Your gurus were wrong

I'm not going to dance on the graves of my enemies here, but Alan Lichtman was wrong about a Kamala Harris win.  The thing is, his model wasn't wrong, just him.  The old saying garbage in - garbage out explains why he was wrong, even though he clearly doesn't get that yet.

He claimed 9 of his 13 predictors favored Harris. For the record, here the are, in a nutshell:

Key 1 (Party Mandate): After the midterm elections, the incumbent party holds more seats in the U.S. House of Representatives than it did after the previous midterm elections.

Key 2 (Contest): There is no serious contest for the incumbent-party nomination.

Key 3 (Incumbency): The incumbent-party candidate is the sitting president.

Key 4 (Third party): There is no significant third-party or independent campaign.

Key 5 (Short-term economy): The economy is not in recession during the election campaign.

Key 6 (Long-term economy): Real per-capita economic growth during the term equals or exceeds mean growth during the previous two terms.

Key 7 (Policy change): The incumbent administration affects major changes in national policy.

Key 8 (Social unrest): There is no sustained social unrest during the term.

Key 9 (Scandal): The incumbent administration is untainted by major scandal.

Key 10 (Foreign/military failure): The incumbent administration suffers no major failure in foreign or military affairs.

Key 11 (Foreign/military success): The incumbent administration achieves a major success in foreign or military affairs.

Key 12 (Incumbent charisma): The incumbent-party candidate is charismatic or a national hero.

Key 13 (Challenger charisma): The challenging-party candidate is not charismatic or a national hero.

Here's the real score, via Tim Pool. And to be fair, Tim Pool is even being generous to Lichtman in his interpretation of the factors, and it still lands on Trump:

Just to top on that, Ann Selzer predicted, rather insanely, that Harris would win Iowa. How anyone believed her makes me incredulous.

The Iowa poll, conducted by Selzer & Company for The Des Moines Register/ Mediacom, found Harris had a three-point lead over Trump in the state, 47 percent to the Republican's 44 percent.

The poll surveyed 808 likely voters in Iowa between October 28 and 31 with its shock result falling within the poll's 3.4 percent margin of error.

Announcing her findings on Saturday, Selzer, previously described as "the best pollster in politics" by aggregator FiveThirtyEight, said Harris had "clearly leaped into a leading position."

However, the Iowa poll turned out to be wrong by 16 points following the results of Tuesday's election. The former president has won Iowa by 55.9 percent to Harris's 42.7 percent, a difference of 13.2 percent so far.

Sorry media, your gurus were wrong. Go get some new ones.

State of the race: October 14th

 

Click to enlarge.

Taken from the RCP polls, and including only polls that are likely voters, with a sample size greater than 500 or more, and a margin of error less than or equal to 3.5% and polls taken on September 30th or later.

This map would lead to a Trump win with 306 electoral college votes, the same as in 2016.  I think this time around he may do better than this shows at the moment.

NOTE:  Applying the same logic to national polls, Kamala Harris appears to still be leading in the popular vote, but not by a great margin.

Pre-Trudeau vs. Post Trudeau, as predicted

Former Canadian prime minister Stephen Harper told Canadians this would happen under Trudeau.  Unfortunately not enough Canadians listened.  Thankfully in another year or less, Justin Trudeau will be in our rearview mirror.

Based on believable polls

As of yesterday, this is what I was seeing:

(Very) Early 2024 prediction

Via Red Eagle Politics, this is how we EXPECT it to turn out. But uh, maybe not after all the 'counting' is done.

President Trump, comeback kid

Red Eagle Politics says president Trump is in line for an epic comeback:

Senate 2024 update

 Via Red Eagle Politics, a look at the 2024 senate race situation:

Well, this is it.

Well, this is it. Either it was a great Trump run, or it's the middle of great Trump run. Let me start with the Senate and Congress as far as predictions go, I have not spent much time on these races unfortunately.  I expect the Republicans however, to maintain a slim margin in the Senate +/- 1 seat. Congress I have no clue.  I expect Republicans to pick up a handful of seats but they have an outside, long shot of winning back Congress.  That would be awesome.

