oblong Posted April 24 Share Posted April 24 I think if you are searching enough for a data point to confirm a theory then you can find it in a primary election like this with so many unusual variables. Two incumbents from the last election, a third party, a passionate issue like abortion, a clear divide on one side with a since departed candidate, early voting.... You can find what you need but I'm not convinced that's reliable. The theory is sound but I don't think these things necessarily show it. That's my best Nate Silver impersonation. Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 The Nates have really gone downhill. Silver probably should have just stuck with statistics instead of punditry. Quote Link to comment Share on other sites More sharing options...
romad1 Posted April 24 Share Posted April 24 15 minutes ago, oblong said: I think if you are searching enough for a data point to confirm a theory then you can find it in a primary election like this with so many unusual variables. Two incumbents from the last election, a third party, a passionate issue like abortion, a clear divide on one side with a since departed candidate, early voting.... You can find what you need but I'm not convinced that's reliable. The theory is sound but I don't think these things necessarily show it. That's my best Nate Silver impersonation. I get what you saying. Any weird data can be dismissed as an anomaly. Any compounding of weird anomalous data can be dismissed as bad faith propaganda. That said, there is something in the numbers that supports aversion to Trump and to an apparent lesser extent, aversion to Biden in the voting population. Quote Link to comment Share on other sites More sharing options...
pfife Posted April 24 Author Share Posted April 24 (edited) I think Nate Silver either could not or chose not to evolve with the field of data analysis. He gained prominence in a time period when data science was in it's infancy and by data science I mean the application of AI to data analysis which had been the industry standard before then. Now, there are free Python libraries that can automatically produce thousands and thousands of models like Nate made, test them all, rank them, add remove/remove all sorts of variables, automatically test which regression models best fit the data (poisson, LSS, etc) all automatically. Plus why wouldn't anyone just chose to be a pundit given that choice? Punditry is one of the least accountable, most upwardly failing jobs you can get. If he's known for predicting elections otherwise, you can get those wrong and look uncredible in your primary goal. As a pundit you might get a bunch of "this you" tweets when you're wrong bfd people don't look to pundits for right and wrong. they look to pundits for reassurance. Edited April 24 by pfife 1 Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 Is Nate still on about Biden being old and needing to be visible on the campaign trail? Is he still quiet about Trump falling asleep in court? Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 The good thing with Nate is that he is never wrong. He predicted Hillary had an 80% chance to win and fell back on well Trump still had a 20% chance. What's the point? It's never going to be 100%. Quote Link to comment Share on other sites More sharing options...
pfife Posted April 24 Author Share Posted April 24 Also, I think he's wise to not stake his career on averaging polls given the time tested principle of GIGO. Quote Link to comment Share on other sites More sharing options...
romad1 Posted April 24 Share Posted April 24 1 minute ago, pfife said: Also, I think he's wise to not stake his career on averaging polls given the time tested principle of GIGO. garbage polls are definitely in the mix. Quote Link to comment Share on other sites More sharing options...
oblong Posted April 24 Share Posted April 24 3 minutes ago, pfife said: I think Nate Silver either could not or chose not to evolve with the field of data analysis. He gained prominence in a time period when data science was in it's infancy and by data science I mean the application of AI to data analysis which had been the industry standard before then. Now, there are free Python libraries that can automatically produce thousands and thousands of models like Nate made, test them all, rank them, add remove/remove all sorts of variables, automatically test which regression models best fit the data (poisson, LSS, etc) all automatically. Plus why wouldn't anyone just chose to be a pundit given that choice? Punditry is one of the least accountable, most upwardly failing jobs you can get. If he's known for predicting elections otherwise, you can get those wrong and look uncredible in your primary goal. As a pundit you might get a bunch of "this you" tweets when you're wrong bfd people don't look to pundits for right and wrong. they look to pundits for reassurance. and he qualifies his statements in such an insfufferable way that he's technically never wrong. I look for pundits to clear out the weeds for me and do the dirty work and give me the straight dope as best they can. It's ok to just come out and say "This is what I think will happen based on....." and if you are wrong, so be it. we can all see it. But don't try to hedge your bets and look for cover. Be a man. Own it. Quote Link to comment Share on other sites More sharing options...
gehringer_2 Posted April 24 Share Posted April 24 2 minutes ago, pfife said: Plus why wouldn't anyone just chose to be a pundit given that choice? Of course the key is being given the choice. You do generally have to have something interesting to say to keep getting that choice. What gave people an interest in what he had to say was the belief that what he said was informed by reliable data analysis. Lose that and IDK how long anyone cares what he has to say! Conventionally at least, what validates most pundits is experience working inside the system, e.g. ex-elected official or having worked inside the system as campaign or admin/Congressional staff or a long time political journalist. Without numbers validation Nate lacks any of that. Quote Link to comment Share on other sites More sharing options...
