Public Research Report

Venue: Kalshi
Market type: Crypto
Series ticker: KXBTC15M
Strategy type: Custom
Loop interval: 10 seconds

Short Disclaimer

This is historical simulation research only. It does not constitute trading advice, a prediction of future returns, or a recommendation to deploy capital.

Intro / Thesis

The working idea was straightforward: BTC 15-minute spot momentum may persist into Kalshi's KXBTC15M market. When Coinbase shows a meaningful 15-minute move, the same-direction YES/NO contract should still have time to converge toward settlement before the market closes. Entering during the move and taking profit before close should, in theory, beat random entry.

We ran 100 parameter variants across a price floor/ceiling grid to see whether that thesis holds up in simulation and whether the edge survives robustness checks. The short answer: there is something here, but it is not bulletproof, and much of the apparent outperformance is concentrated in a narrow band of parameter settings.

Variant and Strategy Explanation

The base strategy uses the following structure:

• Asset: BTC
• Series ticker: KXBTC15M
• Edge feed: Coinbase BTC-USD, 15-minute change, refreshed every 10 seconds
• Position size: 6 contracts per entry
• Max position: 10 contracts
• Entry logic:
  • If change_15m > 0.005 and price is within the allowed band, buy YES
  • If change_15m < -0.005 and price is within the allowed band, buy NO
• Exits:
  • Cancel all orders when time to expiry is ≤ 2 minutes
  • Sell everything if unrealized PnL ≤ -$1.00
  • Sell everything if unrealized PnL ≥ $8.00

The parameter sweep varied two things:

• Price floor: 0.05, 0.09, 0.14, 0.18, 0.23, 0.27, 0.32, 0.36, 0.41, 0.45
• Price ceiling: 0.55, 0.59, 0.64, 0.68, 0.73, 0.77, 0.82, 0.86, 0.91, 0.95

All 100 combinations completed successfully. Each successful variant is saved as a runnable Turbine strategy.

The core question was whether widening the price ceiling toward 0.82 while keeping the floor low would let the strategy capture more directional convergence before the market closes.

Top Results

The top-ranked variants all clustered around the same configuration:

• Price floor: 0.05 to 0.36
• Price ceiling: 0.82

The top eight variants are effectively identical in outcome:

The top result produced a total PnL of $24.40 over 54 trades with a 65.2% win rate and a Sharpe of 0.55. The max drawdown was -$5.97 across all top variants.

The cluster is not subtle: the winning condition was a high price ceiling of 0.82 combined with any floor from 0.05 to 0.36. Below 0.36, the floor didn't matter much — the strategy was buying contracts cheap enough that the upside from momentum convergence dominated.

Bottom Results

The worst-performing variants all involved a low price ceiling of 0.55, which cut off the strategy's ability to buy contracts that still had meaningful upside. The bottom of the table shows what happens when you constrain the entry band too tightly:

The worst variant (rank 100) actually lost money: -$1.70 over 18 trades with a 37.5% win rate. That configuration used a floor of 0.45 and a ceiling of 0.55 — essentially buying expensive contracts with almost no room to run before the ceiling rule blocked entry.

The broader pattern: when the ceiling dropped to 0.55, total PnL fell to $7.23 or lower across most floor settings. When the floor rose to 0.45, total PnL collapsed to -$1.70, $2.90, or $14.93 depending on the ceiling. The strategy needed both cheap entry and room to sell into convergence.

Conclusion

The thesis had merit in the specific region where the price ceiling was high enough to allow convergence capture and the floor was low enough to keep entry prices cheap. The top results cluster at a ceiling of 0.82, and the parameter sweep shows a clear degradation as the ceiling drops toward 0.55.

However, the robustness statistics require caution.

The permutation test shuffled the edge feed (the Coinbase 15-minute change signal) while holding market prices fixed. The real best PnL was $24.40. Across 162 time-scrambled versions of the same feed, the p-value was 0.0245. That means the strategy's edge-feed timing beat roughly 97.5% of the scrambled versions. That is below the 0.05 threshold and suggests the timing signal is doing real work.

But there are caveats:

• The permutation scheme is explicitly flagged as degraded in the output. The market price series was not permuted, so price-based conditions in the rules were not tested by this design.
• The deflated Sharpe was 0.95, which is above the expected max Sharpe of 0.25 under a null of no edge. That is encouraging, but it comes from a single family of 100 parameter combinations.
• The neighborhood degradation was 0.049, meaning nearby parameter settings degrade only slightly. That is a positive sign — the top result is not an isolated spike. However, the bottom of the sweep shows the edge is fragile when the ceiling is constrained.

The honest read: the edge is real in simulation but narrow. The strategy worked best with a price ceiling of 0.82 and a floor of 0.36 or lower. It worked poorly with a ceiling of 0.55 or a floor of 0.45. The permutation test supports the idea that the momentum signal timing added value, but the degraded status of the test means we shouldn't overstate confidence.

This is not a "strong, validated, promising" strategy in the sense that it would survive all market conditions. It is a directional momentum pattern that worked in simulation under specific entry-band settings. Whether it persists out-of-sample is an open question.

Each successful variant is

About this Kalshi strategy backtest

What question does BTC 15m momentum investigate?

BTC 15-minute spot momentum persists into Kalshi's KXBTC15M market. When Coinbase spots a meaningful 15-minute move (0.3%-1.0%), the same-direction YES/NO contract still has time to converge toward settlement, so entering during the move and taking profit before close beats random entry.

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.