As for the presidency, I have two views.  My technical model view, and my tweaked view where I apply some common sense removal of outlier polls. Here we go.

The numbers call -- The day before the election and my polling model is predicting a narrow Trump win with 275 electoral college votes, Biden with 249 and 14 (Minnesota and New Hampshire) too close to call.  Interestingly in the discounting and unbiasing of polls that I perform on the RealClearPolitics polls of swing states, I'm seeing president Trump winning Michigan and Arizona but losing Pennsylvania and Wisconsin.  The two closest states are Arizona (a win of just ~8,000 votes) and North Carolina (~15,000).

So that's my model.  But it's based on a lot of late breaking polls that once again show Biden with a larger advantage than in the last few days of polls.  So I don't trust it and I have to tweak it a little to come up with my actual projection.

My tweaks -- After tweaking my results to discount polls that look valid on the surface but are very clearly outliers, here's what I see:

Texas (38 electoral college votes): Trump +4, with a vote lead >200,000.

Florida (29): Trump +5.5%, with a vote lead greater than his win in Texas, approaching 300,000.

Pennsylvania (20): Trump +0.9% with a lead of just 32,000 votes, and serious legal (or not so legal) attempts by Democrats to alter the results.

Ohio (18): Trump +5% with a vote lead of >165,000.

Georgia (16): Trump +2%, and a vote lead >50,000

Michigan (16): Trump +2%,  and a vote lead >50,000.

North Carolina (15): Trump +0.8% and a vote lead of just 22,000.  This is another state Democrats could very easily try to steal.

Virginia (13): The pollsters have this a lot higher but I see a Biden win by only about 4.5%.

Arizona: Trump +1% and a vote lead of just 19,000 votes.  Interestingly this is just removing one NYT/Siena poll which is clearly a late stage outlier.

Wisconsin (10): This one is a coin toss, even after removing the late stage NYT/Siena outlier poll. My model calls it for Biden but my gut says Trump. To be cautious, I'll leave this one in the Biden column for my prediction, but I honestly think Trump takes it by about 25,000 votes. Unfortunately I would have to deviate from my methodology too much to make the call for Trump.

Minnesota (10): I have Trump winning Minnesota by 1.7%, or 28,000 votes. 

Colorado (9): With no decent polling I'm just going to leave this in the Biden column. If Trump wins Colorado, the election is going to be a big red wave election. As nice as that would be, I don't see it happening, not to that extent.

Nevada (6): This is a Biden +4% win ( >30,000 votes) according to my model. I think Biden will win it but it's going to be closer than that, perhaps 2% or 15,000 votes.

Iowa (6): Trump +8.  This will not be close according to my model, or me.

New Mexico (5): There are no dependable polls in the state.  They show Biden ahead by double digits. I think he wins by 6% though.

New Hampshire (4): Unfortunately the polls have universally leaned Biden far enough that he seems to have an insurmountable lead here.  While I believe the even the 'valid' polls are overstated, unless the bias is far more pronounced than in 2016, Trump is not going to win the state. I have Biden +8% in my model. My gut says that Biden is inexplicably more likeable to people than was Clinton, so I'm going to leave it as is; a strong Biden win.

Maine (4): This state is going to go for Biden by an even larger margin than New Hampshire.  I have not seen any polls on District 2, but I'm hopeful that president Trump picks it off and gets the 1 electoral vote.

So here's my net prediction - Trump 306, Biden 232, and the president flipping Minnesota in exchange for losing Wisconsin. Trump may even pick up Wisconsin and get 316.

Predicting the presidential race comes down to belief

The 2020 election comes down belief. I'm not taking just who believes in their candidate.  I'm talking about predicting who is going to win depends on what and who you believe.  The pols seem wonky and over-dramatically pro-Biden. But that's what I believe. I believe that man cannot generate enthusiasm.  Then again, you might believe there's a rabid anti-Trump enthusiasm that overshadows Biden himself.

Maybe there's a revenge factor for Democrats who really, really wanted Hillary Clinton to  win.  Conversely Trump supporters have a reason for revenge of their own - 4 years of phony impeachment to oust a duly elected president.

