diff --git "a/backtests.html" "b/backtests.html" --- "a/backtests.html" +++ "b/backtests.html" @@ -1 +1 @@ -Jolash

Research · Validation

Backtest Results

Honest, cost-inclusive results on held-out data the model never trained on. No cherry-picking.

USDCHF — 2024-01-02 -> 2025-12-31 (held out, never trained on)

Outperformed buy & hold

Test bars

12,453

Trades taken

4,275

Strategy return

61.67%

Buy & hold return

-5.50%

Sharpe (annualized)

1.29

Max drawdown

-15.91%

Win rate 54.8%Positive Sharpe and beat buy-and-hold -- a real candidate worth further scrutiny, not yet a confirmed edge on its own.
EURUSD — 2024-01-02 -> 2025-12-31 (held out, never trained on)

Underperformed buy & hold

Test bars

12,454

Trades taken

3,816

Strategy return

-24.47%

Buy & hold return

6.26%

Sharpe (annualized)

-0.86

Max drawdown

-32.94%

Win rate 48.7%Negative risk-adjusted return, consistent with the win rate above. Confirms no exploitable edge for this pair/horizon after realistic spread costs.
GBPUSD — 2024-01-02 -> 2025-12-31 (held out, never trained on)

Underperformed buy & hold

Test bars

12,451

Trades taken

1,784

Strategy return

4.88%

Buy & hold return

5.70%

Sharpe (annualized)

0.29

Max drawdown

-9.02%

Win rate 51.8%Technically positive Sharpe, but a simple buy-and-hold still did better -- not compelling enough to pursue as-is.
AUDUSD — 2024-01-02 -> 2025-12-31 (held out, never trained on)

Underperformed buy & hold

Test bars

12,454

Trades taken

3,185

Strategy return

-12.12%

Buy & hold return

-1.99%

Sharpe (annualized)

-0.30

Max drawdown

-26.70%

Win rate 50.2%Negative risk-adjusted return, consistent with the win rate above. Confirms no exploitable edge for this pair/horizon after realistic spread costs.
\ No newline at end of file +Jolash

Research · Validation

Backtest Results

Honest, cost-inclusive results on held-out data the model never trained on. No cherry-picking.

USDCHF — Held out test set, never trained on (1h, horizon 4, submitted 2026-07-09) -- The one real candidate -- also confirmed on a separate 2021-22 window (Sharpe 0.51)

Outperformed buy & hold

Test bars

12,453

Trades taken

4,275

Strategy return

61.67%

Buy & hold return

-5.50%

Sharpe (annualized)

1.29

Max drawdown

-15.91%

Win rate 54.8%Positive Sharpe and beat buy-and-hold -- a real candidate worth further scrutiny, not yet a confirmed edge on its own.
GBPUSD — Held out test set, never trained on (1h, horizon 4, submitted 2026-07-09) -- Positive Sharpe but underperforms buy-and-hold

Underperformed buy & hold

Test bars

12,451

Trades taken

1,784

Strategy return

4.88%

Buy & hold return

5.70%

Sharpe (annualized)

0.29

Max drawdown

-9.02%

Win rate 51.8%Technically positive Sharpe, but a simple buy-and-hold still did better -- not compelling enough to pursue as-is.
EURUSD — Held out test set, never trained on (1h, horizon 12, submitted 2026-07-09) -- Testing the new auto-push flow end to end

Underperformed buy & hold

Test bars

12,446

Trades taken

4,293

Strategy return

-12.39%

Buy & hold return

6.25%

Sharpe (annualized)

-0.19

Max drawdown

-53.39%

Win rate 50.4%Negative risk-adjusted return, consistent with the win rate above. Confirms no exploitable edge for this pair/horizon after realistic spread costs.
AUDUSD — Held out test set, never trained on (1h, horizon 4, submitted 2026-07-09) -- Batch cheap-screen promoted this; full validation says no

Underperformed buy & hold

Test bars

12,454

Trades taken

3,185

Strategy return

-12.12%

Buy & hold return

-1.99%

Sharpe (annualized)

-0.30

Max drawdown

-26.70%

Win rate 50.2%Negative risk-adjusted return, consistent with the win rate above. Confirms no exploitable edge for this pair/horizon after realistic spread costs.
EURUSD — Held out test set, never trained on (1h, horizon 4, submitted 2026-07-09) -- Original hypothesis, basic technical indicators -- confirmed no edge

Underperformed buy & hold

Test bars

12,454

Trades taken

3,816

Strategy return

-24.47%

Buy & hold return

6.26%

Sharpe (annualized)

-0.86

Max drawdown

-32.94%

Win rate 48.7%Negative risk-adjusted return, consistent with the win rate above. Confirms no exploitable edge for this pair/horizon after realistic spread costs.
\ No newline at end of file