I will happily accept that Wang and Silver have performed equally well. As far as Wang having performed better, given the small sample size of races where they disagreed, I greatly doubt he could establish superiority at any statistical level of high confidence.
1 Like
I feel like I am enjoying this way too much
3 Likes
If they have performed equally well, then Wangâs models are better, because they are simpler. In science itâs fundamental that we employ the simplest among a set of models that explain (or predict) with equal power.
Wang understands this. Silver either doesnât understand it, or is blowing smoke.
7 Likes
Fair enough. But Wang, as a scientist, should not claim his model performs better unless he can establish that statistically, and I doubt he can.
SCOtUS Uphoalds GAY maRRY. SiLVer HAtes WAng, LoveS wANg. TPM goes GAy.
Well, personally speaking I am gay in the sense of happyâŚ
1 Like
I think this spat highlights the difference between an intellectual (Wang) and a nerd (Silver).
1 Like
Nate Silver is an entertainer. Prof. Wang got into this to help activists more effectively campaign. The reasons to do meta poll analysis is to get more accurate information on where you need to spend resources, and to entertain.
In light of this, the question of who is right is not who guessed the day before the election. Itâs more complicated. If your goal is to win elections, itâs which model directed resources best weeks before the election. If your goal is to entertain, then it is who gets the most hits.
2 Likes
Speaking as someone whoâs been a personal friend of Sam Wangâs for twenty years, I can say with confidence that Sam is genuinely all about the data. Sure, he thinks his model is better than Silverâs. However, if heâs proved wrong, heâll check, adjust, and move on.
6 Likes
I think thatâs inaccurate. His model has performed better. This better performance may be due simply to random chance, and yes the sample set is too small to determine that. But it has, factually, performed better thus far and that remains as evidence in his methodological favor, albeit quite inconclusive.
1 Like
âPast performance is the best indicator of future resultsâ
Wang wins.
One or the other needs to pull out their pocket protector and slap the other across the face, and challenge them to a no holds barred, winner take all game of Badminton.
In my opinion, Wangâs explanation is the most probable. The problem I have with the special sauce approach, is that is should already or will be soon, baked into the polling numbers.
I am still not, nor have I all season, seen a net 6 states where republicans hold a polling lead. WV, MT, and SDâŚthats it. They are tied in LA, GA, AK, AR and probably IA. They have a lead in KY. They are behind in every other battleground, and now we can throw KS into that pile as well. That means they have to sweep LA, AK and AR from us, while not losing GA.
I have a hard time giving them 50% odds at doing that.
2 Likes
I hate to quibble (actually I donât) but if your model is not closer to the correct answer by some statistical test of significance then you simply cannot say itâs better. If you test a drug vs placebo and more survive with drug but the p value is 0.2, no journal would accept your saying that the drug was better than placebo.
All the data says is that Sam and Nate are both better than pundits throwing darts. It doesnât allow one to say one is better than the other. As the data set grows large enough it may, but I donât think weâre there yet.
âMy Wang is bigger than yoursâŚ?â
Take this outside.
1 Like
Speaking of poll controversies, todayâs SurveyUSA has Alison Lundergan Grimes up on Senator Yertle (R-Turtle), by a margin of 46-44.
Prepare to eat it, cons.
Now, if only Edward Snowden would weigh in on this, weâd have a TPM confluence of awesomeness.
1 Like
sorry, meant to reply to the group, not to you specifically. Love, love, love this new comment system.
It seems to me that Silver is denying the facts and arguing out of the imaginary.
He disagrees with Wangâs approach/model or whatever it is that has him all riled up, but then there are results. And denying them makes him seem petty or like he is stomping his feet and saying that I am better because I say I am.
Silver either canât handle the competition or is intentionally making this a bigger deal than it is just for the public exposure? Maybe they are both playing us and have cooked up their own Revenge of the Nerds scenario? You have to be able to think like a nerd and get into their skin to figure that out and I not only canât but donât have any desire to.
As brainy as they both are, they should be able to work this out and move forward with the objective of providing extremely accurate, useful information. But, brainy and all, they are merely human and have sads just like everyone.
Take both of their models with a grain of salt, throw them in a blender and enjoy the mix that comes out. Thatâs my input in the process sans emotion.
1 Like
Okay, I have to askâŚis anyone else picturing this video when reading about this latest tiff?
3 Likes