Sizing¶

PositionSizer (ABC)¶

Abstract base for position sizing strategies.

from replaybt import PositionSizer

class MySizer(PositionSizer):
    def get_size(self, equity, side, price, symbol="", stop_loss_pct=0.0):
        return equity * 0.1  # 10% of equity

Abstract Method¶

def get_size(
    self,
    equity: float,       # current portfolio equity
    side: str,            # "LONG" or "SHORT"
    price: float,         # current price
    symbol: str = "",     # asset symbol
    stop_loss_pct: float = 0.0,  # SL distance (for risk-based sizing)
) -> float:  # position size in USD

FixedSizer¶

Constant USD position size.

from replaybt import FixedSizer

sizer = FixedSizer(size_usd=10_000)

Parameter	Type	Default	Description
`size_usd`	`float`	`10_000`	Fixed size

EquityPctSizer¶

Size as a percentage of current equity.

from replaybt import EquityPctSizer

sizer = EquityPctSizer(pct=0.10, min_size=100, max_size=50_000)

Parameter	Type	Default	Description
`pct`	`float`	`0.10`	Fraction of equity (0.10 = 10%)
`min_size`	`float`	`100`	Minimum position size
`max_size`	`float`	`0`	Maximum size (0 = no cap)

RiskPctSizer¶

Size so that a stop-loss hit loses at most risk_pct of equity.

Formula: size = (equity * risk_pct) / stop_loss_pct

from replaybt import RiskPctSizer

sizer = RiskPctSizer(risk_pct=0.01, default_sl_pct=0.035)

Parameter	Type	Default	Description
`risk_pct`	`float`	`0.01`	Max equity at risk per trade (1%)
`min_size`	`float`	`100`	Minimum position size
`max_size`	`float`	`0`	Maximum size (0 = no cap)
`default_sl_pct`	`float`	`0.035`	Default SL if none provided

Example: With $10,000 equity, 1% risk, and 3.5% SL:

size = ($10,000 * 0.01) / 0.035 = $2,857

If the trade hits SL, you lose $2,857 * 3.5% = $100 = 1% of equity.

KellySizer¶

Position sizing using the Kelly criterion.

Formula: f* = win_rate - (1 - win_rate) / payoff_ratio

from replaybt import KellySizer

sizer = KellySizer(
    win_rate=0.60,
    avg_win=0.08,
    avg_loss=0.035,
    fraction=0.25,  # quarter-Kelly
)

Parameter	Type	Default	Description
`win_rate`	`float`	`0.60`	Historical win rate (0-1)
`avg_win`	`float`	`0.08`	Avg winning trade %
`avg_loss`	`float`	`0.035`	Avg losing trade %
`fraction`	`float`	`0.25`	Kelly fraction (1.0 = full Kelly)
`min_size`	`float`	`100`	Minimum size
`max_size`	`float`	`0`	Maximum size (0 = no cap)
`max_equity_pct`	`float`	`0.25`	Max fraction of equity

Properties¶

Property	Type	Description
`kelly_fraction`	`float`	Raw Kelly before fractional multiplier

Using a Sizer¶

Pass to the engine config:

engine = BacktestEngine(
    strategy=my_strategy,
    data=data,
    config={
        "initial_equity": 10_000,
        "sizer": RiskPctSizer(risk_pct=0.01),
    },
)

When a sizer is set, default_size_usd is ignored and the sizer determines position size for each trade.