I am dropping out of 538 questions. 538 is going to adjust and drop polls to get and keep a Biden lead. But not clear how much manipulations will be made. But not a clear prediction with transparent methodology

They don't release the horse race poll adjustments but the Biden approval adjustments sum to negative 60 something, so if anything, they're making him look worse. Their methodology is spelled out explicitly. It's RCP that cherry picks, drops polls in ways that help Republicans, uses Rasmussen which is pure garbage, and tells us nothing about their methodology.

if this were true, surely you should just go all in on biden winning?

Your specious leap in logic bores me.

I want an accurate measurement that tells me how it works and not some POS right wing bullshi using Rasmussen. I want an average that weights for quality, sample size, and recency. I don't even mind the house effect adjustments or any of the other tricks, because they give me the data and I can work it out without those things.

538 and realclearpolitics divergence. 538 dropped Rasmussen three months ago. Not sure what other pollster selection adjustments made by 538.

Ras is the only poll 538 has banned in the GEM era. RCP, OTOH, only uses the polls they cherry pick and says nothing about what criteria determines their selection.

@traders this just got interesting! Biden has taken the lead in the average for the first time. We're adding subsidy and making this a prize market:

It’s a bit hard to call one being ahead on 0.1% when the metric is so imprecise to begin with. I think ahead definition should at least account for confidence intervals?

So if the confidence isn't high enough the market would just resolve N/A? I feel like just specifying whose winning on July 1st according to 538 is pretty decisive.

That’s a choice. Another is to say no, Trump is not beating if the delta is within CIs.

It’s still decisive.