Up or Down About

Strategy backtest: Alphabet (Google)

Each model's prediction at today's close sets tomorrow's position (long or short). Returns are daily log returns over the test period (9 May 2022 to 25 Sep 2026), net of costs charged on every position change.

Verdict

0 of 7 strategies beat buy & hold on Sharpe ratio

With hindsight, the best on the test period is Extra Trees at 0.72 against 0.80 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 +174.3% +26.0% 32.1% 0.72 1.06 33.5% 0.78 52.3% 43 4.25%
Logistic Regression +126.9% +20.7% 32.1% 0.58 0.84 23.9% 0.87 52.1% 292 29.15%
Linear SVC +95.4% +16.6% 32.1% 0.48 0.68 22.8% 0.73 52.1% 290 28.95%
Random Forest +65.8% +12.3% 32.1% 0.36 0.51 40.7% 0.30 51.2% 176 17.55%
SVM (RBF) +31.8% +6.5% 32.1% 0.20 0.27 45.9% 0.14 51.4% 382 38.15%
Bagging (KNN) +21.9% +4.7% 32.1% 0.14 0.20 44.7% 0.10 50.2% 417 41.65%
Gradient Boosting -55.0% -16.7% 32.1% -0.57 -0.79 69.4% -0.24 47.8% 471 47.05%
Buy & hold +207.4% +29.3% 32.1% 0.80 1.17 31.7% 0.93 52.8% 1 0.05%

Costs and risk

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

Takeaway With free trading, none of 7 strategies beat buy & hold on Sharpe; at 50 bp, none do. Extra Trees changed position 43 times in 1,100 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