Position sizing strategies: fixed-fractional, ATR and Kelly
Most traders obsess over entries and ignore the one decision that actually determines survival: how much to bet on each trade. Position sizing turns a raw signal into a quantity of contracts or coins, and it is what keeps a losing streak from wiping you out. Bet too big and a normal run of losses ends the account; bet too small and a real edge barely compounds. This guide covers the three sizing frameworks every bot should know — the fixed-fractional 1% rule, volatility (ATR) sizing, and the Kelly criterion — with the code for each and why sizing matters more than your indicator.
Why sizing beats entries
Two bots with the identical signal can have opposite fates purely from sizing. The one that risks 20% per trade is one bad streak from zero; the one that risks 1% survives the same streak and compounds the edge. Sizing is the link between a strategy on paper and an account that is still alive next year — it is the operational core of risk management, and it matters more than the entry rule you spent weeks tuning.
Fixed-fractional — the 1% rule
The workhorse of retail sizing: risk a fixed fraction of equity (commonly 1%) on every trade, where “risk” is the distance from entry to your stop. Because the risk is measured to the stop, a wide stop gives a small position and a tight stop a large one — every trade risks the same dollar amount regardless of where the stop sits. As the account grows the dollar risk grows with it; as it shrinks, the risk shrinks, which dampens drawdowns automatically.
Volatility (ATR) sizing
A refinement that sets the stop distance itself from market volatility. Instead of an arbitrary percentage, the stop is a multiple of the Average True Range, so a volatile coin gets a wider stop and a smaller position while a calm one gets a tighter stop and a larger position. The dollar risk stays constant at 1%, but the position scales inversely with volatility — you take the same risk on a wild market and a quiet one. This is the sizing behind most serious trend systems.
The Kelly criterion
Kelly is the mathematically optimal fraction to bet for fastest long-run growth, derived from your win rate and reward-to-risk. Full Kelly maximises growth but produces brutal, stomach-churning drawdowns and is hypersensitive to estimation error — overstate your edge and Kelly tells you to bet ruinously large. In practice serious traders use fractional Kelly (a quarter or a half), which captures most of the growth with a fraction of the volatility.
Sizing in code
python · sizing.pydef fixed_fractional(equity, risk_pct, entry, stop):
risk_per_unit = abs(entry - stop)
dollar_risk = equity * risk_pct # e.g. 0.01 = 1%
return dollar_risk / risk_per_unit # units / coins
def kelly_fraction(win_rate, reward_risk):
# f* = W - (1-W)/R · use a quarter of this live
f = win_rate - (1 - win_rate) / reward_risk
return max(0.0, f) * 0.25
The interactive position-sizing calculator runs the fixed-fractional formula for you from account size, risk percent, entry and stop.
Which strategy to use
Beginners should use the fixed-fractional 1% rule — it is simple, robust and forgiving of a mis-estimated edge. Graduate to ATR sizing once you want the position to respect volatility. Treat full Kelly as a theoretical ceiling, not a setting: it assumes you know your true win rate (you don’t) and its drawdowns will shake you out. Quarter-Kelly is the most aggressive any disciplined bot should run.
Frequently asked questions
What is the most important factor in position sizing?
Survival. Position sizing determines how much a losing streak costs you, so betting small enough to survive the inevitable run of losses matters more than the entry signal. Two bots with the identical strategy can have opposite outcomes purely because one risked 1% per trade and the other risked 20%.
What is the fixed-fractional or 1% rule?
The fixed-fractional rule risks a constant fraction of equity — commonly 1% — on every trade, measured as the distance from entry to the stop-loss. A wider stop produces a smaller position and a tighter stop a larger one, so every trade risks the same dollar amount, and the risk scales automatically with the account balance.
How does ATR position sizing work?
ATR sizing sets the stop distance as a multiple of the Average True Range, so a volatile asset gets a wider stop and a smaller position while a calm asset gets a tighter stop and a larger one. The dollar risk stays fixed at, say, 1%, but the number of units scales inversely with volatility, equalising risk across markets.
Should I use the Kelly criterion to size positions?
Only a fraction of it. Full Kelly maximises long-run growth but produces violent drawdowns and is hypersensitive to over-estimating your edge, which can lead to ruinous bets. Most disciplined traders use quarter- or half-Kelly, capturing most of the growth with far less volatility, and beginners are better off with the simpler fixed-fractional 1% rule.