Research Report: SOL 15-Minute Momentum Strategy

Short Disclaimer

Historical simulation only. No guarantee of future performance. This is not investment advice.

Intro / Thesis

We tested whether near-expiry momentum in SOL's spot market could predict the resolution direction of Kalshi's 15-minute SOL markets. The core idea was simple: if SOL has been moving up over the last five minutes, showing positive one-minute velocity, and BTC is also showing positive one-minute velocity, the YES side of an uncertain SOL 15-minute market may be underpriced during the final five minutes. The opposite set of signals would favor NO.

Variant and Strategy Explanation

The strategy runs on a 10-second loop against Kalshi's KXSOL15M series. It waits until the market has five minutes or less to expiry. During that window, it looks for three conditions to align:

• SOL's 5-minute change is positive (for YES) or negative (for NO)
• SOL's 1-minute velocity is positive (for YES) or negative (for NO)
• BTC's 1-minute velocity is positive (for YES) or negative (for NO)

It only enters when the market price is between 0.45 and 0.55, meaning the market is still genuinely uncertain. Position size is $50 per contract side, with a stop-loss at -$7.50 unrealized P&L and a hard exit five seconds before expiry.

We swept two parameters: price floor (0.05 to 0.45) and price ceiling (0.55 to 0.95). All 100 combinations completed successfully. Each successful variant is saved as a runnable Turbine strategy.

Top Results

The top eight variants all shared the same ceiling of 0.68, with floors ranging from 0.05 to 0.36. They delivered identical results because the entry price band was wide enough at those floors that no additional trades were blocked.

Best variant: SOL strategy · floor 0.05 / ceil 0.68

• ROI: 177.84%
• Total P&L: $88.92
• Trades: 146
• Win rate: 61.7%
• Max drawdown: -$30.16
• Sharpe: 0.36

The winning parameter combination was a ceiling of 0.68, meaning the strategy refused to buy contracts priced above 0.68. This is notable — it deliberately avoided the most confident-looking setups and focused only on mid-range uncertainty.

Bottom Results

The worst variants all used the base ceiling of 0.95, allowing entry into expensive contracts up to $0.95. The absolute worst was:

Worst variant: SOL strategy · floor 0.45 / ceil 0.95

• ROI: -199.40%
• Total P&L: -$99.70
• Trades: 317
• Win rate: 65.9%
• Max drawdown: -$100.44
• Sharpe: -0.07

The pattern is clear: higher ceilings allowed more trades, but those additional trades were often expensive contracts near $0.90+. When those lost, the per-trade loss was much larger than the per-trade gain from the mid-range contracts the winners focused on. The worst variant actually had a slightly higher win rate than the best variant — 65.9% vs 61.7% — but its losses dwarfed its wins.

Conclusion

The raw numbers for the top variants look attractive: $88.92 total P&L, 61.7% win rate, and a 177.84% ROI. The parameter sweep shows a consistent pattern where lower price ceilings performed dramatically better than higher ones.

However, the robustness statistics tell a more sobering story. The deflated Sharpe ratio came in at 0.56, which is below the 0.95 threshold needed to distinguish the winner from selection noise. The neighborhood degradation of 0.19 means results dropped off relatively quickly around the winning cell.

The permutation test did run, and it produced a p-value of 0.040. That means the strategy's edge-feed timing beat 96% of time-scrambled versions of the same feed. While this clears the 0.05 threshold, it's not overwhelming evidence. The permutation test only scrambled the edge feed timing, not the market price series, so price-based conditions were not fully stress-tested.

Bottom line: the top results are consistent with selection noise and overfitting. The parameter sweep succeeded in finding a configuration that worked well in this historical window, but the statistical evidence is too weak to call this validated or promising. The strategy is not broken — it simply hasn't demonstrated an edge that we can confidently separate from luck.

Long Disclaimer

This report is a historical simulation research document produced by Turbine's research team. It describes parameter sweeps and statistical tests conducted on historical market data. Nothing in this report should be interpreted as a prediction of future performance, a recommendation to trade, or an endorsement of any strategy.

Key limitations:

• Historical simulation only. All performance figures are based on backtested data. Markets change, and strategies that worked in a specific historical window may perform differently — including losing money — in live trading.
• Selection bias. The parameter sweep tested 100 combinations. When many variants are tested, some will look good purely by chance. The statistical tests in this report attempt to quantify that risk but cannot eliminate it.
• Overfitting risk. The deflated Sharpe ratio (0.56) indicates the top result is not statistically distinguishable from the luckiest outcome of a skill-less sweep. The permutation p-value (0.040) provides weak-to-moderate evidence, but does not account for all forms of data snooping.
• Costs and slippage not modeled. The simulation does not include trading fees, spreads, or slippage that would occur in live markets.
• No future guarantees. Past performance, even when statistically robust, does not guarantee future results.

This report is provided "as is" for informational purposes only. Turbine makes no representations or warranties regarding the accuracy or completeness of the data. Trading involves substantial risk of loss. Always do your own research and consider your own risk tolerance before making trading decisions.