A mean reversion strategy exploiting first-hour selling pressure on SOL/USDT.
SOLUSDT (Binance) | Aug 2020 – Mar 2025 | 301 Trades
Skewness of +3.39 means the distribution of returns has a heavy right tail. Winning trades are disproportionately larger than losers, creating a convex payoff structure. Even with a 58.8% win rate, the asymmetry drives returns.
T-statistic of 3.322 with a p-value of 0.001. The probability these results occurred by chance is roughly 1 in 1,000. With 301 trades over 4.5 years, this isn't a small-sample artifact.
The 12-hour holding window captures the mean reversion move without overnight exposure risk. After the initial selling pressure dissipates in the first hour, price gravitates back toward fair value during the active trading session.
For every dollar lost, the strategy generates $1.89. Combined with an average winning trade of +4.27% versus average losing trade of -3.23%, the risk-reward profile is structurally favorable over a large sample.
The entry condition is binary and mechanical: first hourly candle (UTC 00:00-01:00) must close red with a decline of 1% or more. No discretion, no ambiguity, no curve-fitting of multiple indicators. One rule.
Sharpe of 3.04 and Sortino of 6.28 indicate strong risk-adjusted performance. The Calmar ratio of 2.89 shows CAGR relative to max drawdown is favorable. These ratios hold across the full 4.5-year sample.
At UTC 01:00, check the 1H candle that just closed (UTC 00:00 to 01:00). This is the first candle of the day in UTC time. You're looking for a single condition: did it close red with a decline of 1.0% or more from open to close?
If the condition is met, enter a long position on SOL/USDT Perpetual at the close of the 1H candle. Position size should be 1-2% of NAV. This is a mean reversion play: the first-hour selling pressure tends to reverse over the following 12 hours.
Close the position at UTC 12:00, exactly 12 hours after the entry candle opened. This is a time-based exit with no stop loss. The full intraday exposure means you ride through any volatility within the 12-hour window.
The backtest uses 1-2% of NAV per trade. The worst trade was -15.52%, and max consecutive losses reached 9. With 67 trades per year, there will be losing streaks. Size appropriately for your drawdown tolerance.
This composite performance record and statistics are hypothetical. Past performance does NOT guarantee future results. The strategy has no stop loss and carries full intraday exposure. The worst single trade lost -15.52%, and the maximum drawdown reached -33%. Market conditions, liquidity, and exchange mechanics can all change. Position size accordingly. This is educational content, not financial advice. Never risk more than you can afford to lose.
The strategy runs without a stop loss. This contributes to the positive skew but means intraday drawdowns can be severe. The -15.52% worst trade happened because position was held through a sharp move. Understand this before deploying.
Kurtosis of +29.48 means extreme returns (both positive and negative) are far more common than a normal distribution would suggest. Fat tails are real. Your +60.5% best trade and -15.52% worst trade illustrate this.
Mean reversion works until it doesn't. If SOL enters a sustained downtrend, the first-hour selling pressure may continue rather than reverse. Monitor whether the edge persists by tracking rolling win rates and expectancy.
With an EV/trade of +1.18%, execution costs matter. Funding rates, taker fees, and slippage on Binance perpetuals can eat into the edge. Factor in realistic transaction costs before going live.
301 trades over 4.5 years is reasonable but not enormous. The T-statistic of 3.322 provides confidence, but continue validating on out-of-sample data. Market structure in crypto evolves fast.
Hypothetical results don't include real-world frictions: exchange downtime, liquidation cascades, withdrawal restrictions, or sudden liquidity drops. Paper trade first. Forward test before allocating real capital.