Only 3 days since I published the midterms forecast and the House already moved a lot. A batch of polls in districts that had none pushed Democrats to 91%. It’s a good thing I bought D House a long time ago and never sold.
The Senate slipped a bit to 64%, mostly on Nebraska, where OMEN is more bullish on Ricketts than the market.
And Brazil votes tomorrow. My model has Flávio Bolsonaro at 82%, slightly above the market (again).
Everything below 👇
Replying to @HodlMagoo
Apparently so is your brain when faced with probabilities.
Replying to @cryptobasenji
I got mislead a bit when the news mentioned both sides wanting to extend voting... I managed to switch sides eventually and turn green, but it was a long night
Replying to @cryptobasenji
Events with 18% chances of happening happen roughly 18% of the time 🤷♂️
The fact that both Lula and Flavio asked for the extension makes me feel ok about it... Not sure why the markets are dumping though, doesn't look justified.
Replying to @bernardbulletin @DavidD_Chapman
I found that most early vote projections are bogus…
Replying to @JoshHousden
Yeah, it’s impossible to push a poll-based model much above 90% with the data we have
Replying to @globalelectsim
Ballsy forecast… I’m at 82% and I’m not sure how could you push a model above 90% with the polls we have
Replying to @lthlnkso
The law of large numbers is surprisingly difficult to grasp for many…
And now the market is on par with my forecast..
Polymarket@PolymarketOct 24🇺🇸
United StatesConnected via WebAccount-level information, not a live location or per-post device.
BREAKING: Flávio Bolsonaro surges in Brazil’s presidential election odds, with less than 24 hours until voters head to the polls.
82% chance.
Replying to @PersElection
Vivek to maksymalnie ma 33% szans na wygraną, rynki dają nawet mniej, bo 18%. Ostatnia reklama przeciwko Acton to był strzał w kolano.
I've taken a day off after this forecast, but tomorrow I'm back with a small update on the midterms and the Brazil presidential runoff.
It's here! After weeks of work, my midterms model is live. 506 races modeled, with commentary to the most contested ones across Senate, House and Governor elections.
It is the most ambitious project of mine to date—4.5k words and plenty more code lines, that hopefully made a good forecast!
It differs a bit from mainstream models, but has no overall party-lean. And it seems that at least a few important races are mispriced, most notably Alaska Senate race!
Look below for the full model deep dive 👇





