the "neither" category. In Wisconsin's case, we also have what is currently the closest state in our poll averages. Yes, there is a Trump lead in the poll average, but just barely.

Now, given the history of how the final poll averages did in 2008-2020, I translate a 0.1% Trump lead into a 63.3% chance of Trump winning the state if the election was today, simply because looking over all of those years, more often than not a nearly tied poll average resulted in a Republican win.

Finally Nevada. Nevada had been straddling the center line since April. The jump toward Trump in this case really didn't hit until the end of October rather than the beginning. That may just be related to when there was polling and when there wasn't though. We still aren't at the point in the cycle where all the close states are being polled every week or anything.

This move just puts Nevada back where it had been in the first part of the year though.

All four of these states are in the situation where a small number of polls favoring Biden could flip the average back to the other side quickly. So the situation is volatile. That will always be true in a situation where the tipping point state is "close". There is lots of opportunity for the situation to change.

But as of right now, the national situation looks like this:

All of this looks incredible for Trump, and horrible for Biden.

As of right now Trump is ahead in all the critical states, and in the probabilistic views, to look good for Biden, he would not only have to be leading in those states, but leading by a couple percent.

Looking at how the probabilistic ranges have evolved in terms of the Biden win odds:

  • 2023-03-25: Between 17.8% and 25.8%
  • 2023-04-24: Between 24.9% and 32.3%
  • 2023-05-21: Between 34.9% and 35.9%
  • 2023-06-25: Between 43.9% and 46.0%
  • 2023-08-13: Between 42.9% and 44.2%
  • 2023-10-01: Between 35.7% and 44.2%
  • 2023-11-26: Between 3.0% and 23.1%

This is the lowest Biden has been this cycle.

Looking at the two odds charts:

These vary by how the eventual errors between poll averages end up correlating between states.

That is, in the independent states version, the polls can be off from historical norms in one direction in one state, and in the other direction in the next state and there is no relation between what happens in one state and what happens in another. Whereas in the uniform swing model, if polls underestimate one side, they do so in the same way in every state. Of course the reality is somewhere in between.

The general shape is the same, but the Independent States version is more dynamic, so it is nicer to visualize trends. At the moment it is also worse for Biden and better for Trump, basically because the polls have to be underestimating the Democrats in multiple states to get to a Biden win, where with Uniform swing, you just have to have the nationwide error be missing Biden support.

Moving away from win odds to the Electoral College, here are the current views for each of the three ways EG has of showing the range of possible outcomes:

We are just showing all the different ways of looking at the same picture here. The median cases with Trump ahead by several states, but Biden wins are possible if he flips a few of the key states.

The other way we have of measuring how hard those flips would be besides the probabilistic models, is looking at the "tipping point", which is the margin in the state that would put the winner over the top.

Here is how that number has evolved:

  • 2023-03-25: Trump by 0.7% in Wisconsin
  • 2023-04-24: Trump by 0.5% in North Carolina
  • 2023-05-21: Trump by 0.2% in Wisconsin
  • 2023-06-25: Biden by 0.9% in Georgia
  • 2023-08-13: Biden by 0.7% in Michigan
  • 2023-10-02: Biden by 0.7% in Pennsylvania
  • 2023-11-26: Trump by 1.2% in Pennsylvania

Or in graph form:

One point that shows up more in this view than in the others, is that there has already been some movement back toward Biden after a peak around the beginning of November.

And it is a good time to point out again that 1.2% is not much. The tipping point went from Biden by 1.6% at the end of September to Trump by 1.9% at the end of October. These numbers can be volatile and move quickly. If the election was today, Biden would be an underdog, but there is time.

Having said that, the overall trends compared to 2016 and 2020 are dismal for Biden. For more than a year, Trump's polling has been consistently better than the previous cycles looking at both the median electoral college result and the tipping point:

Even with all the reasons I gave that Biden isn't out of the game even though numbers don't look great right now, this last observation should concern Biden folks the most.

Maybe the polls are all just wrong, and completely missing a large swath of Biden support. But if not… then Biden is not only running behind the pace of his narrow win in 2020, but is also lagging Clinton's loss in 2016.

