Some Dems Up, Some Dems Down

It has only been eight days since the last update. Still, there have been new polls in Nevada, Michigan, Pennsylvania (x2), Wisconsin (x2), Virginia, New York, Missouri (x2), Maryland, Montana, New Hampshire, Arkansas, California (x2), Texas (x2), and South Carolina.

So we might as well get in another update before South Carolina primary results start coming in, and Super Tuesday results three days later.

There are lots of reasons not to pay attention to current head-to-head polls against Trump when making decisions about primary choices. Most pointedly, things change and change quickly, so where things are at the end of February do not necessarily correspond to where they will be at the beginning of November. And of course, things like policy and character should also play a role.

But for those for whom "How might the general election go?" is an important decision making factor, here is the latest from Election Graphs, based on state poll averages.

Let's start with the "odds of winning the electoral college" based on the state level head-to-head poll averages, and a Monte Carlo model using the historical accuracy of the final Election Graphs poll averages to determine how far off the polls tend to be. Keep in mind this is "if the election was today." Which it is not.

Dem 21 Feb 29 Feb 𝚫
Biden 98.4% 97.7% -0.7%
Sanders 77.9% 82.0% +4.1%
Bloomberg 59.9% 54.1% -5.8%
Warren 38.3% 40.9% +2.6%
Buttigieg 9.3% 15.9% +6.6%

The last week of polling has improved Sanders, Warren, and Buttigieg's prospects against Trump.

Meanwhile, Biden and Bloomberg have both slipped.

The order has not changed, though.

  • Biden still is the strongest against Trump by far.
  • Sanders is more of a gamble but still significantly favored.
  • Bloomberg is a little better than a coin toss, but not much.
  • Warren is a bit worse than a coin toss but still has a decent chance.
  • Buttigieg would be a long shot. About the same as Trump in 2016.

Now would be a good time to talk a little about a Twitter thread by Johnathan Mummolo, a political scientist at Princeton. The thread summarizes a paper by Westwood, Messing, and Lelkes titled "Projecting Confidence: How the probabilistic horserace confused and demobilizes the public."

The bottom line is that the vast majority of people do not understand probabilities.

I have repeatedly ranted both here and on my Curmudgeon's Corner podcast ever since the 2016 election about people looking at a 14% chance of Trump winning (the median odds from all the sites I could find that gave odds) and acting as if it was 0%. That 14% is approximately the same as rolling a one on a six-sided die. And while people might be disappointed in that result if they wanted a six, nobody would be surprised by getting a one. Ones happen all the time.

This paper gets at a different but related problem. When looking at a probability of a candidate winning, vs. an equivalent percentage margin in the polls, people looking at the margins will interpret the situation as being a closer race and be more likely to vote than the people looking at probabilities even though the underlying truth is precisely the same.

Here at Election Graphs, we used the historical performance of the final, right before election day, Election Graph poll averages for every state vs. actual election results in 2008, 2012, and 2016 to estimate given a particular margin, how often would each party win? The detailed methodology is in this post from January 2019.

This analysis gives us numbers like if a Democrat is leading a state by 3.0% entering election day, they have a 73.8% chance of winning the state.

But it seems if people see 73.8%, they think it is a sure thing, so why should they bother voting? Whereas if they see a 3.0% lead, they believe it is a close race, and maybe they should vote.

Of course, 73.8% is not a sure thing at all! There is more than a one in four chance things will go the other way!

But human psychology and probabilistic innumeracy win the day!

The thread and paper also mention that at the moment, Democrats are more likely to frequent sites (like this one!) that give probabilistic forecasts. So presenting this sort of information ends up serving as a form of voter suppression for Democrats (if the Democrat is in the lead anyway).

I'll also note that when looking at a national election based on the Electoral College instead, people are going to be confused too. A significant Electoral College margin can rely on a small number of states being just barely on one side of the line or the other. A lead there can disappear in a flash with a slight movement in those states.

But of course, looking at the popular vote isn't a solution either, since as both 2000 and 2016 illustrated nicely, we don't pick presidents by the popular vote.

Here at Election Graphs, we are going to continue to present the probabilistic views anyway, of course. But if you are paying attention to them, you do need to understand what they mean, as well as pay attention to the various caveats about how quickly things change that I repeat endlessly. It is important.

But to get a full view of what is going on, we also present the national picture in three other ways regularly in the blog, and there are even more available on the blog. We let you dig into what is happening in all of the states and see all the individual polls too if you want to get granular.

This stuff is complicated. Dig in. Understand the details.

Anyway, we now turn to the median of the Electoral College margin simulations. Roughly speaking half the time, the Democrat will do better than this, and half the time, they will do worse than this. Maybe that is a little less confusing than the probability of winning?

Dem 21 Feb 29 Feb 𝚫
Biden +104 +100 -4
Sanders +42 +50 +8
Bloomberg +14 +8 -6
Warren -14 -10 +4
Buttigieg -54 -40 +14

The changes here parallel the odds, of course. But does presenting it at a margin make you FEEL differently about the results? Maybe.

Also, of course, the median margin in the model does not alone tell you how about the distributions, and how easily it would be for things to change. That is what the probability helps to understand. Two candidates might show the same median margin, but be in very different situations depending on the margins in the individual states.

Simplifying this even further to look at the margins if each candidate wins exactly the states they lead, you get this:

Dem 21 Feb 29 Feb 𝚫
Biden +198 +178 -20
Sanders +66 +66 Flat
Bloomberg +52 +58 +6
Warren -20 -8 +12
Buttigieg -84 -40 +44

Interestingly, in this view, Bloomberg improves, even though his odds of winning and his median margin got worse. This divergence is because Bloomberg improved his margin by 6 EV by taking the lead in Virginia (13 EV) and losing the edge in Wisconsin (10 EV). But meanwhile, he weakened in other states enough to lower his overall chance of winning, even though the straight-up list of places he is ahead improved.

