🧪 One chart: the most common backtesting trap — curve fitting.
Left: smooth upward in-sample results (rules tuned to the past). Right: the same rules out-of-sample — live noise, a different story.
1️⃣ Easy to fit: add 15 conditions and the past looks perfect
2️⃣ But that’s not a strategy, it’s memorization — future data won’t match that shape
3️⃣ Fix: out-of-sample testing, then paper trading
4️⃣ Sample size: 20 trades is not statistics
A backtest gives you a candidate, not certainty. Illustration — not real data.
$BTC #CryptoEducation #Backtesting #TradingStrategy
Left: smooth upward in-sample results (rules tuned to the past). Right: the same rules out-of-sample — live noise, a different story.
1️⃣ Easy to fit: add 15 conditions and the past looks perfect
2️⃣ But that’s not a strategy, it’s memorization — future data won’t match that shape
3️⃣ Fix: out-of-sample testing, then paper trading
4️⃣ Sample size: 20 trades is not statistics
A backtest gives you a candidate, not certainty. Illustration — not real data.
$BTC #CryptoEducation #Backtesting #TradingStrategy