Team Biden has some work to do. Running on a strategy of "well, we won last time" and/or "Trump's support will eventually collapse" is highly risky.

OK, quick look at the "What if it isn't Biden vs Trump" chart:

At this point there is finally just about enough polling to consider the Biden vs DeSantis numbers trust worthy as well as the Biden vs Trump numbers.

But you can safely ignore all of the others. There just isn't enough polling to say anything meaningful about those combinations. Don't trust those lines.

But we do see that with this latest surge, Trump once again does better against Biden than DeSantis does.

To explore Biden vs DeSantis more, look here.

Finally, as usual, closing with the current Biden vs Trump map:

345.1 days until polls start to close on Election 2024.

400 Days Left (Well, 401)

Welp, I seem to have gotten into a pattern of a blog update every 50 days, even though in my head I was aiming for every 28 days at this point. The last two posts were when there were about 500 and 450 days until the election, now there are about 400. Oops.

Posts will still accelerate as we get closer to the election of course, but the main place you should be looking for the current situation is not these blog posts anyway, it is the 2024 Electoral College page. That is updated regularly whenever new state level polls come out. At this point that is generally at least once a week, and often more frequently than that.

Because of that, I actually have now switched things around so if you just go to you will now end up there instead of the blog. You can always click through to the blog if you want, where I'll post occasional posts like this one summarizing how things have evolved lately. But you can see that from the graphs yourself too of course.

OK, before I go through the details of how things have changed over the last 50 days, here is the quick summary for those who don't want to read the rest:

  • If a Biden vs Trump election was held today, it would still be a dead heat and could easily go either way. A ton of critical states are "too close to call".
  • Biden has a narrow lead if you take the polls at face value, but given that on average in the last four election cycles polls have tended to underestimate the Republican, I still give Trump a slight edge on the probabilistic views.
  • From November until June the trend was in Biden's direction. But some polls I added in August, which covered late June dates, bounced things back in Trump's direction a bit, and things have been pretty steady since then.

OK, now for all the detail for folks who want it.

I'm not going to count and list all the individual polls that have been entered since the last update like I have before. Takes too much time. For those who just want to see a list of all the polls, look here.

If you want updates on each and every poll as it comes out, as well as daily summaries of the status for the best polled matchup (currently Biden vs Trump), follow Election Graphs on Mastodon.

OK, so to get started, we will look at how the spectrum of close states has changed.

This is how it looked when I did the last blog post on 2023-08-13:

And here is how it looks now on 2023-10-02:

The states where the poll average moved toward Trump are:

  • Pennsylvania (19 EV): Biden by 1.6% -> Biden by 0.7% (Trump+0.9%)
  • Arizona (11 EV): Biden by 0.3% -> Trump by 0.3% (Trump+0.6%)
  • Virginia (13 EV): Biden by 7.5% -> Biden by 7.0% (Trump+0.5%)
  • Wisconsin (10 EV): Biden by 4.0% -> Biden by 3.8% (Trump+0.2%)
  • Nevada (6 EV): Biden by 0.5% -> Biden by 0.3% (Trump+0.2%)
  • Iowa (6 EV): Trump by 8.5% -> Trump by 8.6% (Trump+0.1%)

And the states where the poll average moved toward Biden are:

  • Michigan (15 EV): Biden by 0.7% -> Biden by 2.0% (Biden+1.3%)
  • North Carolina (16 EV): Trump by 2.6% -> Trump by 1.3% (Biden+1.3%)
  • New Hampshire (4 EV): Biden by 6.0% -> Biden by 8.6% (Biden+2.6%)

Overall, this time more states in this center part of the spectrum are moving toward Trump than toward Biden.

Only Arizona crossed the center line this time though. Lets see what is happening there:

Last time Arizona was just barely on the Biden side. Now it is just barely on the Trump side. All but one of the 5 polls in the average show a margin under 5%.

The exception is a Prime poll that released 2 results. First was a 4% Trump lead in a straight Biden vs Trump matchup. Second was if you added a hypothetical unnamed No Labels candidate. That expanded Trump's lead to 10%. In the case of polls that release multiple results, I include them as their average within the larger average, so this shows a 7% Trump lead for aggregate of the two Prime results. Absent that, Arizona would still be slightly on the blue side. But I don't pick and choose polls to include. The average is the average.

