Alan Lichtman and Cenk get in intense argument, lichtman accuses cenk of blasphemy

[SteveRogers]

Lol, I always have a simple rebuttal in these discussions - show me your superior model.

Once again:

Can you please explain to me how it even counts as a “model” if one of the keys, namely the third party vote being over or under 5%, cannot be determined until after the election? How can the model claim to predict the election result if one of the keys requires knowing the vote totals from the election?

[Senator Incitatus]

I don’t know why Lichtman gets so much hate TBH, anyone who accurately predicted the 2016 election is pretty impressive IMO.

If aliens landed on Earth tomorrow, I wouldn't suddenly go around listening to every person who claims an alien gave them a colonscopy.

[Long Live The King!]

Lol, I always have a simple rebuttal in these discussions - show me your superior model.

Maybe we should stop fixating on finding the 'correct' formula to predict how millions of people will vote and acknowledge that elections are inherently unpredictable - only the voters, not some obscure scientific formula or some all-knowing person's prediction, will determine who will win. No one can determine the winner until the voters cast their votes in the ballot box.

[Big Jim]

Never liked AL. His "keys" been wrong before. He'll just make up some excuse and keep on keepin on.

[Voice of low info America]

My favorite part is how he lashes out and talks about how long he's been a (history) professor and how many books he's written as if that disqualifies criticisms of his "model". A set of mostly vague, equally weighted, not mutually exclusive true/false factors is not even a model, and would not be taken seriously in any analytically rigorous environment.

Also, it should be obvious Piers Morgan, tabloid editor at heart, set up this panel knowing the potential for fireworks. I'm conflicted on whether I should be thankful for the entertainment or hate him for being a terrible person.

[○∙◄☻¥tπ[╪AV┼cVê└]

I don’t know why Lichtman gets so much hate TBH, anyone who accurately predicted the 2016 election is pretty impressive IMO.

His prediction was specifically that Trump would win the popular vote, which was wrong in 2016.

Doesn’t he claim that he changed his model after 2000 to predict the EC instead of the NPV?

Regardless he got one of 2000 or 2016 wrong as well as 2024 wrong.

[Bush did 311]

I think I could come up with a better model than Allan Lichtman. I call it the 3000 dorknobs. It takes everything from the dewpoint humidity on election day to all the esoteric economic and sociological indicators, and runs them through AI trained on previous US elections to predict an outcome.

[qrdlelections]

If the standard for criticizing a public figure is being able to do it better

I mean, good faith discussion please.   The great thing about models is they are open source, bro.

[qrdlelections]

Regardless he got one of 2000 or 2016 wrong as well as 2024 wrong.

I don't think he got 2024 wrong, I think it just didn't apply.  Arguably, it didn't apply very well in 2016 either though given Trump's unprecedented background.

Like, seriously.  You have to weight and blend diverse models.   Obviously nobody on this thread builds models or does inference for a living.

Anyways, a more interesting question imho - how will the keys apply in 2028!

[SteveRogers]

If the standard for criticizing a public figure is being able to do it better

I mean, good faith discussion please.   The great thing about models is they are open source, bro.

Is. It. Even. A. Model?

Can you please explain to me how it even counts as a “model” if one of the keys, namely the third party vote being over or under 5%, cannot be determined until after the election? How can the model claim to predict the election result if one of the keys requires knowing the vote totals from the election?

[Blazers93]

I correctly predicted the winner of 5 of the last 6 elections on this site. They should have me on the panel!

The keys are a pretty flawed system but I think models like this are interesting. I'd argue that Biden didn't have a major policy achievement and that Trump is a charismatic challenger (despite being so polarizing). I bet he returns in 2028 with revised rules to be a punching bag once again

[President Johnson]

I don’t know why Lichtman gets so much hate TBH, anyone who accurately predicted the 2016 election is pretty impressive IMO.

I don't know either. Yes, he was wrong this time and some of his keys are subjective, but I consider him one of the good guys. Not a fan of Cenk.

[Trump v. Wong Kim Ark]

Facts:  I love the keys, and repeatedly told people it won't work this year.

Lichtman is a smart (if unstable) guy who did a great historical study and came up with a beautiful model similar to what's used for predicting earthquakes.

Are you kidding me???

I did my PhD in an institute which is arguably the best in the world for seismology research. My colleague there all said there is no way to predict earthquakes. These are the best scientists with the best statistics and equipments. You think a crapy "historian" can come up with some "models" based on a few arbitrary keys that does better than them???


Also see USGS https://www.usgs.gov/faqs/can-you-predict-earthquakes
"Can you predict earthquakes?
No. Neither the USGS nor any other scientists have ever predicted a major earthquake. We do not know how, and we do not expect to know how any time in the foreseeable future."

Lichtman is a complete liar, period.

