In 2016 I was listening to the 538 podcast with Silver. Right before the election they were discussing the forecast and someone asked "so Hillary is at 70%, that means she wins right?" Nate immediately countered that 30% probabilities happen all the time. For perspective, another guest added: "Russian Roulette is only 16.6%." As I anxiously waited for results on election day I found it hard to dismiss the thought: "there are two bullets in the gun."
Silver would have been "wrong" if in every world where that election happened Hillary won. to be "correct" Trump has to win in 3 of ten scenarios. We sadly live in one of those 3 scenarios....
I would say that he made the wrong prediction. I doubt that the difference between the election prediction and the outcome came from the polls being unrepresentative because of bad luck, but for systemic reasons of how the models translated polling results into a predicted outcome
UA-cam posts the upload date, but it's not really useful for time-sensitive content. You folks post a lot of timely information. Please post the dates that these interviews were recorded.. Otherwise, thanks for all the great interviews, including this one.
The race has not quantitively changed in any significant way since arguably before the debate. It probably won't change much up to election day. All the variation in the polls has been within the margin of error. Very few people are changing their minds.
28% is also a really high probability but the issue is that any probability besides 0%, 50% and 100% is very unintuitive to most people, and in my experience they just end up rounding to one of the intuitive 3 so when you tell someone Hillary has a 71% chance of winning they think you're saying she is going to win.
honestly i think a major issue is that these projections are often plotted the same way as polling data, and people don't separate them in their minds. being projected to get 60% of the vote vs being projected a 60% chance to win are *very* different things.
With regard to Alastair’s point about Nate being wrong in 2016, a perfectly accurate estimate of a 30% chance will be ‘wrong’ 1 in every 3 elections, and he was also getting a lot of criticism at the time for having Trump’s chances so high and not calling it a dead certainty. He doesn’t conduct polls himself so all he can do is estimate the chances based off of the existing polling data which had massive errors.
A tech bro American who confidently speak like they know everything. Characterising UK as rule-following because our new tube lines have barriers in comparison to NYC which doesn’t have have any new lines. And yet in the UK we don’t have jaywalking, we don’t have strict rules around women’s freedom to abortion, we didn’t implement a series of strict rules which separated the black population from whites, etc…
As far Alastair wondering about voters not thinking about what disqualifies Trump, there is a lot of factors, but alot of it comes down to a an anti-establishment realignment and the conservative media ecosystem. The conservative media really is an ecosystem. Unlile the UK where conservatives are a lot more present in the mainstream media, conservatives in the US have just opted out of that basically entirely. There is less of a link for voters who lean right to get those contradicting perspectives. And attempts to establish those links can be contentious at best because you're poking so many holes in a thing that forms the foundation of a worldview. As a result, the right wing media ecosystem has been able to spin Jan 6 as either a non-concern or justified. The second reason is the anti-establishment realignment which composes of people who are much harder to define politically because their views are kind all over the place, but also includes people who feel burned by the system. Regardless, a lot of the people has coalesed around Trump since so much of the political establishment has gone for Harris due to her defense of institutions.
People think that MAGA is some kind of position or ethos, when it is actually a lifestyle choice. There is no data, no persuasive argument, no appeal to any quality or ideal that has any impact on a lifestyle choice. I can only change from within the adherent. Most of the time that will come too late if it is some kind of time sensitive requirement. America was already broken and teetering in the edge of the cliff. Russia just pushed it a little bit.
Democrats and left leaning people don't get much of right leaning prospective themselves, certainly not the best version of it anyway. I think the anti-establishment angle of it is huge. Things have changed since January 6th, but especially before that, I think there were alot of people who voted for Trump hoping to send a message to Democrats and the establishment Republicans in general to try and force some changes and new thought. People should consider that every person who voted for Trump hasn't gone to a Trump rally and doesn't own a red hat.
Big fan of this series and but as a minor point of order, could we have the recording date placed in the description? Sometimes you mention the date in the interview, but it makes understanding the context easier if we know this interview might be 2/3 weeks old. Many thanks!
I recall Nate Silver saying that the 70-30 probability in 2016 equated to someone missing a 42 yard field goal in the NFL. I’m a giant fan. We beat Buffalo in the Super Bowl because their guy missed a 44 yard field goal in Miami in beautiful weather, no wind. so when I read that, I got chills up and down my spine because I know firsthand how often 4 2 yard field goals are missed.
You guys don't seem understand how Silvers' model works. His model was excellent in 2016. When he calls it 29-71 in Hillary's favor, that is a little less than one in three chance that Trump will win. In addition in the lead up to the election other models were calling it 2-98 in Hillary's favor. He's not predicting who will win. He is saying, based on historic outcomes in similar circumstances, these are the probabilities. If you can go back and read what he was writing in the lead-up to that election, he was warning people that other models were skewed in Hillary's favor. Think of it this way, if he calls ten races at a 30-70 probability, the *expected* outcome is that three of the dark horse candidates will win their races.
Right. But ultimately this is so much dancing about architecture. The statement "Today, X has a 1 in 3 chance of beating Y" is a useless statement. You either get the outcome right or you don't from "predictive models." In 2022 Silver characterized anyone who thought Democrats would win the Senate as taking "copium." He was dead wrong. We have reached the point where polls tell us next to nothing useful.
@@Bazball1080 It's not TOTALLY meaningless... just a lot less meaningful than people think. Unfortunately, I don't know of any better methodologies out there.
The commentary on the Trump chances is quite strange. There’s a 1 in 3 chance of rolling a 5 or 6 on a die, but nobody is surprised when they see a five or a six. That also doesn’t mean the chances were wrong and nobody would think that.
