Research Report: KXBTC Hourly YES Early-Entry Strategy

Short Disclaimer

Historical simulation research only. No future performance is implied. All results below are from completed backtests and parameter sweeps, not live trading.

Intro / Thesis

This report examines a custom Kalshi strategy for KXBTC hourly YES contracts. The underlying idea was that BTC hourly YES contracts printing 0.97–0.98 within the first 10 minutes of the market's life are near-certain to resolve YES, while a sharp drop to 0.95 would act as an early-warning reversal signal. The strategy was designed to buy YES at 0.97–0.98, hold winners to settlement, and exit all positions if price hit 0.95 or below, capping losses at $1.

The sweep completed 100 of 100 planned variants. However, the results are not what the thesis hoped for.

Variant and Strategy Explanation

The base strategy was specified in YAML as a custom Kalshi strategy on the KXBTC series, with a 10-second loop interval. The entry rule required:

• Time to expiry greater than 50 minutes
• Price between 0.97 and 0.98 inclusive
• Position size below the max of 3 contracts

The strategy would then buy 1 YES contract. Two exit rules were defined:

1. Stop loss: If price fell to 0.95 or below while holding a position, sell everything.
2. Hold to settlement: If time to expiry reached zero, sell everything.

Risk limits were set at a maximum position of 3 contracts, maximum loss of $1.00, price floor of $0.05, and price ceiling of $0.98.

The parameter sweep varied two axes:

• risk.price_floor: 0.05 through 0.45 in 10 steps
• risk.price_ceiling: 0.55 through 0.95 in 10 steps

This produced a 10 × 10 grid of 100 total cells, all of which completed.

Each successful variant is saved as a runnable Turbine strategy for future reference or live deployment if conditions warrant.

Top Results

The top 8 variants, ranked by net PnL, are shown below:

Every one of the top-ranked variants had zero trades, zero PnL, zero Sharpe, and zero win rate. This is not a case of a strategy performing well — it is a case of a strategy never entering a position.

Bottom Results

The bottom 8 variants are identical to the top 8, because all 100 cells in the sweep produced exactly the same result: zero trades, zero PnL, zero Sharpe, zero win rate, zero drawdown. The parameter sweep produced no differentiation whatsoever.

The entry condition — price between 0.97 and 0.98 with more than 50 minutes to expiry — was never met during the simulation window. The strategy simply sat idle.

This flatness is not evidence of robustness. The robustness statistics explicitly flag both swept parameters as inert: varying risk.price_floor and risk.price_ceiling produced identical trades in every cell because the parameters were never exercised. The strategy never entered a position, so no exit rules were ever triggered, and no risk limits were ever tested.

Conclusion

The thesis behind this strategy — that early high prints of 0.97–0.98 in KXBTC hourly YES contracts are near-certain winners — was not supported by any actual trade in the simulation. The entry condition was never satisfied during the tested window. As a result, the strategy had zero trades, zero PnL, and zero drawdown across all 100 variants.

The permutation test returned a p-value of 1.000, meaning the sweep's best result beat 0.0% of 1,000 luck-only re-sweeps of the same grid. The top results are entirely consistent with selection noise and overfitting. There is no signal here to report.

The warnings are clear:

• Winner has 0 resolved trades, well below the 30-trade threshold for statistical reliability.
• Winner has only 2 distinct PnL days, well below the 10-day threshold for meaningful Sharpe.
• Both swept parameters were inert and never exercised.

This is a null result. The strategy as specified does not trade under these historical conditions. It should not be described as promising, validated, or even weak — it simply did not engage the market in this simulation window.

Long Disclaimer

This report is historical simulation research produced for informational purposes only. It does not constitute investment advice, a recommendation to buy or sell any security or contract, or a prediction of future performance.

All results were generated from parameter sweeps and permutation tests on historical data. Historical simulation has inherent limitations, including but not limited to: survivorship bias, look-ahead bias, data integrity issues, and the fact that past market behavior does not predict future market behavior. The strategy described may perform differently in live markets due to execution costs, slippage, liquidity constraints, and other real-world frictions not modeled in simulation.

The permutation test p-value of 1.000 indicates that the observed results are indistinguishable from random luck. The zero-trade outcome across all 100 variants means no meaningful statistical inference can be drawn about strategy edge, risk management, or expected returns.

Each successful variant is saved as a runnable Turbine strategy for transparency and reproducibility. However, the existence of a runnable strategy does not imply that it will be profitable or that it should be deployed with real capital. Any user considering live deployment should conduct their own due diligence, including out-of-sample testing, paper trading, and independent risk assessment.

Turbine, the platform, and the author of this report make no representations or warranties regarding the accuracy, completeness, or suitability of this research for any purpose. The user assumes full responsibility for any trading decisions made based on this material. Trading binary options and cryptocurrency derivatives involves substantial risk of loss and is not suitable for all investors.