Tiger337 Posted April 24 Share Posted April 24 3 minutes ago, Motown Bombers said: The good thing with Nate is that he is never wrong. He predicted Hillary had an 80% chance to win and fell back on well Trump still had a 20% chance. What's the point? It's never going to be 100%. 80% is not a particularly high number. If a .200 hitter comes up in the 9th inning and needs a hit to win the game, it's probably not going to happen, but there is a decent chance it will. The people who were wrong were the many who said Trump had no chance to win based on that 80%. I'm not making excuses for Silver. I think his time has passed. He was a very influential man in the field for a number of years, but isn't so relevant anymore. Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 Just now, Tiger337 said: 80% is not a particularly high number. If a .200 hitter comes up in the 9th inning and needs a hit to win the game, it's probably not going to happen, but there is a decent chance it will. The people who were wrong were the many who said Trump had no chance to win based on that 80%. I'm not making excuses for Silver. I think his time has passed. He was a very influential man in the field for a number of years, but isn't so relevant anymore. Basically Nate Silver is the pitcher who gave up the hit that ended Chris Davis hitless streak. 1 Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 Was it Nate Silver who came up with that insufferable liberal alter ego or was that the other Nate? Quote Link to comment Share on other sites More sharing options...
gehringer_2 Posted April 24 Share Posted April 24 9 minutes ago, pfife said: Also, I think he's wise to not stake his career on averaging polls given the time tested principle of GIGO. This is the key. The signal has to be in the available data - if it's not reliably there, no amount analysis is going to help. Learning how to survey America without landlines is still in the Blind men encountering the elephant phase. Quote Link to comment Share on other sites More sharing options...
Tiger337 Posted April 24 Share Posted April 24 (edited) 10 minutes ago, oblong said: and he qualifies his statements in such an insfufferable way that he's technically never wrong. I look for pundits to clear out the weeds for me and do the dirty work and give me the straight dope as best they can. It's ok to just come out and say "This is what I think will happen based on....." and if you are wrong, so be it. we can all see it. But don't try to hedge your bets and look for cover. Be a man. Own it. That is how statisticians talk. Never wrong and never right! I don't think he has the personality to be a pundit. He probably should have further developed in the field and stayed in the background rather than trying to be a pundit. Edited April 24 by Tiger337 Quote Link to comment Share on other sites More sharing options...
Tiger337 Posted April 24 Share Posted April 24 3 minutes ago, Motown Bombers said: Basically Nate Silver is the pitcher who gave up the hit that ended Chris Davis hitless streak. Not really understanding that analogy. Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 1 minute ago, Tiger337 said: Not really understanding that analogy. No I get it but Nate is suppose to the genius and he got 2016 wrong. If he's never wrong, why bother with him? Quote Link to comment Share on other sites More sharing options...
pfife Posted April 24 Author Share Posted April 24 3 minutes ago, gehringer_2 said: This is the key. The signal has to be in the available data - if it's not reliably there, no amount analysis is going to help. Learning how to survey America without landlines is still in the Blind men encountering the elephant phase. I also wonder what it was like when you're trying to filter bad polls out, or weigh them very low, and there's bad actors out there intentionally conducting bad polls and others can try to battle you with that fake crap. That's not to mention that clearly even good faith pollsters are really struggling with sampling these days. Quote Link to comment Share on other sites More sharing options...
Tiger337 Posted April 24 Share Posted April 24 7 minutes ago, Motown Bombers said: No I get it but Nate is suppose to the genius and he got 2016 wrong. If he's never wrong, why bother with him? Probability is not about wins and losses. It's not binary. Quote Link to comment Share on other sites More sharing options...
gehringer_2 Posted April 24 Share Posted April 24 (edited) 12 minutes ago, pfife said: That's not to mention that clearly even good faith pollsters are really struggling with sampling these days. Absolutely. I get a little irritated when you see a poll reported as "1500 likely voters" and the error bar is ususally given the binomial uncertainty for the sample size. That's BS because the sample population is not random, it was balanced out of a larger raw number by the pollster and no-one is giving an error bar on the accuracy of the sample selection. That's where the rubber meets the road in today's environment. Edited April 24 by gehringer_2 Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 2 minutes ago, Tiger337 said: Probability is not about wins and losses. It's not binary. I get it, he's not wrong. You tell me something has a 80% chance of happening and it doesn't and you're still not wrong. I can say right now Biden has a 99% chance of winning and not be wrong. Nate was supposed to be the expert since he correctly predicted every election up until then. Quote Link to comment Share on other sites More sharing options...
Sports_Freak Posted April 24 Share Posted April 24 Donald Trump has been complaining he can't campaign because he has to be in court. Since there's no court today, where is his rally going to be at? Or is he at home, sleeping in? 😁 Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 Here's Nate having a condescending conversation with himself where the alter ego he created to mock Democrats turned out to be more correct than he was. The Case For A Democratic Surprise On Election Night | FiveThirtyEight Quote Link to comment Share on other sites More sharing options...
Tiger337 Posted April 24 Share Posted April 24 9 minutes ago, Motown Bombers said: I get it, he's not wrong. You tell me something has a 80% chance of happening and it doesn't and you're still not wrong. I can say right now Biden has a 99% chance of winning and not be wrong. Nate was supposed to be the expert since he correctly predicted every election up until then. Saying that Biden has a 99% chance of winning would be really optimistic at this point! Quote Link to comment Share on other sites More sharing options...
Motown Bombers Posted April 24 Share Posted April 24 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.