While I understand the hosts don't define their probability terms well, Nate is getting away with something dishonest here. He is quick to (rightly) dismiss the criticism of the 2016 model, but doesn't say a word when his model 'correctly predicted' the outcome in other election years. By his logic, the model was no more 'predictive' in the years when it was credited than it was in 2016. So, for his 2012 model, for instance, he should jump in and say, "actually I didn't *predict* a Barack Obama victory, I just provided a specified range of likely outcomes." But he doesn't say that. He takes credit for predictiveness when it suits him, but denies that's his aim when it doesn't.
doubtful Shapiro would have done ditto for PA. Running mates rarely make a dent, they just need to be a useful surrogate who doesn't actively damage the campaign and drag on the ticket
There is a fairly wide consensus that the choice of vice-presidential running mate makes little difference to the outcome of the election, except perhaps in bizarre cases. Doesn't that suggest that a presidential candidate ought to nominate a VP candidate on the grounds of how effective he will be in office and whom she wants to work with for five years. Perhaps the issue is not "Josh Shapiro might have delivered a vital few thousand votes in Pennsylvania", but rather "Tim Walz will be a good VP for Kamala Harris, if the electoral system chooses them".
@@Dan-y6l3k The VP has very little constitutional authority. On the other hand, a VP who is the trusted delegate of the President (such as Dick Cheney or Al Gore) can swing quite a big stick, can act as a Cabinet-level right-hand-man and minister-without-portfolio. Who better to have beside you at a Cabinet meeting etc.? Someone who thinks you should be President, or someone who thinks he should be President.
@@brettevill9055 Wonder what the odds are of a POTUS dying in office are. Might make the VP pick absolutely critical. Which one would you want takin' the reins?
Have I got this right? I am right when I am wrong, and I am wrong when I am right. I was not wrong because I got closer than I predicted so that makes me right. ???.
His description of the UK as a place where the Media is thriving and Gambling isn't an issue? Gell-Mann Amnesia effect emanating strongly off this guy.
There are a lot of people registering to vote that have never voted before. Pollsters use lists of likely voters. So it will be interesting to see if they vote and who they vote for.
So enticing. There was so much in that interview that should have been explored so much deeper. But alas ... . Also this is a wonderful demonstration of why not being statistically numerate is such a big, and growing, disadvantage. There will be AI soon that will be able to more reliably predict elections based on each and every click, world wide. Scary.
I think it’s important when looking at polling, and as a Democrat it worries me, the resilience of Trump voters They were deeply underestimated in 2016 which is why people like Silver got that election wrong. And in 2020, they were again underestimated and Trump barely lost states like Wisconsin by a few thousand votes where Biden was supposed to win by 3 or 4 points That said, undecided voters are more about whether or not those people will vote. So get your ass out to vote.
One of the bullish cases that can be pointed to is that in basically all special elections in this past cycle, democrats have massively outperformed polls due to turnout caused by issues like abortion in particular, but the caveat is we don't know if this polling error will translate to general election polling where turnout is more likely in general and people are more motivated to come out for issues beyond abortion
Trump is at his ceiling, which must be lower than 2020 if the insurrection and convictions have had even a slight effect on non-crazies. If Harris enthuses her coalition to turn out in the same scale as they did for Biden in 2020, we will see the same result. And she seems a lot more appealing to young, female and black voters than Biden was. It's going to come down to turnout. 2016 Dems stayed home and Trump won. 2020 Dems showed up and Trump lost.
I think there are a lot of people like my parents, conservative Christians who won’t vote for him after January 6. The Stormy Daniels thing was so cringe too. I guarantee they would not have voted for him had they known.
I’m not clear why making Josh Shapiro the Vice-Presidential pick would have been a better choice. If Pennsylvania is critical, why would taking its very popular Governor out and sending him to Arizona, Nevada, North Carolina and New Hampshire be smart? Keeping Shapiro campaigning in PA is clearly the far better move.
@@DGE123Obama killed the shapiro option. A couple of stories were released in the press. He served in the idf, not the US army, plus he tortured Palestinians in prisons. The second is an article he wrote in college, very anti Arabs, critical of Ytsak Rabin for the Oslo process, trying to make peace with the Palestinians. Obama feared others papers would come out.
They should have got Allan Lichtman on. He has a far better track record of predicting outcomes (9 out of 10) with his 'Keys to the White House' model.
Gore won. The supreme court stopped the counting in FL, very unconstitutional. They happened to be black votes too. When all was said and done after the counting, Gore won FL
Right? Silver mentions Labor Day, which was September 2nd. His prediction of 50/50 is reflective of a Kamala polling slump from that time which has been completely erased.
You saying that a 30-odd percent chance is ‘you predicted Trump would lose’ is like if someone tosses two coins and they come up both heads, and you respond ‘but you said this wouldn’t happen!’ (When in reality you predicted a 25% chance of this happening). No one would do that, but because it’s politics we expect probability to mean different things?
It's always interesting to me how many Brits and Europeans mock or downplay when Americans try and claim some connection or brotherhood to Europe via heritage, tradition, etc. But then also act so confused that politics and society in general in the US dont do things just like their country. Almsot like the US is just a different place. I also often find that the same people are very good at practicing cultural relativism with various other cultures and countries but seem offended that the US has different ideas and ways of doing things lol.
Just a plurality of votes would give urban voters way too much influence. Especially when urbanization & market economics are no longer a determinant in 21st Century unlike the last century. Electoral College System is a check and balance that forces the 2 parties to appeal to a wide variety of voters outside of their own stifling echo chambers. Democrats need to appeal to rural voters while the GOP needs to appeal to suburb & exurb voters.