Finally, the tipping point, representing how much of a national shift in polls would change the outcome:

Dem 21 Feb 29 Feb 𝚫
Biden +2.8% +2.6% -0.2%
Sanders +0.5% +0.8% +0.3%
Bloomberg +0.7% +0.7% Flat
Warren -0.5% -0.5% Flat
Buttigieg -2.9% -2.1% +0.8%

If I could only keep two of these charts, it would be the probability of winning and the tipping point. The likelihood of winning tells you what might happen if the election was today. But the tipping point tells you how easy it is for those probabilities to change.

There is not a single one of these five candidate combinations that are further than 3% from the centerline. That means that if there is a systematic bias of 3% in the polls toward the other side, the outcome will change. Similarly, any news event that can move the margins by 3% can change the results.

To kill a 3% margin, only 1.5% of the public needs to change their minds. People deciding to stay home and not vote can also eliminate a 3% margin in an instant.

So yes, the odds here show that if a Biden vs. Trump election were today, Biden would have a 97.7% chance of winning. But a tipping point of only 2.6% tells you that Biden's entire advantage could disappear virtually overnight with the right bit of negative news hitting the headlines, or with a pretty slight polling error in the critical states.

Which brings us back to the importance of correctly interpreting the numbers we share here on Election Graphs.

There is a big difference between "would probably win if the election was today" and "will probably win in November."

And even if the election was tomorrow, 97.7% is not the same as 100%. And 82.0% is certainly not the same as 100%.

No matter which candidate pair you look at, this is still a close and highly contested election.

The results of the South Carolina primary start coming in just a few hours.

248.2 days until polls start to close on the general election.

We have a long way to go.

For more information:

This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a poll-based 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 Post-Nevada Sanders Lead

It took a couple of days, but as of Monday afternoon, the Nevada results were final. It was clear from the moment returns started coming in that Sanders was going to win handily, but the extent of that lead in terms of the delegate estimate moved around a bit as the returns slowly came in.

During this time, the estimated number of Sanders delegates in Nevada ranged from 22 to 28, Biden ranged from 7 to 11, Buttigieg ranged from 0 to 6, and Warren ranged from 0 to 1. But once we knew the final count, these were the results from Nevada:

  • Sanders 24
  • Biden 9
  • Buttigieg 3

Sanders needed to get 19 or more of the Nevada delegates to improve his overall position in terms of the % of remaining delegates needed to win. He did that handily. The updated chart of that metric looks like this:

Remember, for this chart down is positive and up is negative. When you get down to 0%, you clinch the nomination. If a candidate goes up to 100%, they become mathematically eliminated.

Sanders has made a turn downward. He is not only in the lead, but Nevada put him on a winning pace.

Now, the general talk is about how absent a significant change, Sanders may be in an uncatchable position after Super Tuesday. That isn't based simply on today's delegate totals, but also on his polling in South Carolina, the Super Tuesday states, and nationwide, and a bit of knowledge of how a small popular vote lead translates into a massive delegate lead.

For instance, in Nevada, Sanders got about 33% of the popular vote, but that translated into 67% of the delegates. This kind of magnification for the winner is intentional in the delegate allocation rules. The 15% delegate threshold in one cause. The fact that the results in individual congressional districts determine many of the delegates is another.

Rather than look at the prognostications of how future states might go, instead, let's look at how you would expect the % of remaining delegates needed to win chart to change as this progresses. That will help us know how we will identify if it seems like Sanders is on track to a clear win, if we are heading towards a contested convention, or if someone else still has a chance to win.

To illustrate, we'll look at some graphs from previous cycles to compare to where things are now.

Let's look at the contested races in both parties since 2008 when we did the first Election Graphs delegate tracking. We'll look at them in order of how quickly the nominee was pretty clear in each contest.

Which means we will start with the Democrats in 2016:

This chart shows what it looks like when we have a runaway victory that is clear from the beginning.

Sorry, Sanders folks. 2020 is going differently, but in 2016, because of the courting of superdelegates long before Iowa even happened, Clinton built up a delegate lead starting from the very beginning.

From the 0% starting line, Clinton improved her position with every contest, and Sanders's situation got worse. The only exception was a slight bump around the 58% mark when Sanders had one outstanding day. But the overall trend was clear from the very beginning. Clinton was on the road to an outright victory, and Sanders never managed to slow that progress.

Next up, the Republicans in 2008:

It took slightly longer for this one to become apparent. Romney took an early lead, but his line stayed flat, hovering around the 50% line. McCain was heading upward along with the also-rans. But at about the 5% mark, McCain started hitting his mark and improving his position with every contest. He pulled ahead of Romney at about the 10% mark, then when Super Tuesday jumped the race over 40% every other candidate was at the point where they needed 60%+ of the remaining delegates to catch up and win. That was, of course, unrealistic. Except for one short jog just past the 50% mark, McCain kept improving his position in every race.

Unlike Clinton in 2016, McCain had some issues before Super Tuesday but hit his pace quickly, and Super Tuesday made things inevitable.

Next up, Republicans in 2012:

This chart shows an example of a slower burn. Romney was in the lead from the very beginning, but his "% of remaining delegates needed to win" basically stayed flat right around the 50% mark for a long time.

This pattern means that he was accumulating delegates much faster than anybody else, and it was clear the other candidates were not going to win outright. But the other candidates were continuing to take enough delegates to keep a contested convention an active possibility for awhile.

