Honest question: is there a reliable website that takes the polls and runs simulations of the odds of Biden winning without considering the economic “fundamentals”? I think the answer is “no” just because polls aren’t predictive this far out because things change… I’m asking for something probabilistic that allows for the possibility of systematic polling error, the probability of polls not being predictive this early (with changes going in either direction), and that discounts the less reliable polling companies. We have pundits saying that the path to victory has narrowed to just the Blue Wall of WI, MI and PA (and NE-2), but that’s not how these things work. A 20% chance of winning GA might have a meaningful impact on the overall odds. (I just have no idea how meaningful).
the 538 forecast which includes fundamentals like the economy and I think incumbency gives President Biden an optimistic <strike>49%</strike> 51% chance of winning (up from a low of 46 a few days ago). Meanwhile the Economist forecast gives Biden a 26% chance of winning (decreasing dramatically from about 40% in early June with the drop predating the debate). Then there’s The Hill which shows a relatively stable race and Biden having about a 44% chance. All 3 of these simulations use fundamentals. However, lots of the conversation these past weeks has focused on a polls-only type of analysis where a decrease of a few points makes all the difference. What I don’t see though are any somewhat reliable websites that run a polls-only based simulation. For example, I think 538 would provide critical information if it released a summary of what a simulation shows using only the polls while allowing for the possibility of systematic polling error, etc. It’d be even more relevant to also understand the extent that those odds have changed these past 3 weeks.
The campaign’s strongest path to victory does seem to run through WI, MI, PA and NE-2. In the blue wall scenario (of WI, MI and PA), to simplify the math if I artificially give Biden a 100% chance to retain VA, NH, ME and MN (and a 0% chance to win AZ, NV and GA); then Biden has somewhere around a 10% chance of winning. Under an optimistic scenario that focuses only on the blue wall that would go up to a 2 in 10 chance.
Grains of salt
However, the real bonus from an optimistic scenario is putting AZ, NV and GA back in play. Having these States in play opens up several pathways and could have a substantial impact on odds. AZ+NV or just GA alone can replace either of WI or MI. GA together with NV (or AZ) can offset PA. Etc, etc. So just how bad are the odds based on the current polls? The fundamental problem with this question is that polls aren’t predictive this far out; things can get worse or better. But it would be informative.
Further reading:
www.pbs.org/…
“Why 538's forecast hasn't moved much post-debate”
abcnews.go.com/…
elections2024.thehill.com/…
Update: listened to Nate Silver’s recent podcast (risky business), and in minute 38 he said his model shows Biden with a 29% chance to win (allowing for the possibility of changes to public opinion and that Biden can campaign well). But Nate also spitballed that Biden has a 20% chance to win, and Harris a 35 to 40% chance per his model… FWIW.