Excellent point, in Canada a very bad man called Justin keeps winning with only a couple of major cities and is on record as disregarding vast rural and the entire western part of the country and its a majorly divisive and is why all these left woke encourage massive immigration and migration to major cities
The problem with Nate's defense of the 2016 model is it starts to break down when you look at the individual states. Nate projected Hillary was favored to win all of the following states: - 83.5% chance to win in Wisconsin - 78.9% in Michigan - 77% in Pennsylvania - 55.5% in North Carolina - 55.1% in Florida Trump won ALL of these states. According to Nate's model, the chance of that happening is around 1 in 1000. So, Nate wasn't just off by a little bit. He was on the other side of the freaking universe.
That's not true as they are not independent events, and his model does not treat them as such. The probability that a sequence of dependent events happens is not just the product of all the probabilities, but the product of their conditional probabilities, P(X and Y) = P(X)P(Y given X). In other words, the probability that he wins Wisconsin times the probability that he wins Michigan *given that you already know he wins Wisconsin* times the probability that he wins Pennsylvania *given that you know that he wins Wisconsin and Pennsylvania* etc. If you know that Trump won Michigan and Pennsylvania, your best estimate for how likely he is to win Wisconsin as well probably shoots up by quite a lot. That's why the model ultimately came down to a 70%-ish chance, despite the odds being what they are for the individual states.
Alastair at the end of the interview keen to get Nate to say he would bet on Harris to win, but Nate being wiser than to rise to the bait, doubled down on his earlier comment that it is too close to call and he would lose a bit to the bookmaker on what is effectively a coin flip. Why is Alastair so desperate to hear good things about Harris, such as predictions from wise analysts?
Unfortunately over the past few years I've come to judge Nate Silver as a bit of a hack. He's been wrong more times than he's been right recently which is kinda bad for a political forecaster.
@SamSamSamSamSam Well 2016 and 2022 are the big ones. Basically his models are terrible at predicting major shifts in public opinion. Even if it seems obvious if you look beyond the polls and guage public sentiment it would be obvious that the weighting of the poll data needs adjusting. Problem is his models only adjust after a major error which is great at knowing past elections but not at predicting future ones.
@@JornMolt-mf6qofor real, and now he wants to play it safe by saying its a coin toss after being wrong twice in a row. So much for being part of the risk taking river. Alan Lichtman is a much better analyst.
I would have thought Rory/Alastair should interview Allan Lichtman (check out his UA-cam channel) who has a system (based on 13 keys to the Whitehouse) where he categorically names a winner so you can 100% say he was right or wrong - he did in fact predict a Trump win in 2016 and has correctly predicted the winner since Reagan. Some will say he incorrectly predicted Gore over Bush but he refutes that due to now known flaws in the Florida ballot helping Bush - not to mention Republican judges. Maybe a future interviewee?
whenever the pundits are talking about Elon Musk, they always neglect to mention his deep and very personal anti-trans sentiment that is fundamental to his radicalization as a Trumper.
Should’ve pushed back on his free speech comments regarding social media. Free speech only protects against the government limiting speech, not private companies.
Nate looks like he is drugged very well. Many people don't know this, but he is a poker player and a risk taker. Now he is trying to make it in the right wing space. Best of luck to him, but I feel bad for him because that space is full of nutjobs and he needs to turn into one in order to survive.
I think if Elon was as obsessed as Nate is about statistics he'd be a lot more accurate than Nate. Elon's just a little too busy making billions of dollars.
His predictions are totally meaningless. He could say that trumps chances are 1%, then say he was right when trump wins because he predicted it one in a hundred
Considering how obsessed the Americans are in dividing themselves up into different categories of voter types it’s no surprise that they’re so viciously divided.
The US is a huge, diverse country with 330 million people. Idk why people from countries the size of US states act surprised when the US does things differently than they do.
Unless the guest has achieved a 90% accuracy prediction since 1981, as Allan Lichtman has, based less on polling and more on historic keys about the incumbent candidate verses the challenger, then his opinions are less meaningful and do not have the same impact as someone who has correctly the predicted the outcome of the US Election 9 out of the last 10 times; Professor Allan Litchtman.
I've read Nate Silver's book and it is excellent. The reason you think he seems to be saying nothing at all is because you simply don't understand it. It takes awhile to understand the language of particular subject area. There's also nothing in the way he's speaking to give the impression that he is on drugs at all.
Alastair not understanding the concept of free speech and willing Elon Musk to be bad at poker both speak to his insecure, weak, pandering personality.
Love your channels Allister and Rory. If you want to interview a real "prediction expert" check out Dr. Allen Lichtman he has predicted EVERY presidential election since Reagan. Nate silver is, sincerely an amateur compared to this man.
13:17 I can't take someone seriously if their "analysis" uses a description of Musk that is just his carefully manufactured public image. For fun at a conference of trade economics. I use the word "fun" very loosely here I know. Anyway detailed information of prominent people's past activities and results was prepared. Without knowing the identity of the VIP, we were asked to give our opinion on these people and their choices. Then pick Dumbass, Lucky, Sharp or Smart Money or a combination. Soros got Smart Money, Bill Gates got Lucky and Sharp. Musk mostly got Dumbass and a few Lucky Dumbass. One funny guy said we should change his name to ENRON Musk.
Shouldn't it be about time the West realize the whole fight about woke-ness is the right's straw man strategy? Did we learn nothing from the critical race theory episode? It's straight out of Mao's playbook, and people don't give that monster enough credit (he is streets ahead in history's bodycount league table).