Given we are right on the edge, it would be reasonable to expect Arizona to bounce back and forth around that centerline a bit as the next few Arizona polls come in.

The same is true with all the "close states" really, although obviously some are closer to zero margin than others. They should all be considered states that could go either way, which is why in our summary, we show ranges of possibility, not point estimates. So here is what that looks like now:

Note that if you look at ANY of the ranges above, each looking at the situation in different ways, every single one of them includes either candidate winning. This is the nature of a close race.

Both of my probabilistic views show Trump favored , while the straight up averages show Biden slightly favored. This is because the probabilistic views take into account that on average over the last four election cycles, the polls have underestimated the Republican in the closest states. But either way, they all show a race that could go either way.

Lets look at the evolution of the Biden win odds in these blog post updates:

  • 2023-03-25: Between 17.8% and 25.8%
  • 2023-04-24: Between 24.9% and 32.3%
  • 2023-05-21: Between 34.9% and 35.9%
  • 2023-06-25: Between 43.9% and 46.0%
  • 2023-08-13: Between 42.9% and 44.2%
  • 2023-10-01: Between 35.7% and 44.2%

The net result of all the state level changes definitely puts Biden in a weaker position than at the last update. Lets look at the "Independent States" chart first, since it is the most dynamic:

Where are we? Well, you can see that things were generally heading toward Biden from November 2022 until the end of June 2023, then things popped back up toward Trump a bit, and things have actually been pretty stable since then.

But wait, it popped back toward Trump at the end of June?

But the last update was in August I can hear you saying.

Yes. But remember that Prime poll in Arizona? The one with the No Labels alternative result? I had initially entered that data into the Election Graphs system with only the straight Biden vs Trump result when it came out in June. Sometime later I determined that I had made a mistake in not entering these secondary results on a whole bunch of polls, and went through a process over several weeks of fixing that.

Prime had released polls in Arizona, Georgia, Michigan, North Carolina, Nevada, Pennsylvania, and Wisconsin. All close states. These had generally been favorable to Trump anyway, and there already was a bit of a pop back in the Trump direction showing when I did the August update, but their results with the hypothetical No Labels candidate showed the 3P candidate helping Trump in every one of those states except North Carolina (where there was no impact).

On average in the Prime polls in those 7 states, the presence of a No Labels candidate helped Trump (and hurt Biden) by a net change of 4.3% on the margin.

So when I folded in the Prime data including No Labels in August, since these results were all good for Trump in a whole bunch of close states, it moved all the curves in the Trump direction from June through the present.

And we have been pretty stable since then. When these polls roll off the averages as new polls come in, if the new polls don't show the same impact, things will move back in Biden's direction a bit.

Of course, most other polls do not include a No Labels contender. If we actually end up with a No Labels candidate, these polls that included them may end up being the ones that are closer to right.

Once various possible 3P candidates become "real" rather than just hypothetical more polling outfits will include them. 3P candidates also usually fade as the election approaches though.

We shall see.

In terms of the actual Electoral College margin predicted by the Independent States view, the chart looks like this:

The tie line is well within the darkest band, which is the 1σ (68.27%) band. Every view shows a close race, so we should not be surprised.

The two Uniform Swing views look similar, but with a wider range of outcomes and a "blockier" appearance:

Just looking at straight poll averages, and considering any margin less than 5% too close to call, the midpoint has moved around a bit, but the overall range of possibilities hasn't changed all that much.

A better way to understand this "straight poll averages" view doesn't look at the electoral college margin, but instead how much the margin would have to change in the tipping point state to change the winner. Here is some history on the tipping point margin:

  • 2023-03-25: Trump by 0.7% in Wisconsin
  • 2023-04-24: Trump by 0.5% in North Carolina
  • 2023-05-21: Trump by 0.2% in Wisconsin
  • 2023-06-25: Biden by 0.9% in Georgia
  • 2023-08-13: Biden by 0.7% in Michigan
  • 2023-10-02: Biden by 0.7% in Pennsylvania

And here is a graph of how this has evolved:

The picture of course matches the others. At the moment Biden leads in the state that would put him over the edge (Pennsylvania at the moment) by only 0.7%.

