BTC Maker Sweep

Kalshi BTC Maker Strategy Sweep: KXBTC15M

Short Disclaimer

This is historical simulation research only. Past performance does not indicate future results. These findings are for informational purposes and do not constitute investment advice.

Intro / Thesis

We ran a parameter sweep on maker-style quoting for Kalshi's 15-minute Bitcoin binary contract, KXBTC15M. The core question: how do three tactical levers—quote offset, minimum spread threshold, and order refresh interval—affect profitability when you're trying to earn the spread rather than take it?

The strategy is straightforward in concept. Every 10 seconds, if market conditions pass a set of filters (spread wide enough, far enough from expiry, away from economic events, position room available), the strategy posts one-lot bids on both sides of the market at some offset from the best bid. The offset is the main variable we tested: quote at the touch (0¢), or step back by 1¢ to 4¢. We also varied how often orders refresh—10, 12, 14, 16, or 18 seconds—and tested one variant with a 2¢ spread threshold instead of the base 1¢.

All variants share the same risk scaffold: $10 max position, 5-contract portfolio cap, hard stop at –$5, and automatic flattening around FOMC/CPI events plus 15-minute expiry windows.

Variant and Strategy Explanation

Each variant is a concrete, runnable Turbine strategy saved to the platform. The base engine is the same; only the three parameters below change:

The sweep covered 25 parameter combinations (5 offsets × 5 refresh intervals), plus one additional 2¢-spread variant, for 100 completed simulations total. All 25 grid cells ran successfully.

Top Results

The raw leader is clear, but the statistics underneath tell a more sobering story.

Rank 1: 0¢ offset, 1¢ spread, 10s refresh

• Total PnL: +$11.79 (117.9% ROI)
• Sharpe: 0.33
• Max drawdown: –$3.79
• 470 trades, 49.1% win rate

This is the only variant with a Sharpe above 0.20 and positive returns. It quotes at the touch, refreshes fastest, and trades most actively. The high trade count matters: it's capturing small edges repeatedly, and the 0¢ offset means it gets hit whenever the market moves through the bid.

Ranks 2–8 cluster in the +$2.50 to +$4.90 range. A pattern emerges:

Two distinct regimes appear to work: quote at the touch (0¢) with frequent refresh, or step back 3¢ and accept fewer, higher-win-rate fills. The –3¢ variants all show 56–58% win rates versus ~44–49% for 0¢, but they trade roughly one-third as often. The 0¢ variants compensate with volume.

Notably, no offset between –2¢ and –1¢ appears in the top eight. The middle ground is barren.

Bottom Results

The worst performers are concentrated in a narrow band: offsets of –1¢ to –2¢, with scattered –4¢ damage.

Rank 100: –2¢ offset, 1¢ spread, 10s refresh

• Total PnL: –$5.75 (–57.5% ROI)
• Sharpe: –0.32
• Max drawdown: –$5.75
• 64 trades, 40.9% win rate

Rank 99: –2¢ offset, 1¢ spread, 16s refresh

• Total PnL: –$5.69 (–56.9% ROI)
• Sharpe: –0.43
• 71 trades, 40.7% win rate

The –2¢ offset is catastrophic across all refresh intervals tested (ranks 92, 93, 96, 99, 100). It appears to be the worst of both worlds: passive enough to miss the good fills, aggressive enough to get picked off when you're wrong. The –1¢ offset is nearly as bad (ranks 95, 98). The –4¢ offset shows up once in the bottom five (rank 97, –54.9% ROI), suggesting extreme passivity also fails, though less consistently.

The 2¢ spread threshold variant (rank 93, –53.5% ROI) performed poorly, indicating that requiring wider spreads did not filter for better opportunities—it simply eliminated too many trades, and the ones that remained were worse.

Conclusion

The sweep produced a wide dispersion: +117.9% at best, –57.5% at worst, from the same core strategy with minor parameter tweaks. That alone is a warning.

The marginal averages by parameter are instructive. Quote offset of 0¢ averages +$4.86 across refresh intervals; –3¢ averages +$3.34. Every other offset averages negative, with –2¢ the worst at –$5.44. For refresh interval, 10s averages barely positive (+$0.28) while all longer intervals average negative—yet the single best cell is 10s, and the second-best is 18s. The interaction dominates the main effect.