That changed around the 43% mark though. Winner-take-all states on the Republican side undoubtedly helped with this. But also once it is clear that candidates can't win, it becomes tough for them to actively continue a campaign based on the idea of forcing a contested convention where maybe they will be picked, but probably not. So one by one, the other candidates drop out, and then the candidate in the lead starts rolling up the remaining delegates.

So this race had a clear leader way ahead of the rest virtually from the 0% mark, but Romney didn't start hitting a winning pace consistently until 43%.

Now Democrats in 2008:

This graph shows a real two-person race. Both Obama and Clinton maintained flat lines for a long long time. Clinton was even improving a little. But not very much. Obama was getting enough delegates to keep her from hitting the marks she needed to improve the "% of remaining delegates needed to win" number significantly.

Clinton still had the advantage for more than the first half of the campaign delegate wise. Around the 50% mark, though, Obama started consistently hitting the percentages needed to improve his position while Clinton fell further and further behind. For a long time, though, the situation was dynamic. Clinton didn't get mathematically eliminated until around the 96% mark!

Finally, the Republicans in 2016:

This graph shows the closest we have gotten to a contested convention since I started tracking delegates in 2008. The other Republican candidates kept Trump over 50% of delegates needed until more than 70% of the delegates were allocated; around the same time, Cruz became mathematically eliminated. Although there were a couple of ups and downs along the way, Trump didn't start consistently improving his numbers until about the 67% mark. Until then, while Trump was way ahead, the possibility of a contested convention was kept open. But just like 2012, once it was clear that other candidates did not have a realistic path, and they started to drop away, the leader was able to take all or almost all of the remaining delegates, and wrap things up.

This year after South Carolina, we will be at 3.9% of the delegates allocated.

After Super Tuesday, we will be at 38.0%.

We'll hit 50% on March 17th  after Arizona, Florida, Illinois, and Ohio.

We'll hit 67% on April 4th  after Alaska, Hawaii, Louisiana, and Wyoming.

As we pass each of these milestones, the question is, does 2020 look like one of these past patterns? Or something else entirely?

If Sanders (or any other candidate) is under 50% of remaining delegates needed, and the number goes down after each contest, then we are on track for that candidate running away to an outright win with no significant obstacles to that result. (This is like the Democrats in 2016, or the Republicans in 2008.)

If all of the candidates except one are racing up to 100%, but the leader is kept flat around 50%, it means that while one candidate has a chance of winning outright, the other candidates are combining at a level that keeps the possibility of a contested convention open. The tendency in this situation is that once all the opposing candidates are mathematically eliminated, they will drop out, and the leader will be able to at that point hit the marks they need to get to a majority. (This is like the Republicans in 2012 or 2016.)

If two candidates are managing to keep their lines relatively flat, you have a two-person race, with both really still in contention. Until the point where one person's line goes up consistently, and the other person's line goes down, you have a real race. Unless the two candidates are closer together than the sum of the other candidates' delegate totals though, one of the two will end up winning. The only question is how close to the end you get before the winner becomes clear. (This is like the Democrats in 2008.)

So what pattern would we see if we are actually on a path to a contested convention?

If after each contest ALL the candidates' numbers for "% of remaining needed to win" go up, time after time, and every candidate is heading up toward 100%, and no candidate is curving down toward 0%, then we are actually on a path to a contested convention.

Be aware, though, until EVERY candidate has gone over 100%, someone can still win. If every candidate other than the leader drops out and stops collecting delegates, allowing the remaining candidate to claim 100% of the remaining delegates, that remaining candidate can still manage to get the delegates they need to win outright.

A contested convention scenario requires multiple candidates who know they are not on pace to get a delegate majority to keep running and accumulating delegates anyway.

That is a pretty tricky path to follow, especially for the candidates who rely on fundraising to keep going.

So far, Sanders is ahead, but it is hard to classify which of these patterns will hold. The situation is even more apparent if you rescale the 2020 Democratic chart to show the entire race:

We have just barely started. Current polling in South Carolina and the Super Tuesday states is driving the current predictions of how this race will turn out. If those polls are correct, then the projections of Sanders winning outright or having a contested convention where Sanders has a significant plurality are probably right.

But we don't know for sure quite yet.

So after South Carolina, and especially after Super Tuesday, come back here and see which ways these lines are all moving, and we'll know a lot better which kind of pattern 2020 is going to follow.

Depending on what we see, this thing may be mostly over, or we'll have indications it will go on awhile.

It will be a fun week. Keep watching Election Graphs!

138.6 days until the Democratic National Convention.

180.6 days until the Republican National Convention.

This post is an update based on the data on the Election Graphs 2020 Delegate Race page. Election Graphs tracks estimates of the convention delegate totals for both parties. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image for the current interactive versions of that chart, along with additional details.

Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of new blog posts. For those interested in more granular updates of delegate updates or general election polling, follow @ElecCollPolls on Twitter. If you find the information in these posts informative or useful, please consider visiting the donation page.

Nevada Eve

Since the last update on February 11th, there have been new state-level general election polls in Alabama, Wisconsin (x2), Texas, Kansas, Oklahoma, New Jersey, North Carolina, California (x2), Florida, Michigan, and Pennsylvania.

Over on the delegate side, we currently have Buttigieg in the lead with 23 delegates, with Sanders just behind with 21, followed by Warren, Klobuchar, and Biden in single digits. But the Nevada caucuses are tomorrow, and that will all get shaken up again, and we will have a new set of narratives to run with.

In the meantime, let's look at how the various contenders are doing versus Trump with the latest polling updates.

