what's your opinion of 538? do we like it?
No
Nate is cuck
>>139762150
> run by a kike faggot
> models based on (((polls)))
> commentary more kiked than models
> wrong about everything
We love it.
>>139762313
>looks like a typical furry
>logo is a fox
Makes a lot of sense.
>>139762150
>do your best to sort through collected data to construct accurate models
>end up failing to construct accurate models because of extreme bias
>stand by your admittedly biased models anyway
As much as we like to rag on Nate Copper, 538 was one of the few sites that predicted Trump stood a small chance of winning. Everyone else was just "YASSSSSS KWEEN HILLARY SLAY UR GONNA WIN 100%"
>>139762150
no
wtf is wrong with you?
?
>>139764753
this
>>139762150
If you actually read Nate's writing you'll find that he hates MSM political reporting more than most people here. Also every pundit and their mother made fun of them for predicting Trump with 30%+ odds of victory or higher. Turns out the uncertainty of his model was spot on, even though the vast majority of polls had HRC 'ahead'.
Their presidential model is legitimately solid, but my expectation is that there will be even less state polls in 2020 for them to work with. They may need to either start commissioning their own polls, or find a new value proposition. Or just die off.
Off election season, their articles are okay but not particularly insightful. Their sports stuff is alright I guess.
>>139762150
>>139762313
To be fair they did give Trump better odds than most, and defended themselves when other outlets gave them shit for it, saying that the other guys were underestimating Trump.
>>139764753
>Trump had a small chance of winning
It was never in doubt.
>>139765048
kys faggot
>>139764003
He did end up getting me some very good odds on Trump. Got to give him that.
>>139765361
If he wins by less than 100,000 votes overall, then yeah, it was in doubt.
>>139765367
Suicide pact of interest to you? I'd be willing to go if I can take you with me.
>>139765048
junk in, junk out
no matter how good it is it's garbage, because they keep insisting on feeding it garbage
Their articles are boring as expected from pundits who try to be "neutral"
I like their poll models though
>>139766639
I dunno... PEC's model was supposedly better because they focused only on 'quality' pollsters and Sam Wang literally had to eat bugs as a result. 1% odds... he should have had to eat human feces.
Some of the pollsters that people crapped all over 538 for including - Google Consumer Surveys, YouGov, LA Times Daybreak - ended up being pretty helpful in terms of counteracting some of the assumed 'quality' pollsters that were fucking up. The value of the model is adjusting for those assumed biases in the first place. And recognizing that if every poll is very close, then it only takes a small polling error to increase the probability of uncertainty, regardless of the direction that the majority points toward.
Pretending polling doesn't exist is retarded.