Leads that small are ephemeral, and possibly illusory.

The probabilistic views here are all based on looking at the final election graphs averages in all the states compared to the actual election results in 2008 through 2020. Basically, in cases where Democrats were ahead by about 0.7%, they won only about 44% of the time.

So with Biden leading by 0.7% in the tipping point state, this is why the two probabilistic views give Trump an edge, even though the straight up poll averages give Biden a slight advantage.

Either way, the correct summary is "too close to call".

As usual, lets compare to 2016 and 2020 as well:

There are some arguments to be had related to Democratic over performance in midterms and special elections that while the polls may have underestimated the Republicans in 2016 and 2020, they may be underestimating the Democrats in 2024 like they did in 2012.

Maybe. There will be no way to tell until the votes are counted.

But at the moment, Biden continues to not only be weaker than he was against Trump in 2020, but in fact significantly weaker than Clinton was at the same point in 2016. There is a long time to go yet, but the Biden camp should not be happy with this position. Trump on the other hand should be heartened. Even with all of his mounting legal issues, he looks to be in a relatively strong position. For now anyway.

What if it isn't Biden vs Trump? OK, a quick look…

The above compares the median electoral college margin in the Independent States view between the 6 best polled candidate combinations.

Of these six though, Biden vs Trump is STILL the only one where I would say we have good reliable data.

Biden vs DeSantis is getting closer, but still needs to be taken with a grain of salt. There are still lots of "close states" where there haven't even been five polls yet.

And the other four combos on this chart? Just ignore them completely. There are only a smattering of polls on those combinations, not anywhere near enough to even remotely believe anything we see yet.

With those caveats, DeSantis does look stronger against Biden than Trump does on this chart at least.

Odds wise the Independent States view gives Biden only a 19.2% shot against DeSantis compared to 35.6% against Trump.

Looking at the Uniform Swing view however gives Biden a full 57.3% against DeSantis, better than his 44.2% against Trump.

So there is still a lot more uncertainty in the DeSantis picture.

Right now primary polling indicates both Biden and Trump are cruising to their respective nominations. If this ever changes, we'll pay more attention to the other possible combinations and dig a bit deeper.

For full Biden vs DeSantis info, look here.

As I always do, I'm going to end with the current national Biden vs Trump map:

400.8 days until polls start to close on Election Night 2024. Hang tight.

Leveled Off?

I had hoped to do another update four weeks after the last one, but here it is about 50 days after my last post instead. Life gets in the way. Oh well. But it does mean that this update is right on the 450 days away from polls closing mark. So there is that.

The quick summary for those who don't want to read all the details:

  • Once again, if a Biden vs Trump election was held today (it won't be), it would be fundamentally too close to call.
  • Biden's odds are slightly less than they were last time, but in between that update and this one, he was better off, so things have at least leveled off, and may have started moving in Trump's direction again.
  • Biden has a lead if you take the state poll averages at face value, but this relies on several states where his lead is less than the average amount polls have underestimated Republicans in close states from 2008 to 2020, so be wary of trusting those leads.
  • Because of the above, both of my probabilistic views give a slight advantage to Trump despite the face value result.
  • Biden's polling is weaker than both his own polling in 2020, and Clinton's polling in 2016 at the same amount of time before the election.

If you want updates on each and every poll as it comes out, as well as daily summaries of the status for the best polled matchup (currently Biden vs Trump), follow Election Graphs on Mastodon.

OK, now all the details that only the die hard will be interested:

New polling added since 2023-06-25:

  • 24 Biden vs Trump: MIx4, AZx3, NVx2, PAx2, WIx2, NHx2, OHx2, NY, TN, FL, GA, NC, CA, VA, KY
  • 11 Biden vs DeSantis: MIx2, OHx2, NH, WI, TN, NV, FL, AZ, VA
  • 2 Biden vs Scott: NH, OH
  • 1 Biden vs Pence: OH
  • 1 Biden vs Haley: OH
  • 1 Biden vs Christie: OH
  • 1 Biden vs Youngkin: VA
  • 1 Kennedy vs Trump: NH
  • 1 Kennedy vs Scott: NH

Also, after going back and forth a bit on how much I wanted to do it this cycle, I decided that like previous cycles, when a pollster releases multiple versions of their results, like one for Registered Voters and one for Likely Voters, or they do "Definite" supporters vs "Total" supporters or what not. When this happens, I include both versions of the poll, but I weight them so the "sub-polls" together have the same weight as a single poll.