Nonsense. It’s a real threat to sanity, truthseeking, rule of law, and the ability of the country to avoid collapse or irreparable division. Some grifters and elements of the right may exaggerate specific events for their own purposes, but that doesn’t mean it’s not a real threat. Your comments about CRT and Mao support the right, and confirm that social justice ideologies can result in disasterous harm
He literally said at that point in time that Kamala had a slight edge but he's at 50/50 right now when it comes to Nov 5th. If anyone says with the current data we have that they know who's going to win this one, they're lying to you
This guy should stick to modelling data on spreadsheets because his inter-personal communication skills and his logical reasoning skills driving a coherent, lucid train of thought are shockingly poor. The utterly aghast looks on Campbell's and Stewart's faces says it all.
Thought the questioning didn't help him at all. Very superficial and generic questions without ever getting in to much depth. Was still a decent listen.
the whole of europe was and still is profoundly anti-Semitic and prejudiced against the Romani simply because they won't conform. Hunter gatherers and Farmers have always conflicted but it Europe it is knee jerk like their anti Semiticism.
He lost me when he claims that betting markets are good ways of pricing political risk. They really arent. Betting markets tend to confuse money with votes. A guy with 500 dollars is no more of a sage than a guy with 10 but betting markets assumes that higher stake means wisdom. Its why betting markets consistently get things wrong
He means it's a good way for large financial institutions to price in risk based upon their own evaluation - i.e. hedge risk. They are clearly far better placed than the average gambler to make that decision
@@avengemybreath3084 Well thats the point, betting markets are mostly wrong, which is why most punters lose on them! I will get rich but it wont be down the bookies.
YOU DO KNOW THAT NATE SILVER HAS ALREADY OFFICIALLY SOLD OUT AND IS ALL IN FOR TRUMP? I haven’t watched the show yet, but I’m surprised to see this trump apologist on it. His word has been worth nothing for at least 8 years.
I mean, he did sell out. He sold 538 to Disney then left and set up his own company Silver Bulletin, which continues to be non-partisan. You could say it's pro-Trump in the sense that it treats Trump as a legitimate viable candidate rather than as an incoherent insurrectionist felon with dictatorial tendencies, but most of the US media is doing that.
Silver mentions Labor Day, which was September 2nd. Sort of mind boggling that the hosts decided to wait nearly an entire month to release an episode that hangs on current events.
The audio version came out on 15th September, I imagine video editing and scheduling episodes meant that it had to wait till now. It's totally irrelevant anyway since most of the discussion was about the science of polling which doesn't change over the space of months.
In 2016 I was listening to the 538 podcast with Silver. Right before the election they were discussing the forecast and someone asked "so Hillary is at 70%, that means she wins right?" Nate immediately countered that 30% probabilities happen all the time. For perspective, another guest added: "Russian Roulette is only 16.6%." As I anxiously waited for results on election day I found it hard to dismiss the thought: "there are two bullets in the gun."
We had six bullets in the gun.
Silver would have been "wrong" if in every world where that election happened Hillary won. to be "correct" Trump has to win in 3 of ten scenarios. We sadly live in one of those 3 scenarios....
If you had to play Russian roulette every four years you're life expectancy isn't going to be high. This isn't sustainable.
I would say that he made the wrong prediction. I doubt that the difference between the election prediction and the outcome came from the polls being unrepresentative because of bad luck, but for systemic reasons of how the models translated polling results into a predicted outcome
UA-cam posts the upload date, but it's not really useful for time-sensitive content.
You folks post a lot of timely information. Please post the dates that these interviews were recorded..
Otherwise, thanks for all the great interviews, including this one.
The race has not quantitively changed in any significant way since arguably before the debate. It probably won't change much up to election day. All the variation in the polls has been within the margin of error. Very few people are changing their minds.
amen
Yes, one thing about a lot of UA-cam that can grind my gears.
Alistair is phenomenally wrong about what a 71% likelihood to win means. That also means Trump had a 28% chance of winning. That's a nonzero chance.
Alistair lied Iraqis died
28% is also a really high probability but the issue is that any probability besides 0%, 50% and 100% is very unintuitive to most people, and in my experience they just end up rounding to one of the intuitive 3 so when you tell someone Hillary has a 71% chance of winning they think you're saying she is going to win.
honestly i think a major issue is that these projections are often plotted the same way as polling data, and people don't separate them in their minds.
being projected to get 60% of the vote vs being projected a 60% chance to win are *very* different things.
I echo on the last comment - people consider chances if they were voting outcomes
In other words they conclude that whoever gets >50% wins
With regard to Alastair’s point about Nate being wrong in 2016, a perfectly accurate estimate of a 30% chance will be ‘wrong’ 1 in every 3 elections, and he was also getting a lot of criticism at the time for having Trump’s chances so high and not calling it a dead certainty. He doesn’t conduct polls himself so all he can do is estimate the chances based off of the existing polling data which had massive errors.
So he can never be right, and he can never be wrong.
@@Zummbot Of course he can be wrong. But you can't base it on one single event. Duh.
Should have just read this comment before I added mine above. :p
Nate is A Nonperson !
A tech bro American who confidently speak like they know everything.
Characterising UK as rule-following because our new tube lines have barriers in comparison to NYC which doesn’t have have any new lines.
And yet in the UK we don’t have jaywalking, we don’t have strict rules around women’s freedom to abortion, we didn’t implement a series of strict rules which separated the black population from whites, etc…
I’ve been observing these Tech Bro types.
There’s not much humility there amongst them as a breed.
Enough hubris to go around
This is spectacularly interesting, thank you both.
Yup, really good and quite fascinating actually, agreed thank you to all 3
why? What did the man actually say?
As far Alastair wondering about voters not thinking about what disqualifies Trump, there is a lot of factors, but alot of it comes down to a an anti-establishment realignment and the conservative media ecosystem.
