Deep Research

BTC Dual Trend Confirmation

[M5 remote-plane validation #2 - 4 permutation workers] Multi-horizon BTC trend confirmation predicts repricing in Kalshi KXBTC15M: buy YES when both 5m and 1h BTC returns are positive, buy NO when both are negative. Only enter when price is $0.25-$0.75, spread <= $0.03, and >2m remain. Test whether trend agreement produces positive, repeatable historical performance.

Historical research only. Not investment advice.

Top Strategy Variants

Bottom Strategy Variants

Research Report: Kalshi KXBTC15M — Multi-Horizon BTC Trend Confirmation

Short Disclaimer

Historical simulation research only. No future performance is implied. Past results, including negative ones, do not guarantee future outcomes.


Intro / Thesis

This report covers validation #2 for a remote-plane strategy on Kalshi's KXBTC15M market. The core idea was simple: when both the 5-minute and 1-hour BTC returns point the same direction, do short-dated YES/NO contracts in that market reprice predictably enough to extract an edge?

The setup used Coinbase BTC-USD data at a 30-second loop interval, entered only within tight price and spread bands, and flattened positions near expiry or on a stop-loss. The thesis was that trend agreement across two timeframes would filter out choppy conditions and produce repeatable, positive historical performance.

The simulation ran 100 completed variants across parameter changes. The results below are what they are.


Variant and Strategy Explanation

The strategy is custom, built on the KXBTC15M series ticker. It selects the most liquid market in the series and loops every 30 seconds. BTC edge data refreshes every 10 seconds.

Entry logic:

  • Buy YES when change_5m > 0 AND change_1h > 0
  • Buy NO when change_5m < 0 AND change_1h < 0
  • Price must fall between the entry band (initially $0.35–$0.65)
  • Spread must be ≤ $0.03
  • More than 2 minutes must remain to expiry

Position sizing was 1 contract per entry signal, with max position of 100 and max loss of $25. Positions are flattened if unrealized PnL hits -$25 or if time to expiry falls to 2 minutes or less.

The parameter sweep varied entry_low from $0.25 to $0.45 and entry_high from $0.55 to $0.75, with spread caps ranging from $0.01 to $0.04 depending on the variant.

Important note on the sweep: the robustness statistics flagged both entry_low and entry_high as inert axes. That means changing these parameters produced identical trades in every cell during the simulation window. The parameter variations were never actually exercised by market conditions. The flatness across variants is not evidence of robustness — it's evidence that the parameters didn't bind.

Each successful variant is saved as a runnable Turbine strategy.


Top Results

The top eight ranked variants are effectively indistinguishable from each other. Most produced identical performance:

RankBandSpread CapTotal PnLROIMax DDTradesWin RateSharpe
10.25–0.550.02-$9.02-9.02%-$29.754,28151.2%-0.07
20.25–0.550.03-$9.02-9.02%-$29.754,28151.2%-0.07
30.25–0.550.04-$9.02-9.02%-$29.754,28151.2%-0.07
40.25–0.600.01-$9.02-9.02%-$29.754,28151.2%-0.07
50.25–0.600.02-$9.02-9.02%-$29.754,28151.2%-0.07
60.25–0.600.03-$9.02-9.02%-$29.754,28151.2%-0.07
70.25–0.600.04-$9.02-9.02%-$29.754,28151.2%-0.07
80.25–0.650.01-$9.02-9.02%-$29.754,28151.2%-0.07

The best variant lost money. Total PnL was -$9.02, ROI -9.02%, with a Sharpe of -0.07. Win rate was just above 50% at 51.2%, but that thin edge was not enough to overcome fees and spread costs. Max drawdown was -$29.75.


Bottom Results

The worst-performing variant (rank 100) was the tightest spread cap at the lowest entry band:

RankBandSpread CapTotal PnLROIMax DDTradesWin RateSharpe
1000.25–0.550.01-$10.14-10.14%-$44.183,95346.6%-0.04

This variant lost -$10.14, with a lower win rate of 46.6% and a significantly worse max drawdown of -$44.18. Fewer trades were taken (3,953 vs 4,281), consistent with the tighter spread cap filtering out entries — but the ones that remained were worse.

The spread between best and worst variant is narrow: -$9.02 vs -$10.14. That's a $1.12 difference. Nothing in this sweep came close to breakeven, let alone profitability.


Conclusion

This strategy does not work under the tested conditions.

The core hypothesis — that multi-horizon BTC trend agreement predicts short-term KXBTC15M repricing — is not supported by the data. The best variant lost roughly 9% and the worst lost roughly 10%. Win rates hovered near 50%, which is essentially coin-flip territory, and fees consumed more than 100% of the winner's gross PnL (111% to be exact).

The robustness statistics reinforce this. The deflated Sharpe for the winner is 0.35, well below the 0.95 threshold needed to distinguish it from skill-less trial outcomes. The permutation test (edge-feed scramble) returned a p-value of 0.0017, but this number is misleading in context: the null histogram was degenerate (all zeros), the test was flagged as degraded, and the market price series was not permuted — only the edge feed was scrambled. Price-based entry conditions were not tested by this design. The honest read is that these results are consistent with selection noise.

Bottom line: This variant is not viable. I would not deploy it, and I would not allocate further research time to this specific trend-agreement entry logic without a fundamentally different filtering mechanism or cost model.


Long Disclaimer

This report is historical simulation research produced for internal analysis purposes only. It is not investment advice, a recommendation to trade, or a solicitation to buy or sell any financial product.

All performance figures are derived from backtesting on historical data. Backtested results have inherent limitations: they do not account for live market impact, partial fills, exchange outages, API latency differences, or changes in market microstructure that may occur over time. Actual live trading results may differ materially.

The strategies described trade on Kalshi, a CFTC-regulated prediction market. Prediction market contracts carry unique risks including but not limited to binary loss of principal, liquidity fragmentation, and regulatory uncertainty. The risk parameters shown (max position, max loss, price floors/ceilings) are simulation constraints, not guarantees.

The permutation test referenced in this report used a degraded design: the null histogram was degenerate, and only the edge feed — not the price series — was permuted. Any p-value from this test should be interpreted accordingly. The deflated Sharpe ratio of 0.35 indicates the top result is not statistically distinguishable from the luckiest outcome of a skill-less parameter sweep.

Fees consumed 111% of the winner's gross PnL, meaning the strategy was net negative even before considering other real-world frictions. The parameter axes for entry_low and entry_high were inert during the simulation window, which means the apparent flatness of results across variants reflects unexercised parameters rather than robustness.

This report should not be used as the sole basis for any trading decision. Past performance, including negative past performance, does not guarantee future results.

About this Kalshi strategy backtest

What question does BTC Dual Trend Confirmation investigate?

[M5 remote-plane validation #2 - 4 permutation workers] Multi-horizon BTC trend confirmation predicts repricing in Kalshi KXBTC15M: buy YES when both 5m and 1h BTC returns are positive, buy NO when both are negative. Only enter when price is $0.25-$0.75, spread <= $0.03, and >2m remain. Test whether trend agreement produces positive, repeatable historical performance.

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.

This report is generated from historical simulations. Backtests can be wrong or incomplete, and live trading can differ materially because of liquidity, fees, slippage, latency, market resolution, outages, and data quality. Do your own review before running any strategy.

BTC Dual Trend Confirmation · 077439b07372 | Turbine