So I've been slowly going back and reviewing older polls I have in my system, and checking to see if they have these sorts of multiple results, and if so I've been adding the "half-polls" I missed originally. I'm slowly catching up with the current day as I do this. All of the below were old polls that I have updated.

  • 31 Biden vs Trump: IAx4, AZx3, GAx3, PAx3, FLx2, MIx2, WI, CO, WA, AR, KS, LA, MS, MT, OH, OK, RI, SC, TX, NC
  • 20 Biden vs DeSantis: AZx2, GAx2, PAx2, AR, FL, KS, LA, MS, MT, OH, OK,  RI, SC, TX, NC, IA, MI
  • 1 Biden vs Pence: NC

As has been the norm the last few updates, I start with looking at the changes to the "spectrum of close states":

As it was on 2023-06-25:

And now on 2023-08-13:

The states where the poll average moved toward Trump are:

  • Michigan (15 EV): Biden by 2.3% -> Biden by 0.7% (Trump+1.6%)
  • North Carolina (16 EV): Trump by 1.5% -> Trump by 2.6% (Trump+1.1%)
  • Georgia (16 EV): Biden by 0.9% -> Trump by 0.2% (Trump+1.1%)
  • Pennsylvania (19 EV): Biden by 2.6% -> Biden by 1.6% (Trump+1.0%)
  • Virginia (13 EV): Biden by 8.0% -> Biden by 7.5% (Trump+0.5%)
  • Florida (30 EV): Trump by 3.1% -> Trump by 3.3% (Trump+0.2%)

And the states where the poll average moved toward Biden are:

  • Colorado (10 EV): Biden by 6.8% -> Biden by 7.0% (Biden+0.2%)
  • Texas (40 EV): Trump by 5.2% -> Trump by 5.0% (Biden+0.2%)
  • Iowa (6 EV): Trump by 9.2% -> Trump by 8.5% (Biden+0.7%)
  • Arizona (11 EV): Trump by 0.6% -> Biden by 0.3% (Biden+0.9%)
  • Nevada (6 EV): Trump by 0.5% -> Biden by 0.5% (Biden+1.0%)
  • New Hampshire (4 EV): Biden by 3.4% -> Biden by 6.0% (Biden+2.6%)
  • Wisconsin (10 EV): Trump by 0.2% -> Biden by 4.0% (Biden+4.2%)

The last few (roughly) monthly updates have been very lopsided, with a lot more states moving toward Biden than moving toward Trump from this group of close states. (We really don't care about how states move if one candidate has an overwhelming lead.) This time though, 7 states moved toward Biden, 6 states moved toward Trump. That is much more even, indicating we've moved into a more stable period, at least for a little while.

Georgia, Arizona, and Nevada crossed the center line.

Let's look at them in more detail.

This time I won't write a paragraph of commentary on each state, you can look above and draw your own conclusions.

At the moment these are the three closest states according to the straight up averages though, all with margins 0.5% or less. You can expect the closest states to bounce around from one side of the center line a bit.

OK, lets look at the overall summary:

Looking at the evolution of the Biden win odds first:

  • 2023-03-25: Between 17.8% and 25.8%
  • 2023-04-24: Between 24.9% and 32.3%
  • 2023-05-21: Between 34.9% and 35.9%
  • 2023-06-25: Between 43.9% and 46.0%
  • 2023-08-13: Between 42.9% and 44.2%

The trend of Biden being stronger each time than the time before has been broken, with Biden slightly weaker this time than last. Lets look at the two odds charts to see if we can get any additional insights.

In both ways of calculating odds (the two extremes of how correlated state polling errors might be) you can see that for a time the odds actually favored Biden for a bit, but then things popped back in Trump's direction. The net result is that we are pretty close to where we were in the update 50 days ago, but the trend has reversed.