The conservative media really is an ecosystem. Unlile the UK where conservatives are a lot more present in the mainstream media, conservatives in the US have just opted out of that basically entirely. There is less of a link for voters who lean right to get those contradicting perspectives. And attempts to establish those links can be contentious at best because you're poking so many holes in a thing that forms the foundation of a worldview. As a result, the right wing media ecosystem has been able to spin Jan 6 as either a non-concern or justified.
The second reason is the anti-establishment realignment which composes of people who are much harder to define politically because their views are kind all over the place, but also includes people who feel burned by the system. Regardless, a lot of the people has coalesed around Trump since so much of the political establishment has gone for Harris due to her defense of institutions.
That a thoughtful analysis, appreciated.
It’s the conservative media ecosystem. Fox News and the million social media/podcast offshoots
People think that MAGA is some kind of position or ethos, when it is actually a lifestyle choice.
There is no data, no persuasive argument, no appeal to any quality or ideal that has any impact on a lifestyle choice. I can only change from within the adherent.
Most of the time that will come too late if it is some kind of time sensitive requirement.
America was already broken and teetering in the edge of the cliff.
Russia just pushed it a little bit.
Democrats and left leaning people don't get much of right leaning prospective themselves, certainly not the best version of it anyway.
I think the anti-establishment angle of it is huge. Things have changed since January 6th, but especially before that, I think there were alot of people who voted for Trump hoping to send a message to Democrats and the establishment Republicans in general to try and force some changes and new thought.
People should consider that every person who voted for Trump hasn't gone to a Trump rally and doesn't own a red hat.
Big fan of this series and but as a minor point of order, could we have the recording date placed in the description? Sometimes you mention the date in the interview, but it makes understanding the context easier if we know this interview might be 2/3 weeks old.
Many thanks!
I recall Nate Silver saying that the 70-30 probability in 2016 equated to someone missing a 42 yard field goal in the NFL. I’m a giant fan. We beat Buffalo in the Super Bowl because their guy missed a 44 yard field goal in Miami in beautiful weather, no wind. so when I read that, I got chills up and down my spine because I know firsthand how often 4 2 yard field goals are missed.
You guys don't seem understand how Silvers' model works. His model was excellent in 2016. When he calls it 29-71 in Hillary's favor, that is a little less than one in three chance that Trump will win. In addition in the lead up to the election other models were calling it 2-98 in Hillary's favor. He's not predicting who will win. He is saying, based on historic outcomes in similar circumstances, these are the probabilities. If you can go back and read what he was writing in the lead-up to that election, he was warning people that other models were skewed in Hillary's favor. Think of it this way, if he calls ten races at a 30-70 probability, the *expected* outcome is that three of the dark horse candidates will win their races.
Right. But ultimately this is so much dancing about architecture. The statement "Today, X has a 1 in 3 chance of beating Y" is a useless statement. You either get the outcome right or you don't from "predictive models." In 2022 Silver characterized anyone who thought Democrats would win the Senate as taking "copium." He was dead wrong. We have reached the point where polls tell us next to nothing useful.
It’s a totally meaningless approach
@@Bazball1080 It's not TOTALLY meaningless... just a lot less meaningful than people think. Unfortunately, I don't know of any better methodologies out there.
@@Bazball1080 it is the only meaningful approach
It’s meaningless. Ok am calling it 60-40 Harris wins.
Loved this, thank you. Been following 538 for years.
The commentary on the Trump chances is quite strange. There’s a 1 in 3 chance of rolling a 5 or 6 on a die, but nobody is surprised when they see a five or a six. That also doesn’t mean the chances were wrong and nobody would think that.
Great when you both together sharing thoughts.
While I understand the hosts don't define their probability terms well, Nate is getting away with something dishonest here. He is quick to (rightly) dismiss the criticism of the 2016 model, but doesn't say a word when his model 'correctly predicted' the outcome in other election years. By his logic, the model was no more 'predictive' in the years when it was credited than it was in 2016. So, for his 2012 model, for instance, he should jump in and say, "actually I didn't *predict* a Barack Obama victory, I just provided a specified range of likely outcomes." But he doesn't say that. He takes credit for predictiveness when it suits him, but denies that's his aim when it doesn't.
Great Interview
I loved Nate Silver as a guest. He had such an interesting (risk, statistics) view of the world. Fascinating.
A very late thank you to Mr. Silver. His books are excellent reads.
Shapiro has baggage that would have blown up in the campaigns face. Walz is the safer pick.
But Walz only boomed his own state by .5% in favor of Harris
Not strong
doubtful Shapiro would have done ditto for PA. Running mates rarely make a dent, they just need to be a useful surrogate who doesn't actively damage the campaign and drag on the ticket
@@packersmockdraftsimulator4781WTF does "boomed his own state" mean?
@@JeanValjean875 bombed actually
A .5% boost in their state is a joke
Is this current? Doesn’t seem so. It’s been many weeks since Labor Day.
Can you have on Simon Rosenberg and Sarah Longwell? They are both brilliant and so insightful about the Nov election.
There is a fairly wide consensus that the choice of vice-presidential running mate makes little difference to the outcome of the election, except perhaps in bizarre cases. Doesn't that suggest that a presidential candidate ought to nominate a VP candidate on the grounds of how effective he will be in office and whom she wants to work with for five years. Perhaps the issue is not "Josh Shapiro might have delivered a vital few thousand votes in Pennsylvania", but rather "Tim Walz will be a good VP for Kamala Harris, if the electoral system chooses them".
“The office of the Vice President doesn’t do that much.” - Nancy Pelosi (2024)
@@Dan-y6l3k The VP has very little constitutional authority. On the other hand, a VP who is the trusted delegate of the President (such as Dick Cheney or Al Gore) can swing quite a big stick, can act as a Cabinet-level right-hand-man and minister-without-portfolio.
