A strategy with a win rate of 37.96% lost 21.6%, while one with a win rate of 12.86% actually made money—I’ll show you using 22,786 K-lines
First, let me lay out the data: BTC/USDT 1-hour timeframe, from 2024-01-16 to 2026-08-22, 22,786 K-lines, starting capital of $10,000, 10% position size, and a 0.1% one-way trading fee. I ran 6 common strategies. The most counterintuitive pair in the results:
• RSI mean reversion: win rate 37.96% (the highest among the six). Ending value $7,839, down 21.6%
• Turtle trading: win rate 12.86% (the lowest among the six). Ending value $10,227, +2.27%, the only strategy with positive returns
The highest win rate lost the most, but the lowest win rate ended up making money. Why?
Because win rate only answers “are there more winning trades?” It does not answer “how much do you make when you win, and how much do you lose when you lose.” RSI is mean reversion: it looks good because you take small profits and stop early, with a decent win probability—but each trade only grabs a few cents. If the market moves strongly in one direction, one big stop loss can wipe out dozens of small gains. Turtle trading is trend following. 90% of the time it’s making small losses while testing; once it catches the trend, it takes a big bite.
What you truly should look at is the expectation value: expectation = win rate × average profit − loss rate × average loss. Even with a win rate of 80%, if you win 1 and lose 5 per trade on average, the expectation is still negative.
Let AI calculate it in one minute: dump the trade records into it, give it one instruction—“output three columns: win rate, average profit per trade, and average loss per trade, then calculate the expectation value.” If expectation is negative, don’t put real money in just because the win rate looks pretty.
What’s your strategy’s win rate? How much do you typically make or lose per trade on average? Post it in the comments—I’ll help you calculate the expectation value.
$BTC #crypto
First, let me lay out the data: BTC/USDT 1-hour timeframe, from 2024-01-16 to 2026-08-22, 22,786 K-lines, starting capital of $10,000, 10% position size, and a 0.1% one-way trading fee. I ran 6 common strategies. The most counterintuitive pair in the results:
• RSI mean reversion: win rate 37.96% (the highest among the six). Ending value $7,839, down 21.6%
• Turtle trading: win rate 12.86% (the lowest among the six). Ending value $10,227, +2.27%, the only strategy with positive returns
The highest win rate lost the most, but the lowest win rate ended up making money. Why?
Because win rate only answers “are there more winning trades?” It does not answer “how much do you make when you win, and how much do you lose when you lose.” RSI is mean reversion: it looks good because you take small profits and stop early, with a decent win probability—but each trade only grabs a few cents. If the market moves strongly in one direction, one big stop loss can wipe out dozens of small gains. Turtle trading is trend following. 90% of the time it’s making small losses while testing; once it catches the trend, it takes a big bite.
What you truly should look at is the expectation value: expectation = win rate × average profit − loss rate × average loss. Even with a win rate of 80%, if you win 1 and lose 5 per trade on average, the expectation is still negative.
Let AI calculate it in one minute: dump the trade records into it, give it one instruction—“output three columns: win rate, average profit per trade, and average loss per trade, then calculate the expectation value.” If expectation is negative, don’t put real money in just because the win rate looks pretty.
What’s your strategy’s win rate? How much do you typically make or lose per trade on average? Post it in the comments—I’ll help you calculate the expectation value.
$BTC #crypto