Research Report: Kalshi BTC 15-Minute Momentum Strategy

Short Disclaimer

This report is historical simulation research only. No live trading occurred. Past simulated performance does not guarantee, and should not be interpreted to imply, future results.

Intro / Thesis

The core idea was straightforward: on the Kalshi KXBTC15M contract (a 15-minute binary on Bitcoin price direction), when the YES price reaches exactly $0.60 with a tight bid-ask spread (≤$0.03), and Coinbase BTC-USD is registering positive 5-minute momentum, the market is showing enough conviction and liquidity to carry the YES price meaningfully higher within the remaining window. The thesis bet that this specific confluence—a neutral-to-bullish price level, narrow spread signaling agreement, and rising spot BTC—would produce a momentum continuation trade. We entered long 10 contracts and aimed for a quick +$3.00 profit with a matching -$3.00 stop.

Variant and Strategy Explanation

We ran 100 variants of this base concept. All variants shared the same foundational DSL structure: entry triggered at exactly $0.60 YES price, spread ≤$0.03, and positive BTC-USD 5-minute change sourced from Coinbase. Entry size was always 10 contracts. Exits were mechanical: take profit at +$3.00 unrealized P&L, stop loss at -$3.00, and a forced expiry exit when fewer than 60 seconds remained in the contract window.

The 100 variants explored different combinations of parameter values—shifting the entry price threshold, widening or tightening the spread requirement, adjusting the BTC momentum lookback, changing position sizing, and modifying the P&L exit levels. The goal was to see if any tuning could surface a profitable edge that the base parameters missed.

Every variant that produced results was automatically saved as a runnable Turbine strategy, meaning the simulation logic is preserved and reproducible.

Top Results

The highest-ranking variants were, in truth, a study in consistency—unfortunately the wrong kind. The top eight variants (ranks 1 through 8) produced identical performance metrics:

• Total P&L: -$3.96
• ROI: -79.2%
• Sharpe: -0.21
• Max Drawdown: -9.73%
• Total Trades: 12
• Win Rate: 50.0%

Six wins, six losses. The take-profit and stop-loss were each $3.00, but slippage and spread costs on the losing trades slightly outweighed the winners, generating a net loss of just under $4 across the simulation period. A 50% win rate sounds respectable until you see the negative Sharpe and nearly 80% drawdown on deployed capital.

The top eight are effectively identical in economic outcome despite running through different parameter IDs. This suggests the simulation searched a region of parameter space where the core entry logic dominates any marginal tuning—the thesis simply did not hold up under the data.

Bottom Results

Ranks 81 through 88 (the worst-performing variants in the set) show what happens when parameter choices make a fragile strategy worse:

• Total P&L: -$11.32
• ROI: -113.2%
• Sharpe: -0.37
• Max Drawdown: -19.52%
• Total Trades: 12
• Win Rate: 50.0%

Again, a 50% win rate—but the losing trades here were deeper. These bottom variants likely used wider stops, larger position sizes, or entry thresholds that captured the trade at worse average prices, causing the same six losing trades to incur roughly triple the losses of the top group. The drawdown nearly doubled to almost 20%.

The symmetry is instructive: both the "best" and "worst" variants lost money. No variant in the 100-run batch produced a positive total P&L, positive Sharpe, or positive ROI. The absolute best outcome was simply losing less than the others.

Conclusion

The $0.60 entry thesis with tight spread and rising BTC momentum did not translate into a profitable edge in historical simulation. The 50% win rate across all variants tells the real story—this was coin-flip territory, with transaction costs and adverse price movement on exits producing a consistent bleed.

Key takeaways:

1. Momentum at $0.60 is unreliable. Reaching exactly $0.60 with BTC rising did not predict continued upward movement with enough frequency or magnitude to overcome the stop-loss drag.
2. Spread tightness didn't help. Even when the market was liquid (spread ≤$0.03), the edge failed to materialize.
3. Systematic losses across all 100 variants. No amount of parameter tuning rescued the thesis. This is a clean negative result.
4. Low trade count. Twelve trades over the simulation period is a thin sample, but the results were uniformly negative enough that larger sample would be unlikely to flip the sign.

For researchers, this is a useful falsification. The strategy logic is sound mechanically—rapid entry, defined risk, exit before expiry—but the predictive signal is absent. The saved strategies remain available for review or to serve as templates with entirely different entry criteria.

Long Disclaimer

This research report presents historical simulation results generated by Turbine's automated strategy testing infrastructure. No actual capital was deployed, and no real trading occurred. All performance metrics—including P&L, ROI, Sharpe ratio, win rate, and drawdown—are derived from backtesting on historical market data.

Historical simulation performance does not reflect actual trading, does not account for all real-world frictions (such as partial fills, exchange outages, or wider spreads during volatile periods), and does not guarantee that any strategy will be profitable in the future. Simulated results are inherently limited by the quality and completeness of historical data, assumptions about execution, and the specific time periods analyzed.

The strategies discussed may contain parameter choices that overfit to historical conditions. Markets evolve, and strategies that underperform in simulation may underperform or perform differently in live environments. This report is for informational and research purposes only and does not constitute investment advice, a recommendation, or a solicitation to trade.

Turbine makes no representations about the suitability of any strategy for any particular use case. All trading involves risk of loss.