Now is this move toward Trump something that will be sustained? Or is it just a temporary blip before the trend toward Biden continues? Of course there is no way to know other than to wait and see what happens next.

So how does this look in terms of the electoral college? Here are the three views:

The first two are the two probabilistic views based on how correlated the state polling errors may be and taking into account how far off polls have been from the final election results in the past 4 election cycles. The third is just taking the state poll averages at face value.

But for all three the key is to not fixate on the green line in the center, but instead look at the green bands. For the probabilistic views there are graduations that don't exist in the categorization view, but in all three, there is a fair amount of green on both sides of the center line.

This means that either candidate winning is a reasonable outcome to consider, thus the summary that this race would be "too close to call" if the election was today.

It remains close. And in the last 50 days, movements have been mixed, so there isn't a clear trend either. Nobody is pulling away here.

Another way to look at the straight unmodified polls is to look at the "tipping point", which is how much you would have to shift all the polls to change the winner. It has been evolving like this:

  • 2023-03-25: Trump by 0.7% in Wisconsin
  • 2023-04-24: Trump by 0.5% in North Carolina
  • 2023-05-21: Trump by 0.2% in Wisconsin
  • 2023-06-25: Biden by 0.9% in Georgia
  • 2023-08-13: Biden by 0.7% in Michigan

Just another way of showing a super close race.

Time to compare to 2020 and 2016:

Biden still lags where Clinton was at this stage in the 2016 cycle, and is far behind where he was himself in the 2020 cycle. Given the Trump win in 2016, for Biden to be feeling good, I'd think he would AT LEAST want to be doing better in the polling than Clinton was in 2016. He is not there right now.

I have heard many folks state things like "Biden vs Trump is a rematch, Biden won last time, so he will win again" or that "surely with all his legal issues, there is no way Trump can win".

It is unwise to make these sorts of assumptions.

Well, what if Trump's legal issues (or something else) at some point cause the Republicans to give the nomination to someone else? (Yes, I know he is way ahead on that front, but what if?)

The polling for Biden vs Trump is still the only combination where there have been enough polls in the close states that I feel confident making judgements on it. Most combinations just have scattered polling, so looking at the state by state info like we do here is almost meaningless at this point.

But Biden vs DeSantis is getting closer and closer to having good polling in the close states, so lets take a look anyway.

Here is a comparison of the medians for the Independent States probabilistic view:

He may be woefully behind in the Republican Primary polling, but for the general? He does better against Biden than Trump. At least in this view. (Again, still with much more limited polling, so the picture may be incomplete.)

Looking at probabilities from both probabilistic views, the picture is a bit more muddy. While we say Biden has between a 42.9% and 44.2% chance of winning, the spread is much bigger against DeSantis. The odds are anywhere from 15.9% to 57.3%. So basically, we don't really have a good handle on this particular match up.

The reason for such a wide spread between the probabilistic views is because of the specific configuration of close states and their averages for Biden vs DeSantis. If we get to the point where I feel the Biden vs DeSantis polling is more solid, AND it looks like the chances of that matchup are increasing, AND this situation persists by then, I will of course dive into that in more detail. But not today.

For full Biden vs DeSantis info, look here.

To close up, the national map for Biden vs Trump:

450.1 days until polls start to close on Election Day 2024.

500 Days Left (Well, 499)

It has been five weeks since my last post instead of four, but I noticed we were coming up on a milestone, so I waited an extra week.

As I start writing this post, it has been 75 minutes since we passed the 500 day mark until the first states close their polls for Election 2024, and we start seeing real results. It will be a bit less by the time I finish and hit publish, but still, it seemed worth noting.

For those not wanting to read all the details below, here is the TLDR:

  • If a Biden vs Trump election was held today, it would be a dead heat and could easily go either way. A ton of critical states are "too close to call".
  • Biden has a narrow lead if you take the polls at face value, but given that on average in the last four election cycles polls have tended to underestimate the Republican, I still give Trump a slight edge on the probabilistic view.
  • Since November the trend has been in Biden's direction, and this trend has continued in the last month.

OK, now some details:

Since the last post there have been 29 new data points added.