Now election  prediction - that's a different sort of beliefs, going back to what you believe about the polls.  In 2016 they were wrong. Badly wrong.  Oh, granted in the final days they got close. But that was right at the end.  I pointed this out yesterday.  The real question is: "Is the same thing happening again in 2020"? It's hard to believe that the Democrats chose a candidate that is even more lackluster than Hillary Clinton was, but they actually did it. So yes, I believe that the polling is again in 2020 either erroneous or deliberately misleading.  How much so?

That's the real question. I'm seeing some truly dramatic election impacts if the reasonable and recent polls (margin of error less than 4.3%, done in the last 13 days and by reasonably reputable pollster) are scaled to the same level of bias (or less) compared to how off they were in 2016.

For example, if the polls are only 80% as biased towards Biden as the were to Hillary. Biden has almost won.  In this scenario he probably gets all of the states that cannot be called based on polling, except maybe Iowa and ends up 305 electoral college votes - a solid win.  However, in order to ensure that he wins the level of bias would have to drop all the way to 45% where Biden would also capture Florida.

But if the bias is 85% of the Clinton bias, Trump is actually leading right now.  And interestingly, Pennsylvania is the first state to flip from Biden to Trump's column.

At 90% of the Clinton bias in polls applied to 2020, North Carolina flips from Biden to Trump as well, but he still has not won.

At a bias level in late October 2020 matching that of October 2016, Michigan flips and Trump wins.

If there is zero bias in the polls Biden has 308 electoral votes and likely garners most of the 48 undecided states. His potential ceiling is 356 electoral college votes.  That shows this is definitely a matter of belief.  After all the bias could be even worse this time around. Biden is less energizing than Hillary Clinton was, right?  If the bias is 125% of what Hillary got in October 2016, Trump get 288 electoral votes comfortably and at possibly 28 of the undecided 48 undecided states (where the polling is just not sufficient to make a plausible call), for a total of 316 electoral college votes.

What do you believe? The polls are right, or close, or wrong?  That's what leads to an inability to make a reasonable prediction right now and we are less than a week away.

For the record, I see no evidence that the pollsters recognized and adjusted their methodologies from 2016. As a result, I believe Trump is going to win (minus any cheating). I just don't know how soundly he is going to win. My hope is that it's enough to overcome any potential "insurance policies" on the part of the left.

Nevada GOP caucus predictions

Predicting that Trump is going to win in Nevada is not exciting given recent polling. Who comes in second or third also isn't all that big a deal.  What I'm going to focus on is who is going to drop out after tonight.

Ben Carson.  Just a hunch.  But he can now say he lasted longer than Jeb Bush. 

And that's it.  Even though he should drop out to benefit the establishment and also because his chances of winning are approaching zero, Kasich is going to stay in for at least another week.  It's pointless for him to continue, but I don't believe he's going away.  For better or worse, Trump is benefiting from a split field.  Until Ted Cruz drops out, the eventual departure of  candidates will benefit Rubio first, and Cruz second.  

The longer this carries on, the more likely it is that Trump will be the nominee.

Fearless predictions about Iowa caucus results

My predictions can best be qualified as "wonky".  That's on a good day. But what the heck, I might as well take a stab at what is going to happen in the Iowa caucuses today/tonight.

2016 seems to be shaping up as the year of the outsiders on both sides.  My predictions are being made in light of that assumption. So they should be relatively obvious. 

Bernie Sanders is trailing Clinton in the latest polling but has rapidly been gaining ground.  I think his momentum will carry him across the finish line first. It'll be close, but I think he's going to win.

On the GOP side, weather impacts aside (snow is expected), I think Trump is going to win.  It'll be a test of how deep his support runs, rather than how wide.  How loyal will his caucus-goers be if the weather is bad, or if the caucus gets heated?  We'll see.  But I think he'll pull it out.  What will be more interesting is to see how the rest of the rank ordering plays out. I think Cruz, Christie and Rubio fare the best and Jeb Bush will do decidedly poorly.

Let's see if I'm right tonight.  Tomorrow is Ground Hog Day, so maybe I'll get to be Bill Murray and get a do-over.