Using the poll recency weighted by polling margin metric I use to determine which candidate pairs are the "best polled" with this update Bloomberg vs. Trump finally has better polling than O'Rourke vs. Trump, so even though that polling is still a bit more sparse than we would like we'll start including that matchup in these summaries. (This may be temporary, this only happened because at the moment the Bloomberg vs. Trump margin happens to round to zero in Wisconsin.)

Starting this time with the bottom line, the odds of winning in our probabilistic model:

Dem 11 Feb 21 Feb 𝚫
Biden 98.4% 98.4% Flat
Sanders 74.2% 77.9% +3.7%
Bloomberg —– 59.9% —–
Warren 42.5% 38.3% -4.2%
Buttigieg 11.4% 9.3% -2.1%

First of all, Bloomberg debuts doing better against Trump than Buttigieg and Warren, but not as well as Sanders or Biden. At 59.9% he is better than a coin toss, but not by a lot.

Now, caveats on Bloomberg. So far he has only been polled against Trump in 17 states, and for some reason, many of those states are ones like California or Utah, where we pretty much know who will win and the poll doesn't change the picture very much. In the close states, there still isn't even a single state where he has been polled five times, which means that we are still heavily reliant on past election results rather than 2020 polls to estimate where things are.

So maybe consider the Bloomberg numbers provisional.

Bloomberg has been spending hundreds of millions of dollars on advertising campaigns that for the most part have not been directed at the other Democrats, but instead have been attacking Trump. Bloomberg has been trying to make the case that he is better positioned to win against Trump than the other candidates in part due to his ability to do that. At this week's debate, he also specifically claimed that nominating Sanders would guarantee a Trump win.

The numbers so far just don't back that up.

Where there have been polls including Bloomberg, he generally performs in the middle of the pack against Trump. He isn't clearly doing better than the other Democrats.

Now, the others.

Sanders is the only Democrat who has improved their position over the last 10 days, moving from a 74.2% chance of winning to 77.9%.

Biden still does better than anybody else against Trump, down from his highs, but still at a very strong 98.4% with no significant change since last we looked.

Warren and Buttigieg continue to fade. Current polling shows both to be more likely to lose than to win, and the picture is getting worse with time.

OK, now the view that shows the median electoral vote result in the Monte Carlo simulations for each candidate pair. As always, keep in mind that the simulations actually show that a very large range of results is possible for each candidate, this is essentially just the line marking the middle, where half the time the Democrat does better, and half the time the Republican does better.

Dem 11 Feb 21 Feb 𝚫
Biden +106 +104 -2
Sanders +34 +42 +10
Bloomberg —– +14 —–
Warren -8 -14 -6
Buttigieg -52 -54 -2

With this view, you can see Biden also dropped a little bit, even though his odds of winning didn't change within the rounding limit.

That's it for the probabilistic model.

For those of you who prefer the older and simpler categorization model, where we just look at what would happen if each candidate won every state where they lead the averages, here you go:

Dem 11 Feb 21 Feb 𝚫
Biden +178 +198 +20
Sanders +26 +66 +40
Bloomberg —– +52 —–
Warren -20 -20 Flat
Buttigieg -84 -84 Flat

Interestingly, in this view Biden actually improves a bit. How can he improve here while getting worse or staying flat in the last two views?

Well, without digging up all the specific state by state details, it boils down to flipping from just barely losing to just barely winning in one state (Wisconsin) while getting a bit worse in a few other close states without the state actually crossing the centerline. In the probabilistic view, the declines in the other states slightly outweigh the change in Wisconsin. But in the categorization view, the change in Wisconsin is the only change that matters at all.

Finally, the tipping point, which is in the margin in the state that would put the winner over the top (if the results all matched the polling averages and you sorted the states by the margins). Basically, this shows you how much polling would have to shift across the board in all states to change the winner.

Dem 11 Feb 21 Feb 𝚫
Biden +3.0% +2.8% -0.2%
Bloomberg —– +0.7% —–
Sanders +0.1% +0.5% +0.4%
Warren -0.5% -0.5% Flat
Buttigieg -2.2% -2.9% -0.7%

In terms of the tipping point, Bloomberg actually does slightly better than Sanders. Otherwise, this looks similar to the patterns seen elsewhere. One way to interpret this result is that while Sanders is in a stronger position than Bloomberg in the current polling, that position is a bit more tenuous, and could change more easily.

But bottom line, ALL of these tipping points show a very volatile race. In 2016 the tipping point showed it could swing 5% in just weeks as a reaction to campaign events in the news. None of these candidates are further than 2.9% from the centerline. So with the right stories in the news, they could go from winning to losing or vice versa very quickly.

And that is where we are a day before Nevada.

256.3 days until polls start to close.

For more information:

This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a poll-based 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 to go to a page with the current interactive versions of that 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.

New Hampshire Tie Leaves Buttigieg Ahead

So, New Hampshire went much more smoothly than Iowa, so a few hours after polls closed, the delegates are already locked in.

A lot of coverage has talked about popular vote totals in both Iowa and New Hampshire. This is a mistake. That should be ignored. Just like winning the popular vote did not make Hillary Clinton president, winning the popular vote means nothing in the nomination process. What matters is delegates, and only delegates.

In New Hampshire, the delegate breakdown is:

  • 9 for Buttigieg
  • 9 for Sanders
  • 6 for Klobuchar

It was a tie.

No matter what happened with the popular vote.

Now, what does that mean for the race overall?

Time to look at the "% of remaining delegates needed to win" graph:

Remember, on this graph, down is good, up is bad. When a candidate gets down to 0%, they have won the nomination. If they go up to 100%, they have been mathematically eliminated.

Everybody is still going up!

This reflects the fact that nobody is even getting a majority of the delegates in contests yet, let alone performing at the (currently slightly higher) levels they would need to in order to actually start bringing these lines down.