  • 16 for Biden vs Trump [NVx2, CAx2, GAx2, KY, TX, IA, VA, NC, WA, MD, MI, AZ, PA]
  • 11 for Biden vs DeSantis [TX, IA, NV, NC, WA, CA, MD, MI, AZ, GA, PA]
  • 1 for Harris vs Trump [TX]
  • 1 for Harris vs DeSantis [TX]

If you want updates on each and every poll as it comes out, as well as daily summaries of the status for the best polled matchup (currently Biden vs Trump), follow Election Graphs on Mastodon.

Let's start again with the "spectrum of close states" for Biden vs Trump.

As it was on the last post on 2023-05-21:

And now on 2023-06-25:

The states where the poll average moved toward Trump are:

  • North Carolina: Trump by 0.5% -> Trump by 1.5% (Trump+1.0%)
  • Nevada: Biden by 0.4% -> Trump by 0.5% (Trump +0.9%)

And the states where the poll average moved toward Biden are:

  • Michigan: Biden by 2.2% -> Biden by 2.3% (Biden+0.1%)
  • Pennsylvania: Biden by 1.8% -> Biden by 2.6% (Biden+0.8%)
  • Texas: Trump by 6.0% -> Trump by 5.2% (Biden+0.8%)
  • Iowa: Trump by 10.0% -> Trump by 9.2% (Biden+0.8%)
  • Arizona: Trump by 1.8% -> Trump by 0.6% (Biden+1.2%)
  • Georgia: Trump by 0.7% -> Biden by 0.9% (Biden+1.6%)
  • Virginia: Biden by 5.6% -> Biden by 8.0% (Biden+2.4%)

Of these there were two that changed categories, so lets take a closer look at those.

In the last update I said this about Nevada:

You can see that what happened here is that what looks like an outlier leaning toward Trump was replaced in the "last five polls" by an outlier that leans toward Biden, thus moving the average significantly toward Biden. But this leaves us JUST BARELY on the Biden side of the fence. The next poll in Nevada could easily flip the state back to the red side. Or it could show that this was no outlier, but a harbinger of a new trend. Too soon to tell.

Basically, the next poll in Nevada flipped the state back to the red side, just as I said might happen. It was another poll that looks like an outlier, this time on the Trump side. The Election Graphs poll averages don't exclude polls, but just glancing at the chart above and using human judgement, I'd probably throw out any of the polls showing either candidate with a lead over 5%. All the other polls show a close race, and all but one of those show a small Trump lead. So if I had to guess the "true" state of Nevada, I'd say there is a narrow Trump lead.

The one caveat would be that if you look at only those non-outlier polls, the rest seem to be trending slowly toward Biden over time. If you mentally draw that trend line, Nevada is actually pretty close to a tie right now. So maybe the real situation is a tie race in Nevada right now, but trending toward Biden.

Georgia doesn't seem to have the same outlier issue as Nevada, but there does seem to have been a shift around the midterms in November 2022.

Before that point, there was only a single poll showing a Biden lead, and most polls didn't just show a Trump lead, they showed a Trump lead of more than 5%.

After the midterms, there has only been a single poll showing a Trump lead, with all of the others showing narrow Biden leads.

So what we are seeing here is the Election Graphs average finally having enough post-midterms polls for the trend line to reflect this change and show a narrow Biden lead, which seems about right for Georgia right now.

The actual poll average right now is an 0.9% Biden lead in Georgia. But based on looking at how Election Graphs poll averages correlated to actual election results from 2008 to 2020, a Democratic lead of 0.9% only translates into a 46.1% chance of the Democrat winning, at least based on that historical performance.

Simply put, more times than not, the polling average has underestimated the Republican, which means the Democrat needs to be leading the poll average by more than 1.3% for me to list them as having a better than 50/50 shot of winning. If anybody wants details of how I calculate that, look at this post from January.

As everyone knows, "past performance is not a guarantee of future results", but looking at past performance is the best I can do.

Bottom line though, Georgia is looking super close at the moment, and like Nevada, could easily go either way if the election was held today. (Which, obviously, it won't be, and things will change a thousand times before we get there.)

