Content
@
https://warpcast.com/~/channel/dropglobalnews
0 reply
0 recast
0 reaction
Murtaza Hussain
@mazmhussain
I find that polling and other types of political forecasting have lost meaningful predictive power but for what it’s worth models are showing Kamala Harris narrowing the gap with Trump and potentially moving past him. Trump had a huge advantage in 2016 which is that he was a novel figure. Americans love change, excitement, and the unknown and he fed that sentiment. Harris now ironically enjoys that to a greater degree since unlike Trump she’s never been president before. So despite the fact that she had been an underwhelming VP I would not count her out as a strong POTUS candidate.
4 replies
0 recast
3 reactions
MJC
@mjc716
the paradox of political forecasting is that there will never be enough sample size to validate any model
2 replies
0 recast
0 reaction
Eddie Wharton
@eddie
Challenging, but possible when someone issues hundreds of forecasts each election cycle Assessment of 538’s forecast calibration while Nate was there https://projects.fivethirtyeight.com/checking-our-work/
1 reply
0 recast
1 reaction
MJC
@mjc716
yea i read and love nate and he's the best at it clearly, but i guess the point i was making it that politics is a lot more idiosyncratic than baseball
1 reply
0 recast
1 reaction
Eddie Wharton
@eddie
For sure and politics is a way smaller sample since states have high correlation each cycle All their presidential forecasts since 2008 — way wider credible intervals, but also shows some level of calibration
1 reply
0 recast
1 reaction
Connor McCormick ☀️
@nor
Beautiful graph
1 reply
0 recast
1 reaction
Eddie Wharton
@eddie
Agreed, calibration plots are one of the cooler graphs and 538 really nails the aesthetic with clean CIs
0 reply
0 recast
1 reaction