There are clear differences between the candidates of course. We have two groupings at the moment.

The leaders are Buttigieg and Sanders, because they are going up more slowly than the others.

Buttigieg is doing the best, but Sanders is just behind him.

Sanders is also disputing some results in Iowa. If Sanders gets the best result he can hope for from those disputes, Sanders gains a delegate and Buttigieg loses a delegate, which would move Sanders and Buttigieg into an overall tie. (If that is resolved before Nevada anyway.)

Then there is a gap, followed by Warren, Klobuchar, and Biden (in that order) in the second tier, grouped pretty closely together.

The only actual change in the ordering caused by New Hampshire was Klobuchar overtaking Biden. This puts Biden in 5th place, which is clearly not where he wanted to be at this point.

Lots of people are making prognostications on how the rest of the race will play out based on these two contests. And it certainly does look like Biden's standing in future states has been hurt by his poor performance so far. But it is important to remember that only 1.63% of the delegates have been allocated so far.

All of the candidates still only need between 50% and 51% of the remaining delegates in order to be on pace to win. That is better than any of them have done so far of course, but that is not an outrageous or impossible number.

There is a long way to go. A lot can happen. And we haven't even gotten to the states where Bloomberg has been dumping money yet.

OK, especially at this stage, it may also be helpful to look at the chart in some more familiar ways before we close up.

Here are the results so far in terms of total delegate count :

And in terms of percentage of the delegates so far:

Or for those who prefer tables:

And broken down by state:

Bottom lines:

Buttigieg is the leader, with Sanders nipping at his heels.

Warren, Klobuchar, and Biden are behind, but it is so early, all three of them, and also candidates with no delegates yet for that matter, still have plenty of time to catch up… if they can get ahead of the rapidly growing narrative that the first 1.63% of the delegates have already determined their destinies.

152.5 days until the Democratic National Convention

194.5 days until the Republican National Convention

This post is an update based on the data on the Election Graphs 2020 Delegate Race page. Election Graphs tracks estimates of the convention delegate totals for both parties. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image to go to a page with the current interactive versions of that chart, along with additional details.

Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of new blog posts. For those interested in more granular updates of delegate updates or general election polling, follow @ElecCollPolls on Twitter. If you find the information in these posts informative or useful, please consider visiting the donation page.

New Hampshire Eve

Yes, we had Iowa, but on the general election side of things, since the last update on February 3rd, we have had new polls for Tennessee, North Dakota, South Carolina, New Hampshire (x2), and Kansas.

With the exception of New Hampshire, these are solidly red states and so didn't really do anything to change the national picture. But things did jiggle just a little bit with the two New Hampshire polls, so we'll review the changes on all the major candidate pairs.

But first, for the first time in one of these updates, we actually have delegate leaders in both parties. This may well change as soon as we get results from New Hampshire, but for the moment, the delegate leaders are Buttigieg and Trump.

So before we look at comparisons of the various Democrats vs. Trump, let's look specifically at Buttigieg vs. Trump.

Looking at the Election Graphs probabilistic model, from September through the end of 2019, Buttigieg's position just kept getting worse. It seems to have leveled off a bit so far in 2020, so maybe Buttigieg is hitting a bottom. But currently, things don't look good for him, with the median case in our simulations being a 52 electoral vote defeat by Trump and only an 11.4% chance of an actual victory.

The categorization view, where every state just goes to whoever leads the average, regardless of how close it is, looks even worse. Buttigieg loses by 84 electoral votes.

If the election was today, things would look very grim for Buttigieg. However, looking at the center part of the spectrum of states, you can see that to flip this back to a Buttigieg lead, you need to pull Nevada, Michigan, Ohio, and Flordia back across the line without losing any currently blue states. (There are other combinations of states that would do it too, but this would be the "easiest" group.)

This may seem like a lot, and that 11.4% chance of victory may seem small. But remember, that is "if the election was held today". It does not account for potential future movement.

Of the four states mentioned, Buttigieg is doing worst in Florida. But he is only losing Florida by 2.2% in the polling average. So an across the board margin gain of 2.2% would put him in the lead again. In 2016, this metric (the tipping point) moved more than 5% in a month on several different occasions.

So while the current state of play for Buttigieg does not look great, this far out from the election, for that matter even a month out from the election, a lot can still happen to change a picture like this.

But for the moment, at this very instant, a Buttigieg vs. Trump general election looks like a pretty easy Trump win. If Buttigieg became the nominee, he would have some work to do in order to change his odds.

OK, now let's look at our top four candidate combinations on our national metrics:

Dem 3 Feb 11 Feb 𝚫
Biden +178 +178 Flat
Sanders +26 +26 Flat
Warren -12 -20 -8
Buttigieg -84 -84 Flat

In the "expected case" where everybody wins exactly the states where they lead the average, only Warren loses over these last 8 days, as New Hampshire slips from "Weak Warren" to "Weak Trump".

The tipping point doesn't move at all for any of these four between 3 Feb and 11 Feb, so we'll skip that one and move on to the probabilistic model.

Dem 3 Feb 11 Feb 𝚫
Biden +108 +106 -2
Sanders +36 +34 -2
Warren -6 -8 -2
Buttigieg -52 -52 Flat

In the median Monte Carlo simulation of our probabilistic model, every Democrat except Buttigieg slips by two electoral votes. Slipping a bit, but not a lot. Certainly less change than we've seen in previous updates.

Dem 3 Feb 11 Feb 𝚫
Biden 98.5% 98.4% -0.1%
Sanders 75.3% 74.2% -1.1%
Warren 43.9% 42.5% -1.4%
Buttigieg 11.4% 11.4% Flat

Finally, in terms of chances of winning the electoral college, Buttigieg is flat, but the others all continue to fade a bit.

