Parameter Sweep¶

Parallel grid search across parameter combinations using multiprocessing.

Complete Example¶

from replaybt import CSVProvider, Strategy, MarketOrder, Side
from replaybt.optimize.sweep import ParameterSweep


class EMACrossover(Strategy):
    def configure(self, config):
        self._prev_fast = self._prev_slow = None

    def on_bar(self, bar, indicators, positions):
        fast = indicators.get("ema_fast")
        slow = indicators.get("ema_slow")
        if fast is None or slow is None or self._prev_fast is None:
            self._prev_fast, self._prev_slow = fast, slow
            return None

        crossed_up = fast > slow and self._prev_fast <= self._prev_slow
        self._prev_fast, self._prev_slow = fast, slow

        if not positions and crossed_up:
            return MarketOrder(side=Side.LONG)
        return None


sweep = ParameterSweep(
    strategy_class=EMACrossover,
    data=CSVProvider("ETH_1m.csv", symbol_name="ETH"),
    base_config={
        "initial_equity": 10_000,
        "indicators": {
            "ema_fast": {"type": "ema", "period": 5, "source": "close"},
            "ema_slow": {"type": "ema", "period": 10, "source": "close"},
        },
    },
    param_grid={
        "take_profit_pct": [0.02, 0.04, 0.06, 0.08, 0.10],
        "stop_loss_pct": [0.01, 0.02, 0.03, 0.04],
    },
    n_workers=10,
)

results = sweep.run()
print(results.summary(top_n=10))

Parameters¶

ParameterSweep(
    strategy_class=MyStrategy,    # Strategy CLASS (not instance)
    data=data_provider,           # DataProvider instance
    base_config=config_dict,      # base engine config
    param_grid={                  # parameters to sweep
        "take_profit_pct": [0.04, 0.06, 0.08],
        "stop_loss_pct": [0.02, 0.03, 0.04],
    },
    n_workers=10,                 # parallel workers (default: cpu_count)
)

The sweep evaluates every combination of parameter values. With 5 TP values and 4 SL values, that's 20 combinations evaluated in parallel.

SweepResults¶

Top/Bottom Results¶

results = sweep.run()

# Top 10 by net PnL
best = results.best("net_pnl", n=10)
for combo in best:
    print(f"TP={combo['take_profit_pct']}, SL={combo['stop_loss_pct']} → ${combo['net_pnl']:,.0f}")

# Bottom 5
worst = results.worst("net_pnl", n=5)

Filter¶

# Only combos where TP >= 0.06
filtered = results.filter(take_profit_pct=0.06)
print(filtered.summary())

DataFrame Export¶

df = results.to_dataframe()
print(df.sort_values("net_pnl", ascending=False).head(10))

Available Metrics¶

Each combo dict includes:

Key	Description
`net_pnl`	Net profit/loss
`net_return_pct`	Return percentage
`max_drawdown_pct`	Maximum drawdown
`total_trades`	Number of trades
`win_rate`	Win rate percentage
`profit_factor`	Gross profit / gross loss

Plus all parameter keys from param_grid.

Formatted Summary¶

print(results.summary(metric="net_pnl", top_n=20))

Prints a table of the top 20 combinations sorted by the specified metric.