However, the statistical validation is weak. The deflated Sharpe ratio is 0.25, well below the 0.95 threshold and close to the expected maximum Sharpe of 0.45 for a random strategy in this search space. The DSR warning is explicit: the top result is not distinguishable from the luckiest outcome of skill-less trials. No permutation p-value is available—the test did not run—so we cannot even quote a formal significance level.

What we can say: the parameter surface is unstable. Small moves from 0¢ to –1¢ or –2¢ flip the strategy from profitable to deeply unprofitable. The "working" variants are few and geographically isolated in parameter space. The neighborhood degradation of 0.94 means that adjacent parameter combinations perform drastically worse—exactly the signature of an overfit peak.

Practical takeaway: If one were to deploy anything from this family, the 0¢/10s variant or the –3¢/18s variant are the only candidates with positive simulated history. But the evidence does not support confidence that either will persist. The strategy's profitability appears fragile, and the top results are consistent with selection noise and overfitting rather than genuine edge. Each variant is saved as a runnable Turbine strategy for further paper testing or live deployment at your own risk.

Long Disclaimer

This report presents historical simulation results only. All performance figures—including returns, Sharpe ratios, win rates, and drawdowns—are derived from backtests using past market data. They do not represent actual trading results, and they do not guarantee or imply any future performance.

The research was conducted by systematically varying parameters across a predefined grid. This process introduces multiple comparison problems: with 25+ variants tested, some will appear superior by chance alone even if no true edge exists. The deflated Sharpe ratio and related robustness statistics are designed to account for this, and in this case they indicate that the top-ranked variant does not meet conventional thresholds for statistical reliability. The absence of a completed permutation test means no p-value is available to further assess significance.

Market conditions change. Liquidity, volatility, fee structures, and competitor behavior in Kalshi's KXBTC15M market may differ materially from the simulated period. Slippage, latency, and execution assumptions in simulation may not match live trading reality. The strategy's reliance on maker rebates and post-only execution assumes favorable queue positioning that may not hold under competition.

The strategies described involve substantial risk of loss. The base configuration includes a –$5 stop loss, but this is not guaranteed to execute at the specified level in fast markets. Position limits and economic-event blackouts are risk controls, not guarantees.

Turbine provides these research tools for informational and educational purposes. We are not a registered investment adviser, broker-dealer, or fiduciary. Nothing in this report constitutes a recommendation to buy, sell, or hold any financial instrument, nor does it constitute legal, tax, or financial advice. Users should conduct their own due diligence and consult qualified professionals before deploying capital.

Each variant referenced is saved as a runnable strategy within the Turbine platform. Saving or running a strategy does not imply endorsement of its profitability. Live deployment is at the sole discretion and risk of the user.

About this Kalshi strategy backtest

What question does BTC Maker Sweep investigate?

Exploring how quote offset, min spread, and refresh interval affect maker profitability on KXBTC15M.

Read the analysis alongside the individual variant results. Check which entry, exit, price-band or risk settings changed, which settings stayed fixed, and whether the report documents enough historical coverage to support its comparison. A different parameter value defines a different simulation; it does not establish a new source of live trading returns.

How many strategy variants are included?

This report lists 100 completed and 0 failed strategy variants. Failed comparisons do not count as successful tests with zero profit. Review trade counts and drawdown alongside net P&L, win rate and Sharpe, and read any coverage or execution caveats in the analysis. A result based on very few trades needs particular care, even when its return looks high.

Does a profitable Kalshi backtest predict live performance?

These are historical simulations. Live results can differ because of liquidity, fees, slippage, latency, market resolution, outages and data quality. Selecting the best result after comparing many variants can also overstate how reliably it will perform on new data. Compare nearby settings, inspect losing variants and test on a separate period before considering a live run. The report is research, not a promise of profit or investment advice.

How can I inspect or test a strategy from this report?

Use an available Open strategy link to inspect that variant. Review its rules, market, date window, fill assumptions, fees and trade history before changing parameters or rerunning it. Confirm which data and execution assumptions the selected market actually supports; a replay does not establish that every historical order could have filled live. For the workflow, read the Kalshi and prediction market backtesting guide. The Kalshi backtesting software page explains the product's supported testing workflow.