The Democratic weakening we have been seeing since September may be slowing, but it has not stopped.

And that is where we are on the eve of NewHampshire. As I write this, Dixville Notch, Hart's Landing, and Millsfield have already posted their results. In less than 17 hours, we'll start getting results from the rest of the state. The Democratic nomination race is in full swing.

266.6 days until polls start to close.

For more information:

This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a poll-based 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 to go to a page with the current interactive versions of that 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.

Iowa! (Finally)

I'm sure anyone reading blog posts on Election Graphs already knows about all the drama about getting results from the Iowa caucuses. So suffice it to say that they had some issues.

The 2020 Delegate Race page has been updating whenever new results have become available over the last few days. For the most part, we use the excellent delegate breakdowns at The Green Papers as our definitive source for this information. You can find their current detailed status of the Democratic results in Iowa here. Note that they present a more conservative estimate at the top of the page, but a more aggressive estimate (using more provisional data) at the bottom of the page. Election Graphs uses the second estimate.

The delegate estimates here represent the best estimates for national delegates at the point the Iowa Democrats reported 100% of the vote counted. These may change slightly if there are corrections or recanvasses based on some of the irregularities that were found. And they almost certainly will be adjusted months down the line when national delegates are actually selected at the Iowa Democratic Convention in June.

With that in mind, let's jump right in and explain the central chart that Election Graphs uses to examine the delegate race. It isn't a straight forward chart of the number of delegates each candidate has accumulated either at the present moment or over time. You can find that kind of chart, and others, on the 2020 Delegate Race page. But the most important chart to watch is actually this one:

Rather than the date on the horizontal axis, we show the percent of available delegates that have been allocated so far. In the case of the Democrats this year, this is only the PLEDGED delegates (no superdelegates) since superdelegates will not be able to vote in the first round at the convention unless it is mathematically impossible for them to change the winner.

Using % allocated rather than date gives us a better idea of how far along we are in the race, given that primary and caucus dates are scattered across the calendar irregularly, and the number of delegates available on different dates varies wildly based on which and how many states are handing out delegates that day.

Even more critically though, the vertical axis is not simply a count of delegates. We do have that graph too. But the headline graph shows something that gives a much better idea of how the race is going.

Namely: The percentage of the remaining delegates each candidate would have to win in order to have a majority of the delegates (and therefore clinch the nomination).

If you support a particular candidate, you want this number to go DOWN. When it reaches 0%, a candidate has clinched the nomination. If it goes above 100%, on the other hand, then a candidate has been mathematically eliminated. (Absent pledged delegates being released from their pledges and voting a different way than they were "supposed" to.)

In practice, a candidate can be in a position where they have not yet been mathematically eliminated, but it becomes harder and harder to envision a scenario where they would win. For instance, if a candidate would need 60% of the remaining vote to win, but their percentage of the vote so far is only 40%, unless you know that they are really heavily favored in the remaining states, their chances are actually very slim.

Candidates who are on a pace to win will see their lines moving down.

Candidates who are not on a winning pace will see their lines moving up.

So, what do we see so far after the preliminary results from Iowa?

Well, everybody is moving up. This is quite simply because nobody got over 50% of the available delegates in this first round, which is where you start when nobody has any delegates yet. To move your line down, you need to collect delegates faster than your current "% of remaining needed". If you don't, your line keeps going up, as it becomes harder and harder to catch up.

This is just like how if you are behind in a race, to win you have to not just go faster than the car that is in the lead, you have to go enough faster to catch up with them before the finish line.

As of this writing, the best estimate of the delegate breakdown is:

  • 14 for Buttigieg
  • 12 for Sanders
  • 8 for Warren
  • 6 for Biden
  • 1 for Klobuchar

This seems like an absolutely huge difference between the top and bottom of this list until you realize that only 41 delegates out of 3979 have been allocated so far. That is only 1.03%.

So the "% of remaining delegates needed" varies from 50.18% for Buttigieg, to 50.51% for Klobuchar. (It would be 50.53% for any candidates who still have zero delegates.) These numbers are still very very close to each other.

The news has been filled with pronouncements of the possibility of Biden being doomed by this result or hyperventilation about the momentum for Buttigieg or Sanders. If such a small percentage of the delegates have been allocated so far, and everybody is still pretty close to each other, why is this?

Well… How candidates do in Iowa impacts their perception in New Hampshire. And New Hampshire impacts their perception in Nevada. Which impacts South Carolina. Which impacts Super Tuesday. And perhaps even more importantly, their performance in each state impacts fundraising and media coverage.

In these early stages, the "narrative" dominates. It does matter. A lot.

But in the end, it is all about the delegates. And so far, there is still not all that much difference between the candidates. Anything can still happen.

In terms of the graph above, look for when one of the curves starts heading down instead of up. That's when someone is really getting some momentum. It means that in every new contest, they don't even have to do as well as they have before in order to win. They can just keep chugging along how they have been, and they will end up winning.

For now, though, things can still get crazy.

Finally, before wrapping up, there was another surprise in Iowa besides Biden doing badly and Buttigieg doing well. The surprise was on the Republican side, where Iowa was actually the third state to allocate delegates (after Hawaii and Kansas).

In the Republican Iowa Caucuses, Bill Weld got 1.29% of the vote. Which was enough to get him one delegate out of the 40 available. So we have a race on the Republican side too!

Yeah, OK. Not really. But hey. Weld got a delegate.

157.4 days until the Democratic National Convention.

199.4 days until the Republican National Convention.

