Up or Down About

Strategy backtest: Microsoft

Each model's prediction at today's close sets tomorrow's position (long or short). Returns are daily log returns over the test period (7 Jun 2021 to 24 Sep 2026), net of costs charged on every position change.

Verdict

2 of 7 strategies beat buy & hold on Sharpe ratio

With hindsight, the best on the test period is Extra Trees at 0.62 against 0.52 for buy & hold. Drag the cost slider to see how quickly turnover erodes any edge.

Risk and return

Annualised with 252 trading days. Sharpe and Sortino assume a zero risk-free rate; Calmar = annual return รท max drawdown.

StrategyTotal returnAnnual returnVolatilitySharpeSortinoMax drawdownCalmarHit rateTradesCosts
Extra Trees +145.1% +18.5% 27.5% 0.62 0.96 31.4% 0.59 50.1% 369 36.85%
Gradient Boosting +116.6% +15.7% 27.6% 0.53 0.80 37.3% 0.42 49.1% 562 56.15%
Bagging (KNN) +51.2% +8.1% 27.5% 0.28 0.43 36.7% 0.22 49.1% 603 60.25%
Logistic Regression +24.7% +4.3% 27.6% 0.15 0.23 41.7% 0.10 49.5% 485 48.45%
Random Forest -2.6% -0.5% 27.6% -0.02 -0.03 54.8% -0.01 49.5% 548 54.75%
Linear SVC -3.9% -0.8% 27.6% -0.03 -0.04 49.4% -0.02 48.1% 431 43.05%
SVM (RBF) -48.0% -11.6% 27.6% -0.45 -0.63 65.7% -0.18 48.6% 508 50.75%
Buy & hold +112.2% +15.3% 27.5% 0.52 0.76 37.1% 0.41 52.0% 1 0.05%

Costs and risk

Would it survive real trading, and was the return worth the risk?

Takeaway With free trading, 3 of 7 strategies beat buy & hold on Sharpe; at 50 bp, none do. Extra Trees changed position 369 times in 1,332 trading days.

Takeaway In long/short mode every strategy is always fully invested (long or short), so they all carry the stock's volatility. Only the return differs.

Inside one strategy

Change the model with the "Detail model" selector