Who better to have beside you at a Cabinet meeting etc.? Someone who thinks you should be President, or someone who thinks he should be President.
@@brettevill9055 Wonder what the odds are of a POTUS dying in office are. Might make the VP pick absolutely critical. Which one would you want takin' the reins?
Have I got this right? I am right when I am wrong, and I am wrong when I am right. I was not wrong because I got closer than I predicted so that makes me right. ???.
What did Churchill say about stats?
His description of the UK as a place where the Media is thriving and Gambling isn't an issue? Gell-Mann Amnesia effect emanating strongly off this guy.
There are a lot of people registering to vote that have never voted before. Pollsters use lists of likely voters. So it will be interesting to see if they vote and who they vote for.
So enticing. There was so much in that interview that should have been explored so much deeper. But alas ... . Also this is a wonderful demonstration of why not being statistically numerate is such a big, and growing, disadvantage. There will be AI soon that will be able to more reliably predict elections based on each and every click, world wide. Scary.
I think it’s important when looking at polling, and as a Democrat it worries me, the resilience of Trump voters
They were deeply underestimated in 2016 which is why people like Silver got that election wrong. And in 2020, they were again underestimated and Trump barely lost states like Wisconsin by a few thousand votes where Biden was supposed to win by 3 or 4 points
That said, undecided voters are more about whether or not those people will vote. So get your ass out to vote.
One of the bullish cases that can be pointed to is that in basically all special elections in this past cycle, democrats have massively outperformed polls due to turnout caused by issues like abortion in particular, but the caveat is we don't know if this polling error will translate to general election polling where turnout is more likely in general and people are more motivated to come out for issues beyond abortion
Trump is at his ceiling, which must be lower than 2020 if the insurrection and convictions have had even a slight effect on non-crazies. If Harris enthuses her coalition to turn out in the same scale as they did for Biden in 2020, we will see the same result. And she seems a lot more appealing to young, female and black voters than Biden was. It's going to come down to turnout. 2016 Dems stayed home and Trump won. 2020 Dems showed up and Trump lost.
I think there are a lot of people like my parents, conservative Christians who won’t vote for him after January 6. The Stormy Daniels thing was so cringe too. I guarantee they would not have voted for him had they known.
I’m not clear why making Josh Shapiro the Vice-Presidential pick would have been a better choice. If Pennsylvania is critical, why would taking its very popular Governor out and sending him to Arizona, Nevada, North Carolina and New Hampshire be smart? Keeping Shapiro campaigning in PA is clearly the far better move.
plus hes a loose cannon
@@DGE123Obama killed the shapiro option. A couple of stories were released in the press. He served in the idf, not the US army, plus he tortured Palestinians in prisons. The second is an article he wrote in college, very anti Arabs, critical of Ytsak Rabin for the Oslo process, trying to make peace with the Palestinians. Obama feared others papers would come out.
I love these interviews - but in this fast paced election season, please date them - this was obviously a pre debate interview?
Strange thumbnail when he said in the same sentence that waltz was a mistake too.
They are biased unfortunately
It appears that hatred for one candidate influences objectivity.
Great discussion.
Nate has an each way bet on every election.
He never really makes a call.
Just listen to Professor Lichtmann He puts it on the line.
They should have got Allan Lichtman on. He has a far better track record of predicting outcomes (9 out of 10) with his 'Keys to the White House' model.
Gore won. The supreme court stopped the counting in FL, very unconstitutional. They happened to be black votes too. When all was said and done after the counting, Gore won FL
@@caroleeb1997 actually once they counted the votes Gore won - handily
Too bad this is already outdated
How so?
@@mandel94 Nate silver referred to the time in which they were speaking and said it was between the DNC and the debate.
@@VirginiaBronson Has much really changed since lol
@@SamSamSamSamSam quite a few things, yes. In terms of his model output? Also yes.
It's all still relevant
how old is this?
Right? Silver mentions Labor Day, which was September 2nd. His prediction of 50/50 is reflective of a Kamala polling slump from that time which has been completely erased.
7 minutes in and I’ve already heard the terms ‘happy accident’ and ‘telephone penetration’…. god bless America 😬
You saying that a 30-odd percent chance is ‘you predicted Trump would lose’ is like if someone tosses two coins and they come up both heads, and you respond ‘but you said this wouldn’t happen!’ (When in reality you predicted a 25% chance of this happening).
No one would do that, but because it’s politics we expect probability to mean different things?
It's always interesting to me how many Brits and Europeans mock or downplay when Americans try and claim some connection or brotherhood to Europe via heritage, tradition, etc. But then also act so confused that politics and society in general in the US dont do things just like their country. Almsot like the US is just a different place. I also often find that the same people are very good at practicing cultural relativism with various other cultures and countries but seem offended that the US has different ideas and ways of doing things lol.
Just a plurality of votes would give urban voters way too much influence. Especially when urbanization & market economics are no longer a determinant in 21st Century unlike the last century. Electoral College System is a check and balance that forces the 2 parties to appeal to a wide variety of voters outside of their own stifling echo chambers. Democrats need to appeal to rural voters while the GOP needs to appeal to suburb & exurb voters.
Excellent point, in Canada a very bad man called Justin keeps winning with only a couple of major cities and is on record as disregarding vast rural and the entire western part of the country and its a majorly divisive and is why all these left woke encourage massive immigration and migration to major cities
The problem with Nate's defense of the 2016 model is it starts to break down when you look at the individual states. Nate projected Hillary was favored to win all of the following states:
- 83.5% chance to win in Wisconsin
- 78.9% in Michigan
- 77% in Pennsylvania
- 55.5% in North Carolina
- 55.1% in Florida
Trump won ALL of these states. According to Nate's model, the chance of that happening is around 1 in 1000. So, Nate wasn't just off by a little bit. He was on the other side of the freaking universe.