It is going to be a fun ride…

For more information:

This post is an update based on the data on the Election Graphs 2020 Delegate Race page. Election Graphs tracks estimates of the convention delegate totals for both parties. The charts, graphs, and maps in the post above are all as of the time of this post. Click through on any image to go to a page with the current interactive versions of that chart, along with additional details.

Follow @ElectionGraphs on Twitter or Election Graphs on Facebook to see announcements of new blog posts. For those interested in more granular updates of delegate updates or general election polling, follow @ElecCollPolls on Twitter. If you find the information in these posts informative or useful, please consider visiting the donation page.

On the Eve of Iowa

Since the last blog post on January 20th, there have been new polls in Florida, California, Iowa, New Hampshire, Delaware, Alabama, Texas (x3), Missouri, and Utah.

As I write this we are less than 24 hours away from the Iowa caucuses, so we'll move the usual "how things have changed since the last update" summaries to the end of the post, and instead, we will look at how each of the best-polled Democrats is doing against Trump as of the day before the Iowa caucuses.

There are many factors that go into voter decisions on who to vote for in primary and caucuses. But this cycle, time and time again, "who can beat Trump" comes up as a critical part of what Democratic voters are considering.

Election Graphs obviously looks at this question in a variety of ways. Our snapshots are "if the election was held today" though. We don't try to prognosticate how things will change in the months between now and November. We recognize things will indeed change. And we've seen very convincing arguments that a big part of the differences we see between how different Democrats do vs. Trump is due directly to name recognition and that if the lesser-known candidates actually won the nomination, they would, therefore, do better than current polling indicates.

None the less, the best view we can get right now is by looking at the current state by state polling. So we will do that. Just keep in mind the limits of that approach.

So here we go. One by one.

Biden vs. Trump

Let's begin the look at Biden's polling by comparing it to the 2016 election results.

If Biden won every state where he leads in the Election Graphs poll average (even by a little bit), then compared to Clinton in 2016, he picks up Florida (29 EV), Pennsylvania (20 EV), Ohio (18 EV), Georgia (16 EV), Michigan (16 EV), North Carolina (15 EV), Arizona (11 EV), and Maine-CD2 (1 EV) while not losing any states that Clinton won.

Clinton's final earned result was a 74 electoral vote loss (it was a 77 electoral vote loss after the faithless electors did their thing). If Biden were indeed to win every state he leads right now, he would have a 178 electoral vote victory.

Now, it is unlikely that Biden would actually win EVERY state where he leads in the poll average, even if the election was indeed today. Many of the states listed above show Biden just barely ahead. So if you look at the probabilistic model, which takes into account the chances of all the close states flipping one way or another, you end up with a median case of only a 108 electoral vote victory for Biden. This is much less than the 178 you get with the more naive view, but still pretty substantial.

In terms of odds of winning based on the past accuracy of Election Graphs averages going into election night, Biden is at a 98.5% chance of winning the electoral college.

Keep in mind though, on this exact date four years ago, Election Graphs showed Clinton with an 80 electoral vote lead if she won all the states where she was ahead in the polls. Later in the year that rose to 188 electoral votes, then it collapsed to only an 8 electoral vote lead on election eve, followed by her 74 (or 77 counting faithless electors) electoral vote loss in reality. So again, never forget that things can and do change, sometimes very rapidly. Most of Clinton's collapse happened in the last few weeks before the election.

Sanders vs. Trump

Compared to Biden vs. Trump, Sanders loses Florida (29 EV), Pennsylvania (20 EV), Georgia (16 EV), and Arizona (11  EV).

Now, all of those states are still close. But they flip to the other side of the line, and Sanders doesn't pull anything in the other direction to compensate.

This leaves Sanders only ahead by 26 electoral votes when you just count up the states in which he is ahead.

Unlike Biden, the probabilistic model helps Sanders. In the Biden case, there were several big "just barely blue" states. With Sanders, there are several big "just barely red" states. So the median in the probabilistic model is actually a 36 electoral vote win.

The odds of a Sanders win come in at 75.3%.

So Sanders would still be a favorite to win over Trump but is significantly weaker against him than Biden.

Warren vs. Trump

This time starting with Sanders vs. Biden, in comparison Warren loses Ohio (18 EV), North Carolina (15 EV), and Nevada (6 EV), but gains back Pennsylvania (20 EV) in exchange. That is a net loss of 19 electoral votes compared to Sanders.

But that 19 electoral vote change puts Warren on the other side of the win line. If both Trump and Warren won exactly the states where they lead the Election Graphs averages, you end up with a 12 electoral vote Trump win.

The probabilistic model helps Warren a little bit, but not enough. When you take into account how the close states might behave, you end up with the median case being only a 6 electoral vote win for Trump.

With things this close, and a slight Trump lead, we're currently showing Warren with a 43.9% chance of winning the electoral college.

This combination is so close that we're currently showing a 2.1% chance of an actual 269-269 tie for Warren vs. Trump, a scenario that would throw the election into the House of Representatives.

43.9% is not zero. Warren could win. But she would be an underdog. You'd be better off with a coin flip at the moment.

Buttigieg vs. Trump

This time starting with Warren vs. Trump, Buttigieg loses Pennsylvania (20 EV) again, as well as Michigan (16 EV). Adding all this up, you have Buttigieg losing to Trump by 84 electoral votes.

The probabilistic model helps slightly, with the median case being only a 52 electoral vote loss. But the odds of winning outright are a dismal 11.4%.

It would not be impossible for Buttigieg to win with these odds, but it would be a bigger upset than Trump's win in 2016 when the median of the odds given by sites that gave odds was a 14% chance of a Trump victory.

