We backtested two variations of the same core strategy this week: buying longshot YES positions in the 5-15¢ range (think "unlikely but possible" events) with tight time windows before expiration. The only difference? One closed positions within 24 hours, the other within 6 hours. The results were so good we had to double-check the data.
Both strategies took roughly 300 trades at a ~17% win rate and somehow turned $100 into $290–$330. Yes, you're reading that right. A 17% win rate doesn't usually feel like a party, but here's the magic: longshots move fast. When you're betting on something priced at 5-15¢ that you think is undervalued, even small wins compound quickly. The market was clearly misprice-pricing low-probability events across every category—weather, sports, crypto, politics, you name it.
The 6-hour version edged out the 24-hour version by $40, and the win rate ticked up 0.7 percentage points. Why? Shorter windows mean less time for the market to correct your edge. If you've identified an underpriced YES bet at 12¢, you want to close before the crowd catches on. By hour 24, sentiment can shift. By hour 6, you're probably already out.
That said, the difference is marginal. Both versions worked, which suggests the underlying edge (spotting cheap YES contracts on genuinely mispriced events) is the real driver, not the timer itself.
This is simulated performance on historical data—not actual money deployed. Past performance doesn't predict future results, blah blah, the usual disclaimers. More importantly: a 17% win rate requires you to be very confident in your event assessments. One unlucky week where you're wrong on 80% of your picks instead of 83%? The math flips fast. Also, you're doing 300+ trades. Execution risk, slippage, and the mental bandwidth to handle that volume are real factors that don't show up in backtest results.
The breakdown isn't detailed here, but the strategy ran across weather, sports, crypto, financials, politics, economics, and culture. That's a refreshingly broad set—suggests the mispricing was systemic, not isolated to one event type. Everyone seemed to be undervaluing unlikely-but-real outcomes across the board.
Did this work because of genuine market inefficiency, or because longshots as a category got a lucky break this week? If you ran this exact strategy against the *previous* week's data, would it hit 230% returns again? That's the stress test that matters.
Bottom line: Longshot YES betting with tight exit windows had a phenomenal week in simulation. It's worth tracking whether this holds up or if we're just seeing mean reversion settle in.
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.