That's not true as they are not independent events, and his model does not treat them as such. The probability that a sequence of dependent events happens is not just the product of all the probabilities, but the product of their conditional probabilities, P(X and Y) = P(X)P(Y given X). In other words, the probability that he wins Wisconsin times the probability that he wins Michigan *given that you already know he wins Wisconsin* times the probability that he wins Pennsylvania *given that you know that he wins Wisconsin and Pennsylvania* etc. If you know that Trump won Michigan and Pennsylvania, your best estimate for how likely he is to win Wisconsin as well probably shoots up by quite a lot. That's why the model ultimately came down to a 70%-ish chance, despite the odds being what they are for the individual states.
35:29 Is he being tortured?
Alastair at the end of the interview keen to get Nate to say he would bet on Harris to win, but Nate being wiser than to rise to the bait, doubled down on his earlier comment that it is too close to call and he would lose a bit to the bookmaker on what is effectively a coin flip. Why is Alastair so desperate to hear good things about Harris, such as predictions from wise analysts?
October 17th 2012 Romney was ahead in a Gallup poll by 6%. He lost.
Unfortunately over the past few years I've come to judge Nate Silver as a bit of a hack. He's been wrong more times than he's been right recently which is kinda bad for a political forecaster.
Any specifics? What has he gotten wrong?
@SamSamSamSamSam Well 2016 and 2022 are the big ones.
Basically his models are terrible at predicting major shifts in public opinion.
Even if it seems obvious if you look beyond the polls and guage public sentiment it would be obvious that the weighting of the poll data needs adjusting. Problem is his models only adjust after a major error which is great at knowing past elections but not at predicting future ones.
@@JornMolt-mf6qofor real, and now he wants to play it safe by saying its a coin toss after being wrong twice in a row. So much for being part of the risk taking river. Alan Lichtman is a much better analyst.
On the NYT 2016 election night needle: Rory, wrong Nate!!
These are some really basic questions. With Nate Silver you could have really dug into the details of the horse race.
I would have thought Rory/Alastair should interview Allan Lichtman (check out his UA-cam channel) who has a system (based on 13 keys to the Whitehouse) where he categorically names a winner so you can 100% say he was right or wrong - he did in fact predict a Trump win in 2016 and has correctly predicted the winner since Reagan. Some will say he incorrectly predicted Gore over Bush but he refutes that due to now known flaws in the Florida ballot helping Bush - not to mention Republican judges. Maybe a future interviewee?
hahah the 2000 election was obviously stolen. I don't think that's even debatable
whenever the pundits are talking about Elon Musk, they always neglect to mention his deep and very personal anti-trans sentiment that is fundamental to his radicalization as a Trumper.
Yeah, I think it links to his obsession with [white] birth rates lol
My money was on RFK, Jr. was going to win California!!! Blast!!!
Love you guys but you need to study statistics and probability a bit better. Might i suggest nate silvers new book?
….aaaand then you have the Electoral College and everything goes berserk.
Disappointing how Alastair doesn't understand the threat from the wokeness and the anti-free speech movements in the US and the UK.
No leftist does
what a deranged convo
I found him a bit hard to follow by the way he spoke. Also, I wished he took off his hat, hard to see his eyes. 😐
Why does Rory make such weird faces all the time, really distracting lol
Shapiro is really more interested in being president than vice president.
How tiresome are these loud overconfident bros... so at odds with the vibe of the podcast
Rory's keeps doing that fake smile looks strange
Should’ve pushed back on his free speech comments regarding social media. Free speech only protects against the government limiting speech, not private companies.
Has the government or government agencies ever pressured or attempted to influence private companies' policies on speech?
this must be 3 weeks old really, so misleading that it has just come out, and the predictions could well be out of date.
These guys are as villagey as they get.
Nate looks like he is drugged very well. Many people don't know this, but he is a poker player and a risk taker. Now he is trying to make it in the right wing space. Best of luck to him, but I feel bad for him because that space is full of nutjobs and he needs to turn into one in order to survive.
Nate is cool but I have a whole lot more admiration for Elon. Seems petty for Nate to criticize him.
why?
I think if Elon was as obsessed as Nate is about statistics he'd be a lot more accurate than Nate. Elon's just a little too busy making billions of dollars.
elon does fuck all
@@N13TheGeneralcope
Gosh, that was really boring
His predictions are totally meaningless. He could say that trumps chances are 1%, then say he was right when trump wins because he predicted it one in a hundred
Agreed
Then it's always only really 50/50. He wins, or he doesn't.
He is doing statistical analysis based on polling data.
Considering how obsessed the Americans are in dividing themselves up into different categories of voter types it’s no surprise that they’re so viciously divided.
The US is a huge, diverse country with 330 million people. Idk why people from countries the size of US states act surprised when the US does things differently than they do.
Unless the guest has achieved a 90% accuracy prediction since 1981, as Allan Lichtman has, based less on polling and more on historic keys about the incumbent candidate verses the challenger, then his opinions are less meaningful and do not have the same impact as someone who has correctly the predicted the outcome of the US Election 9 out of the last 10 times; Professor Allan Litchtman.
This guys talks fast but seems to be saying almost nothing at all. It's mostly gibberish and none of it answering the questions. Was he on drugs?
I've read Nate Silver's book and it is excellent. The reason you think he seems to be saying nothing at all is because you simply don't understand it. It takes awhile to understand the language of particular subject area. There's also nothing in the way he's speaking to give the impression that he is on drugs at all.