Now, as mentioned up top, there is a lot of evidence that even at this late date, part of why someone like Buttigieg is doing so badly is simply that a lot of people still don't know who he is. And presumably, that would change if he was nominated.

But counting on that seems like a big gamble.

How about Bloomberg, Yang, Klobuchar, and the rest?

Yes, Election Graphs does have data for these folks. And more. You can go to the main 2020 Electoral College page and select any of these Democrats and see the data. You can even see combinations with non-Trump Republicans. If there has been a state-level poll including a candidate, they are listed.

But frankly, of the currently active candidates, the four big ones (Biden, Sanders, Warren, and Buttigieg) are the only ones that have enough state-level data to say anything meaningful. Harris was there too. But she dropped out. O'Rourke had a bit less, but still, maybe you could say something. He's out too.

But these others just plain still don't have enough state-level polling. Even though it has been quite some time since Harris and O'Rourke dropped out, we still have better data about them than any of these others.

Of the rest of the active Democrats, Klobuchar vs. Trump has the next best volume of polling, followed by Bloomberg vs. Trump. Look if you want to. The links are right there. But take anything with you see with a huge grain of salt. There has been so little polling, in so few states, that you will not be looking at a reliable picture of anything.

Presumably, if any of these candidates start to look like serious contenders, we'll get a bunch of new polling including them, and that will change. But as of now, state-level polling on these combinations just doesn't have the volume to be meaningful.

The Choice for Democrats

There are many many reasons Democrats may have for picking one candidate over another in the primary process. Policy preferences and character being two of the biggest.

But for those whose biggest deciding factor is simply who is best positioned to beat Trump, the numbers above are pretty clear cut. It is Biden. By a wide margin.

Sanders is next, but it is a bit more of a gamble. Then Warren is slightly worse than a coin toss. Finally, Buttigieg is just a long shot, much more likely to lose than to win.

Now, again, this is all based on current state-level polling. Things will change.

Perhaps if Buttigieg runs the gauntlet and somehow ends up the Democratic nominee this will be accompanied by a huge rise in the polls against Trump. Maybe the Democrats would rally with excitement around a Warren nominee and you would see that big improvement against Trump there too. Maybe something else will happen that will cause Trump to plummet in the polls against all the Democrats.

Or in the other direction, perhaps Biden will completely collapse, and his advantage against Trump will disappear.

Any of the above could happen.

There is a strain of thought that says that because of this kind of uncertainty, the ultimate comparative electability of these candidates against Trump is fundamentally unknowable, and therefore this kind of examination should be ignored.

It is true that head-to-head state-level polls, or corresponding national level polls for that matter, can only tell you where things stand NOW. They are not a crystal ball into the future. But there is still a substantial amount of data there.

It probably isn't wise for anybody to make their primary choices ONLY on this kind of information. It must be balanced against the other factors that lead people to like or dislike particular candidates.

But based on where things are now, there are clear differences in how these four candidates fare against Trump, and it seems to me that it would be folly to ignore that information.

Since the Last Update

OK, that's it for the snapshot in time examination of the Democrats on the verge of the Iowa Caucuses. Now time for the standard look at how things have evolved since my last Election Graphs blog post.

Dem 19 Jan 3 Feb 𝚫
Biden +178 +178 Flat
Sanders +26 +26 Flat
Warren -12 -12 Flat
Buttigieg -84 -84 Flat

So, in the straight-up, what happens if everybody wins all the states they lead, there were no changes this time around.

Dem 19 Jan 3 Feb 𝚫
Biden +3.0% +3.0% Flat
Sanders +0.1% +0.1% Flat
Warren -0.5% -0.5% Flat
Buttigieg -1.6% -2.2% -0.6%

In the tipping-point view, only Buttigieg declined.

Looking at the numbers above, keep in mind that in 2016 the tipping point moved more than 5% in mere weeks on more than one occasion. Things can be very volatile once we get into the final stages of the campaign. A movement of just a couple of percent in the tipping points above would lead to a completely and totally different picture of the race.

Frankly, in the end, ALL of these tipping points show a close race.

With the categorization view, things have looked more static, and we see changes less frequently, because states have to jump categories or move past the existing tipping point to make a change.

So let's move to the probabilistic view, which is much more sensitive to individual poll changes, regardless of any category changes which may or may not happen.

Dem 19 Jan 3 Feb 𝚫
Biden +126 +108 -18
Sanders +48 +36 -12
Warren +4 -6 -10
Buttigieg -50 -52 -2

All four Democrats worsened their positions against Trump in the median case for the Election Graphs probabilistic model. Biden drops the most, while Warren flips to the loss side of the centerline.

Dem 19 Jan 3 Feb 𝚫
Biden 99.2% 98.5% -0.7%
Sanders 79.9% 75.3% -4.6%
Warren 52.4% 43.9% -8.5%
Buttigieg 14.2% 11.4% -2.8%

Finally, the win odds.

Once again, all four Democrats slipped in these last 15 days.

This continues the movement away from the Democrats and toward Trump that started in the fall. Judging from the graphs, there are some signs that the movement is slowing. But it does not appear to have stopped yet.

The Democrats are still bleeding.

Perhaps the start of the delegate battle with the Iowa Caucuses will be a turning point. Or not. We shall see.

In the meantime, the 2020 Delegate Race page is itching to get some real data. As soon as results start coming out of Iowa, that page will start being updated with the delegate totals as they develop, along with calculations on what that means for the remainder of the race. And of course, there will be blog posts here after each major primary or caucus with an analysis of what it all means.

Buckle up, here we go.

274.6 days until polls start to close.

For more information:

This post is an update based on the data on the Election Graphs Electoral College 2020 page. Election Graphs tracks a poll-based 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 to go to a page with the current interactive versions of that 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.