Yeah, with earthquakes, the best we can do is monitor for when they start and then send out warnings immediately. Very helpful to know if you need to stop a pipeline or a train or something, but it’s not something you can wake up and see if there’s going to be one at 12:23 PM today and plan lunch accordingly.

I’m glad that Lichtman’s “model” is going down in flames. Hell, my vibes weren’t right for this election, but I was going on national TV saying everyone else was full of BS while trying to justify it using some hocus pocus that he pretends hasn’t changed due to misses 2000 and 2016.

The stats somehow showed themselves to be decently accurate this time and I know I’ll focus more on that when I look at 2028 predictions.

[Torie]

 His career is over,and he knows it. That was why he was so nasty. He sees it all going away. We wont't be hearing about keys anymore.   

[Whale Psychiatrist]

The keys are never wrong, only the one turning them

[Chief Justice PiT]

     I'll admit that I can understand Lichtman's frustration. Building up a position is much harder than tearing it down, and Cenk's complete dismissal of Lichtman's work when he likely couldn't do better himself has to sting. Not saying that we can't criticize him, but the asymmetry there is real. The problem for Lichtman is that the keys clearly have flaws and he has to take the criticism with more grace and humility. His explanations for the miss were ridiculous and it's more likely that, as Nate Silver suggested, he simply applied his method incorrectly.

[iceman]

this Allan Litchman guy should be put to oblivion and never to receive attention again in the next elections. He is a desperate, bitter hack that tainted his own reputation, if there’s one.

He deserves all the ridicule and mockery that he is getting now. Schadenfreude for this guy.

[Asta]

Nearly no one has argued this, but the model's biggest flaw is not the subjectivity, but that each one of the checkbox is apparently given equal weight. Things like social unrest or charisma doesn't outweigh economic fundamentals, but if one insists, grade it.

Was there a recession/inflation at some point? Grade it based on the severity on a scale of 1-5. This should have been given the biggest weight by far. Lichtman gave the point toward Harris for inheriting a good economy from Biden, but that's where his model was incorrect.

Had he given huge weight to economy under his model, he would have predicted a Trump win.

[Averroës]

lolcow vs. lolcow moment - really depressing to think of how much attention these guys get

[BaldEagle1991]

I correctly predicted the winner of 5 of the last 6 elections on this site. They should have me on the panel!

The keys are a pretty flawed system but I think models like this are interesting. I'd argue that Biden didn't have a major policy achievement and that Trump is a charismatic challenger (despite being so polarizing). I bet he returns in 2028 with revised rules to be a punching bag once again

He isn’t. He has no charisma, just a promise that he can dumb down to his supporters.

[Chief Justice PiT]

Nearly no one has argued this, but the model's biggest flaw is not the subjectivity, but that each one of the checkbox is apparently given equal weight. Things like social unrest or charisma doesn't outweigh economic fundamentals, but if one insists, grade it.

Was there a recession/inflation at some point? Grade it based on the severity on a scale of 1-5. This should have been given the biggest weight by far. Lichtman gave the point toward Harris for inheriting a good economy from Biden, but that's where his model was incorrect.

Had he given huge weight to economy under his model, he would have predicted a Trump win.

     This is an excellent point. I do predictive modeling as part of my work, and one of the key parts of making a model work is weighting the factors correctly. Reading about his keys it sounds like he developed it as some kind of predictive model, so it's absurd to pretend that all of the keys just somehow happened to be weighted the same. It's strange too that he presented it as some kind of open source model so anyone could use it to try and divine the future, but then he gets angry if someone draws different conclusions from his own about who it should predict.

[Stranger in a strange land]

this Allan Litchman guy should be put to oblivion and never to receive attention again in the next elections. He is a desperate, bitter hack that tainted his own reputation, if there’s one.

He deserves all the ridicule and mockery that he is getting now. Schadenfreude for this guy.

I had thought that a silver lining of this election was that we would never have to hear about Lichtman and his keys again, but then someone on another forum noted that he can just say his model is valid but the keys he chose were just wrong, so he'll just have to forge new keys and all will be good.

[qrdlelections]

Lol, I always have a simple rebuttal in these discussions - show me your superior model.

Maybe we should stop fixating on finding the 'correct' formula to predict how millions of people will vote and acknowledge that elections are inherently unpredictable - only the voters, not some obscure scientific formula or some all-knowing person's prediction, will determine who will win. No one can determine the winner until the voters cast their votes in the ballot box.

Nothing is inherently 'unpredictable', there is just some margin of error.

It's just a lot of hate from people who don't understand the nature of modeling and predicting.  

I mean, you all don't understand, the keys were correct - it's the voters that voted wrong.

Anyways, more interesting discussion is what the Atlas Intel(AI) guys did.  whoo!

(yes, the above prior was a joke.  Seriously people, chill.  it's just another model, just like selzer is just another poll.  if you were unable to weight them properly, that's not their fault.)