OK, so how does all this change the national picture? Here is the current overall summary:

Looking at the evolution of the Biden win odds first:

  • 2023-03-25: Between 17.8% and 25.8%
  • 2023-04-24: Between 24.9% and 32.3%
  • 2023-05-21: Between 34.9% and 35.9%
  • 2023-06-25: Between 43.9% and 46.0%

Each month Biden has been stronger than the month before.

Visually, the two odds charts:

Trump still has a slight advantage, but only due to what I mentioned earlier about the Election Graphs polling average tending to underestimate the Republicans in the last few cycles.

Taking that factor into account here is what the trend in the range of possibilities looks like assuming the states are completely independent:

And assuming any polling error will be the same in every state:

Both of these views show Trump with a slight advantage, but with a huge range of possible outcomes very possible. The darker the green, the more likely the outcome is in that range, with the three shades being 1σ (68.27%), 2σ (95.45%), and 3σ (99.73%) probability bands.

In both models above, even the 1σ bands include either candidate winning, so this remains a situation where either candidate could win easily.

Removing the "well, polls have underestimated Republicans in the past" factor and just looking at who is ahead in each state and what the close states are, you get this:

Looking at snapshots of this on the days I've done blog posts:

Categorization Trump Best / Expected / Biden Best:

  • 2023-03-25: Trump+96 / Trump+18 / Biden+162
  • 2023-04-24: Trump+96 / Trump+18 / Biden+162
  • 2023-05-21: Trump+96 / Trump+6 / Biden+162
  • 2023-06-25: Trump+96 / Biden+14 / Biden+162

It is easy to just fixate on that middle number moving toward Biden. Unlike the two probabilistic models, Biden is indeed in the lead for the first time since September.

But once again the full picture here is that there is a wide range of reasonably possible options, and we essentially have a tie race right now.

A more subtle way to look at the straight unmodified polls is to look at the "tipping point", which is how much you would have to shift all the polls to change the winner. It has been evolving like this:

  • 2023-03-25: Trump by 0.7% in Wisconsin
  • 2023-04-24: Trump by 0.5% in North Carolina
  • 2023-05-21: Trump by 0.2% in Wisconsin
  • 2023-06-25: Biden by 0.9% in Georgia

Once again, this is just another way of showing a very close race, but one that is moving slowly in Biden's direction.

But how does this compare to the last couple of cycles?

So…  yes, things have been moving away from Trump and toward Biden. But Trump is polling better now than he was either 4 or 8 years ago. I've said this before, and I'll say it many more times I'm sure: Nobody should underestimate Trump in this situation.

But what about people other than Trump?

The only other candidate combination with more than scattered polls here and there is Biden vs DeSantis. But the polling at the state level for this combination is still woefully inadequate.

For Biden vs Trump there are enough polls for a full poll average that doesn't rely on old election results as a baseline in every single one of the close states (not counting ME and NE CDs).

For Biden vs DeSantis, the only close states with that much polling are Georgia, Nevada, and Arizona. (DeSantis also has that level of polling in Florida, but while Florida is a close state for Trump, it is not for DeSantis.)

Bottom line, continue to treat Biden vs DeSantis with a big grain of salt until there is a lot more polling. And completely ignore all the OTHER possible combinations. The polling for those is still just completely useless.

Having said that, here is a look at the median case for the independent states model that compares the most polled combinations:

Like I said, ignore everything besides Biden vs Trump and Biden vs DeSantis, and take Biden vs DeSantis with a grain of salt, but…

The recent trend does show that while Trump has been falling vs Biden, DeSantis has been rising. Based on this particular view of the current state level polls, DeSantis would do much better against Biden than Trump would.

Again though, until there are a good number of Biden vs DeSantis polls in all of the close states, I would not completely trust that Biden vs DeSantis picture.

I'm also showing the single chart that shows the best picture for DeSantis here. While DeSantis is improving on the other metrics too, it is nowhere near as dramatic as this one. For full Biden vs DeSantis info, look here.

That is enough for now. As usual, I'll close with the update national map for Biden vs Trump:

I mentioned at the top that there would be less than 500 days left by the time I got the post published. I was right. It is a whole day later. 🙂

499.0 days until polls start to close on Election Day 2024.