Discussion for article #227782
I guess we’ll know who’s right and wrong in 6 weeks.
11 Likes
Alt headline:
Nate Silver Critiques Wang
14 Likes
Wasn’t Sam Wang more accurate on the Senate races last time around than Silver?
9 Likes
No. Can’t know right or wrong by looking at a model correlation to one election result.
3 Likes
True. I thought about that right after I posted it.
Nate has my vote for accuracy after 2012’s predictions.
3 Likes
Polls today are not any better than they were in 2012. Saying that, you still have to go with the polling averages of at least the last three polls done in the same timeframe. Even if the polls are widely different and pollsters methodology and weight of the poll is way off that is the only way to get a read on what is going on. I don’t think that generic polling is a decider on who will come out to vote. It might give a single picture in one timeframe. The reason why 2010 was so different was because democrats and republicans were both mad about Obamacare for totally different reasons. Democrats stayed home because they were mad and republicans went out to vote because they were mad. I just don’t see the same dynamic for 2012.
1 Like
He’s got my vote for accuracy in POTUS election years. I haven’t paid enough attention to predictors’ forecasts in midterm years to vouch for anyone’s accuracy. Like the first responder said,we’ll know in 6 weeks who is right or wrong.
2 Likes
Nerdfight!
12 Likes
So, Nate’s getting testy as he fades into obscurity and misses the limelight.
How good was Nate in 2012 on the Senate races? Nothing like the big enchilada as I recall.
5 Likes
Fame doesn’t sit as easily as it once did…
3 Likes
What storm said. When you say there is an 80 percent chance of something happening, it means that there is a 20 percent chance of it not happening, or that 8 times out of every 10 that the odds line up that way, it will happen, and 2 times out of every 10 it will not. If it doesn’t happen, it doesn’t mean the odds were wrong, it just means that in that particular instance, it was the low probability possibility that occurred. You can only say that someone is “wrong” if the outcomes they assign high probability to don’t come out that way as often as the odds say they should.
So that Wang got all of the Senate forecasts right in 2012, while Silver missed two that were close races doesn’t mean Wang was right and Silver was wrong, it just means that Wang was a little luckier that particular time.
As much as I would like to believe Wang’s projection, I find it hard to believe that Democrats at this point have a 70 percent chance of winning the Senate. Their path to that result is a lot narrower than the path the Republicans have to a Senate majority. Doesn’t mean the Democrats can’t do it, but pretty much everything has to go right for them for that to happen.
6 Likes
I don’t know about that - I think polling consistently under represents dems, so I think there will be some surprises on election night. And I think the odds of us keeping the senate are a lot higher than the “everything has to go right for us” scenario.
9 Likes
I am less concerned with the model predictions at this point than I am with how each model tracks momentum. Currently both models are moving towards the Democrats in the Senate.
11 Likes
“…That model is wrong…”
Relative to reality, all models are wrong.
It’s coincidence if any effectively predict numbers, because no model can predict actual human behavior – especially not weeks from now – any better than meteorologists can predict next weekend’s weather pattern.
1 Like
You are arguing that Nate is right.
Nate’s is saying that the error bars (Nate’s are much wider) more closely match his model than Wang’s.
NERD FIGHT!!! In the science lab News at 11
1 Like
wang was right in 2012 he didn’t miss anything unlike Silver
2 Likes
Crowns neither.
1 Like