January 19, 2026
Place your bets, lose your chill
The Microstructure of Wealth Transfer in Prediction Markets
Kalshi’s “Optimism Tax” has bettors fuming, quants flexing, and bot rumors flying
TLDR: A huge study of Kalshi trades says longshot “YES” bettors lose while calm market-makers profit. Comments exploded: some cheer the “optimism tax” finding, others say the math is wrong, and a few blame bots—proof prediction markets run on hype, psychology, and who plays house.
Forget Vegas—on Kalshi, traders are paying more than slot machines’ house edge to chase longshots, and the comments are a full-on math fight. One top reply drops a mic: 72.1 million trades show a longshot bias where tiny “YES” prices win even less than they imply, while makers quietly farm an Optimism Tax from impulsive takers. Translation: the chill sellers get richer; the hype buyers get rinsed.
Then the thread splits. The skeptics blast the study as “probability 101 fail,” insisting efficient prices mean you shouldn’t lose—cue a flurry of corrections and eye-rolls. Sports fans pile in, puzzling over YES/NO asymmetry like it’s playoff overtime, while quants explain that the crowd loves affirmatives (“YES” at long odds) and pays for it. Another curveball: someone asks about interest rates—if a contract definitely pays $1 next year, why not cap it at ~96¢? Cue memes of “Where did my 4¢ go?”
And the bot conspiracy hits: are these markets just LLMs brawling with bankrolls? The thread devolves into jokes about “GPT vs ApeTrader” and “quants vs vibes.” Verdict from the crowd: prediction markets aren’t just about probabilities—they’re about humans, hype, and who’s willing to be the house.
Key Points
- •Study covers 72.1 million Kalshi trades totaling $18.26 billion from 2021–2025.
- •Evidence of longshot bias: low-probability contracts are overpriced, with some longshots returning as low as 43 cents on the dollar.
- •Systematic wealth transfer observed: liquidity takers overpay for YES outcomes; makers capture an “Optimism Tax.”
- •YES/NO asymmetry identified: takers disproportionately favor YES at longshot prices, exacerbating losses.
- •Inefficiency varies by category: strongest in Sports/Entertainment; Finance approaches near-perfect efficiency; data cutoff 2025-11-25.