Position Sizing — 7 Modes

Position sizing is the single configuration decision with the greatest impact on your backtest's risk profile. Two strategies with identical entry and exit logic can produce opposite outcomes — one compounding steadily, another blowing up — based solely on how much capital each trade consumes. StratBase.ai gives you seven distinct models, each solving a different problem.

Every time your strategy generates an entry signal, the engine calculates the position size before executing the trade. The calculation runs against your current simulated balance, not the opening balance, so growth and drawdowns feed into sizing dynamically wherever the mode supports it.

Position sizing is set during backtest creation. To access it:

1. Open New Backtest from the dashboard
2. Expand the Capital & Risk panel
3. Select a sizing mode from the Position Size dropdown
4. Fill in the parameters that appear below the selector
5. Confirm under Review & Launch

You cannot edit sizing on a saved backtest — clone it first (⋯ → Clone), then adjust the cloned copy.

1. Fixed Amount

Every trade allocates the same absolute sum regardless of account balance or volatility.

When to use: Simple strategies where you want total predictability in exposure per trade.

Parameters: | Field | Example Value | |-------|--------------| | Amount (USD) | $500 |

A $500 trade on BTC/USDT at $60,000 buys 0.00833 BTC. If your account doubles, the trade size stays $500 — no compounding.

2. Fixed Units

Trades a constant quantity of the asset rather than a dollar value.

When to use: Futures contracts with standardized sizes, or when you want exposure in asset terms rather than currency terms.

Parameters: | Field | Example Value | |-------|--------------| | Quantity | 0.1 ETH |

Useful when you're benchmarking returns in ETH rather than USD.

3. Percentage of Account

Allocates a fraction of the current balance at signal time.

When to use: The default choice for most strategies. Wins compound naturally; drawdowns reduce exposure automatically.

Parameters: | Field | Example Value | |-------|--------------| | Percentage | 15% |

Starting balance $8,000 → first trade = $1,200. After five winning trades the balance reaches $11,000 → next trade = $1,650. No manual adjustment needed.

4. ATR-Based

Scales position size inversely to recent volatility, keeping dollar risk per trade approximately constant across changing market conditions.

Formula: Size = (Balance × Risk%) ÷ (ATR × Multiplier)

Parameters: | Field | Example Value | |-------|--------------| | ATR Period | 14 | | Risk per Trade | 1.5% | | ATR Multiplier | 2.0 |

If BTC/USDT has a 14-period ATR of $1,800 and your balance is $10,000: Size = ($10,000 × 0.015) ÷ ($1,800 × 2.0) = $150 ÷ $3,600 ≈ 0.0417 BTC

During a quiet market where ATR drops to $900, the size would double — capturing more exposure when conditions are calmer.

5. Kelly Criterion

Derives the theoretically optimal bet size from your strategy's historical win rate and payoff ratio.

Formula: Kelly% = W − (1−W) ÷ (AvgWin ÷ AvgLoss)

Parameters: | Field | Example Value | |-------|--------------| | Kelly Fraction | 0.5 |

A strategy with 55% win rate and 1.4 average win/loss ratio produces a Kelly% of roughly 16%. At half-Kelly (fraction = 0.5), each trade uses 8% of balance. StratBase.ai calculates W and the payoff ratio from your backtest's rolling performance, so the size updates as results accumulate.

Full Kelly maximizes long-run growth mathematically but tolerates large interim drawdowns. Half-Kelly is the standard practitioner choice.

6. Risk-Based

Works backward from a maximum acceptable loss: given your stop-loss distance, it calculates exactly how large the position must be so that a stop-out costs precisely X% of your account.

Formula: Size = (Balance × Risk%) ÷ SL Distance%

Parameters: | Field | Example Value | |-------|--------------| | Max Risk per Trade | 1% | | Stop-Loss | Required — set in Stop Loss panel |

Balance $12,000 · 1% risk · 2.5% stop-loss distance: Size = ($12,000 × 0.01) ÷ 0.025 = $4,800

If the stop triggers, you lose exactly $120 — 1% of balance. This mode is disabled unless a stop-loss is configured.

7. Recovery (Martingale & Variants)

Increases position size after consecutive losses to accelerate balance recovery.

Parameters: | Field | Example Value | |-------|--------------| | Base Size | 10% of account | | Progression | Martingale / Linear / Custom | | Multiplier | 2.0 | | Max Steps | 4 |

With Martingale at multiplier 2.0 and base 10%: loss 1 → 10%, loss 2 → 20%, loss 3 → 40%, loss 4 → 80%. A four-loss streak at 80% wipes nearly all equity. Always set Max Steps. See Recovery Mode for the Linear and Custom progressions.

Recovery modes are unsuitable for assets with persistent trends or high gap risk. Test extensively on bear-market date ranges before using on live capital.

| Situation | Recommended Mode | |-----------|-----------------| | First backtest, unknown edge | % of Account (10–15%) | | Stop-loss is precisely defined | Risk-Based (1–2%) | | High-volatility asset, variable ATR | ATR-Based | | Strategy with 200+ backtested trades | Kelly at half-fraction | | Futures with contract minimums | Fixed Units | | DCA or grid accumulation | Fixed Amount | | Tested mean-reversion with max step limit | Recovery (Linear) |

Can I run the same strategy with two different sizing modes to compare? Yes — clone your backtest, change only the sizing mode, and use the Compare view to place the equity curves side by side on the same chart.

Does the engine account for insufficient balance? Yes. If a calculated size exceeds the available balance, StratBase.ai caps the trade at 100% of remaining capital and logs a warning in the Trade Log tab.

Will Kelly sizing change mid-backtest? Yes. The Kelly calculation refreshes after every closed trade using the running win rate and payoff ratio, so early trades in a short backtest may behave differently than trades 200 bars in.

• Recovery Mode — Full Configuration
• Stop Loss Configuration
• Key Metrics Explained
• Grid Entry (DCA) Setup