This was thoroughly boring
Nate Silver does not know anything about the US elections
Why’s he talking so fast?
Alastair not understanding the concept of free speech and willing Elon Musk to be bad at poker both speak to his insecure, weak, pandering personality.
Your God Musk gives 0 F’s about you 😂 off you pop little one!
What is wrong with that guy’s face? 😂
Love your channels Allister and Rory. If you want to interview a real "prediction expert" check out Dr. Allen Lichtman he has predicted EVERY presidential election since Reagan. Nate silver is, sincerely an amateur compared to this man.
Maybe I’m not bright enough to follow what this guy says but man….!
You’re good, he’s not well read enough about society and politics to say anything intelligible about them
Complicated maths guy who made a lot of money explains probabilities. Jesus this election just needs to happen already.
13:17 I can't take someone seriously if their "analysis" uses a description of Musk that is just his carefully manufactured public image. For fun at a conference of trade economics. I use the word "fun" very loosely here I know. Anyway detailed information of prominent people's past activities and results was prepared. Without knowing the identity of the VIP, we were asked to give our opinion on these people and their choices. Then pick Dumbass, Lucky, Sharp or Smart Money or a combination. Soros got Smart Money, Bill Gates got Lucky and Sharp. Musk mostly got Dumbass and a few Lucky Dumbass. One funny guy said we should change his name to ENRON Musk.
Shouldn't it be about time the West realize the whole fight about woke-ness is the right's straw man strategy? Did we learn nothing from the critical race theory episode? It's straight out of Mao's playbook, and people don't give that monster enough credit (he is streets ahead in history's bodycount league table).
Nonsense. It’s a real threat to sanity, truthseeking, rule of law, and the ability of the country to avoid collapse or irreparable division. Some grifters and elements of the right may exaggerate specific events for their own purposes, but that doesn’t mean it’s not a real threat. Your comments about CRT and Mao support the right, and confirm that social justice ideologies can result in disasterous harm
White men always seem to get “wokeness” wrong Nate seems really level headed I would love him to sit with his concept of wokeness a little more
“White men” aren’t one group. What a weird comment
The irony of your raycist statement.
This is the perfect encapsulation of why people hate "wokeness". The irony
Tell you what folks, Nate Dog reckons Trumps going to win.
He literally said at that point in time that Kamala had a slight edge but he's at 50/50 right now when it comes to Nov 5th. If anyone says with the current data we have that they know who's going to win this one, they're lying to you
@@doughvahkiin2885 There never was a scenario where Trump was going to win. Polls won't tell you that. See: Allan Lichtman.
@@doughvahkiin2885 yes but he thinks trump will win
This guy should stick to modelling data on spreadsheets because his inter-personal communication skills and his logical reasoning skills driving a coherent, lucid train of thought are shockingly poor. The utterly aghast looks on Campbell's and Stewart's faces says it all.
He’s a pollster. Why are you whining like a child that he’s not an expert communicator? He’s a data analyst you wally.
Thought the questioning didn't help him at all. Very superficial and generic questions without ever getting in to much depth. Was still a decent listen.
The French Enlightenment was profoundly antisemitic.
the whole of europe was and still is profoundly anti-Semitic and prejudiced against the Romani simply because they won't conform. Hunter gatherers and Farmers have always conflicted but it Europe it is knee jerk like their anti Semiticism.
How? They emancipated the Jews in 1791.
what a typical american
yeah, the yanks all seem to be genius pollster
Stay mad while your country sinks into irrelevance.
@@chickenfishhybrid44 I’m not even British, cry more
:OOOO
He lost me when he claims that betting markets are good ways of pricing political risk. They really arent. Betting markets tend to confuse money with votes. A guy with 500 dollars is no more of a sage than a guy with 10 but betting markets assumes that higher stake means wisdom. Its why betting markets consistently get things wrong
He means it's a good way for large financial institutions to price in risk based upon their own evaluation - i.e. hedge risk. They are clearly far better placed than the average gambler to make that decision
@@harrysuffolk They certainly do price and hedge risk. But they dont use betting markets to do that
Then go get rich!
@@avengemybreath3084 Well thats the point, betting markets are mostly wrong, which is why most punters lose on them! I will get rich but it wont be down the bookies.
What a shifty interview. Total word salad BS.
YOU DO KNOW THAT NATE SILVER HAS ALREADY OFFICIALLY SOLD OUT AND IS ALL IN FOR TRUMP? I haven’t watched the show yet, but I’m surprised to see this trump apologist on it. His word has been worth nothing for at least 8 years.
No, I didn’t know that. And I still don’t.
He literally said that he would be voting for Harris and that Jan 6th is disqualifying and means that he would never vote for Trump...
@@phueal Amazingly, I was unpersuaded even though it was said in all caps.
William is Wrong.
I mean, he did sell out. He sold 538 to Disney then left and set up his own company Silver Bulletin, which continues to be non-partisan. You could say it's pro-Trump in the sense that it treats Trump as a legitimate viable candidate rather than as an incoherent insurrectionist felon with dictatorial tendencies, but most of the US media is doing that.
18:33 its an update to how the right used the word Communism.
So this interview was before the debate? It would be good if there was a date somewhere - 22 hours ago, it clearly was not.
Silver mentions Labor Day, which was September 2nd. Sort of mind boggling that the hosts decided to wait nearly an entire month to release an episode that hangs on current events.
The audio version came out on 15th September, I imagine video editing and scheduling episodes meant that it had to wait till now. It's totally irrelevant anyway since most of the discussion was about the science of polling which doesn't change over the space of months.