Sometimes the best trading strategies are the ones that feel counterintuitive. This week, our backtest data threw us a curveball: the absolute lowest-probability bets on Kalshi—those sub-20¢ longshots with tight time windows—actually crushed it. We're talking nearly 5x returns on capital, which is the kind of number that makes you squint at the data twice to make sure it's real.
Spoiler: it is. Let's dig in.
Let's pause on that win rate for a second. Eighteen to nineteen percent means you're winning roughly 1 out of every 5 trades. That's not high. That's low. Yet somehow both strategies still printed money—and a lot of it.
This is the beauty (and the trap) of asymmetric payoffs. When you're buying YES contracts at 5-15¢, you're risking a nickel to potentially make 85¢. The odds don't need to be in your favor for the math to work out, as long as when you do win, you win big enough often enough.
Backtesting over 300+ trades across weather, sports, crypto, financials, politics, economics, and culture tells us something important: there was consistent mispricing across categories last week. Whether Kalshi's user base was skewing pessimistic, or certain event categories just had better-than-expected outcomes, the 5-15¢ YES tickets were systematically undervalued.
The 24-hour window strategy edged out the 6-hour version by about $68 per $100 invested. That's a notable difference. More time to resolution = more opportunity for events to resolve YES, apparently. But both made money, so the real story is: tight timeframe, long-shot YES bets worked this week, period.
Zero drawdown on both strategies is also suspicious in the best way possible—it suggests these weren't just lucky; there was genuine edge here.
That said, the consistency across 600+ combined trades and seven different market categories suggests there's something real here, not just variance.
The question for the coming week: was this a temporary inefficiency in Kalshi's pricing, or a durable edge? If the same strategy keeps working, we're probably looking at a genuine gap between how the market prices sub-20¢ YES events and their actual resolution rates. If it reverts, we learned a valuable lesson about sample sizes and survival bias.
Either way, this week proved that sometimes the most boring-sounding strategy—buying cheap longshots with hours left to resolve—can be the one that actually prints money.
Want to test these strategies yourself? Backtest and analyze any of them in chat — 250,000 free credits, no credit card required.
Try Prediction PilotEvery week we test strategies against real Kalshi data and share the profitable ones.