Every binary options trader makes position sizing decisions on every trade, whether they have an explicit system or not. Trading "whatever feels right" is itself a position sizing approach -- it is just one with unpredictable and typically unfavorable properties.
The two main structured approaches are fixed stake sizing (every trade gets the same dollar amount or the same percentage of balance) and variable stake sizing (adjusting your stake based on assessed setup quality, market conditions, or recent performance). Understanding how each behaves mathematically tells you which to use and under what circumstances.
TL;DR -- Key Takeaways
- Fixed percentage sizing (e.g., 1% of current balance per trade) automatically scales down during drawdowns and up during growth -- without any decision required.
- Fixed dollar sizing is simpler but leaves your risk percentage unanchored as your balance changes.
- Variable sizing based on setup quality can improve outcomes -- but only if you have verified data showing your confidence ratings actually predict trade outcomes.
- Variable sizing based on recent performance (e.g., betting more after wins) is Anti-Martingale if done correctly and has different properties than Martingale.
- For most retail traders, fixed percentage sizing is the correct default until a significant verified track record justifies variable adjustments.
1. Fixed Dollar Sizing: Simple but Incomplete
Fixed dollar sizing means placing the same dollar amount on every trade. If you decide to trade $15 per trade, that amount stays constant regardless of whether your account grows from $1,000 to $2,000 or shrinks to $500.
The appeal is simplicity. The calculation is trivial, there are no decisions to make per trade, and the dollar amount directly matches your intuitive sense of what you are risking.
The problem is that the implied risk percentage changes dramatically as your balance changes:
| Account Balance | Fixed Stake | Implied Risk % |
|---|---|---|
| $500 | $15 | 3.0% |
| $750 | $15 | 2.0% |
| $1,000 | $15 | 1.5% |
| $1,500 | $15 | 1.0% |
| $2,500 | $15 | 0.6% |
At $500, $15 is aggressive. At $2,500, $15 is extremely conservative. The same dollar amount produces very different risk profiles depending on where your balance sits, and the profile is most aggressive at exactly the wrong time -- when your balance is low after a drawdown.
2. Fixed Percentage Sizing: The Standard Approach
Fixed percentage sizing (also called fixed fractional sizing) means risking a set percentage of your current balance on every trade. The dollar amount changes as your balance changes, but the risk fraction stays constant.
At 1% per trade:
- $1,000 balance: $10 stake
- After 5 losses: ~$951 balance, $9.51 stake
- After 5 wins from $1,000: ~$1,040 balance, $10.40 stake
The automatic adjustment has two important properties:
During drawdowns, stakes shrink. This means losses during a bad stretch are smaller in absolute dollar terms, reducing the speed at which the balance erodes. A 10-loss streak at 1% reduces the balance by approximately 9.5%, not 10%, because each subsequent loss is on a slightly smaller balance.
During growth, stakes expand. This enables genuine compounding. A 10-win streak at 80% payout and 1% risk increases the balance by approximately 8.6% rather than a flat 8%, because each subsequent win is on a slightly larger balance.
The mathematical effect of these two properties over large trade samples is meaningful: fixed percentage sizing produces better long-run outcomes than flat dollar sizing for any strategy with a genuine positive expected value.
3. Variable Sizing Based on Setup Quality
Variable sizing assigns different stake sizes to different trade setups based on assessed quality or confidence. A trader using variable sizing might place 1% on standard setups, 1.5% on high-confluence setups, and 0.5% on marginal but acceptable setups.
The logic is sound: if some setups genuinely have a higher probability of winning than others, it makes sense to bet more on them. The mathematical basis is the Kelly Criterion, which calculates the optimal bet fraction as a function of edge: the higher your edge on a trade, the larger the optimal bet.
The practical requirement that most traders cannot meet: you need verified data showing that your confidence ratings or setup tiers actually predict different win rates.
The self-assessment problem: If you review your trade journal and find that your "high confidence" setups have a 58% win rate and your "standard" setups have a 55% win rate -- a measurable but small difference -- the additional complexity of variable sizing generates minimal benefit for significant additional decision-making burden. If your "high confidence" setups have a 65% win rate and your "standard" setups have a 52% win rate -- a meaningful difference that can only be verified over 200+ trades -- variable sizing begins to make mathematical sense.
Without this data, variable sizing based on confidence is essentially just the Martingale intuition in a different form: trusting your subjective assessment of when to bet more.
4. Anti-Martingale Variable Sizing
Anti-Martingale (also called reverse Martingale) is a variable sizing approach that increases stakes after wins rather than after losses. It is sometimes dismissed as equally irrational to Martingale, but the mathematics are different in an important way.
Standard Martingale: Bet more after losses, less after wins. Produces a smooth equity curve punctuated by catastrophic losing streaks that produce ruin.
Anti-Martingale: Bet more after wins, return to baseline after losses. Produces a more volatile equity curve but without the ruin dynamic. Catastrophic losses are bounded by your return to baseline after each loss.
A simple Anti-Martingale rule for binary options:
- Start each session at your standard 1% risk per trade
- After 3 consecutive wins, increase to 1.5% for the next trade only, then return to 1%
- After any loss, return immediately to 1%
This approach captures a small amount of upside during winning runs without creating the correlated downside of Martingale. It is not a significant alpha generator, but it avoids the ruin properties of loss-chasing systems.
The key distinction from Martingale: the maximum loss is still bounded by your standard risk percentage because you always return to baseline after a loss. Anti-Martingale cannot produce the catastrophic sequence where each loss requires a larger-than-account recovery stake.
5. Variable Sizing Based on Market Conditions
A different type of variable sizing adjusts stakes based on assessed market conditions rather than recent trade outcomes. This is distinct from confidence-based sizing and from outcome-based sizing -- it responds to the trading environment rather than to recent results.
Examples:
- Reduce to 0.5% per trade during high-impact news release windows (higher uncertainty, strategy signals less reliable)
- Increase to 1.5% per trade on specific high-payout assets where your historical win rate is demonstrably higher
- Reduce to 0.5% when trading outside your primary session hours (e.g., your strategy is calibrated for London session, but you are trading during the Asian session)
Condition-based variable sizing requires the same prerequisite as quality-based variable sizing: data from your trade log showing that your win rate actually differs meaningfully across conditions. Without that data, it is assumption-based rather than evidence-based.
6. Comparing Approaches Over 200 Trades
To illustrate the differences concretely, consider a trader with a 58% win rate, 80% payout, and $1,000 starting balance:
| Approach | Stakes | Ending Balance (Theoretical) | Max Drawdown (Estimated) |
|---|---|---|---|
| Fixed dollar ($10) | $10 constant | ~$1,088 | ~$50-$80 |
| Fixed percentage (1%) | 1% of balance | ~$1,094 | ~$45-$75 |
| Flat 2% | $20 constant | ~$1,176 | ~$120-$160 |
| Fixed percentage (2%) | 2% of balance | ~$1,188 | ~$90-$130 |
The differences between fixed dollar and fixed percentage at comparable risk levels are small but favor percentage sizing due to compounding. The more significant observation is that the higher risk level (2%) produces meaningfully larger returns alongside meaningfully larger drawdowns -- this is the core position sizing trade-off: more risk per trade accelerates both growth and loss.
Variable sizing strategies produce wider spreads around these theoretical expectations depending on how well the variable rules actually identify higher-quality setups. With no real predictive value, variable sizing is just noise around the fixed-percentage outcome. With genuine predictive value, it outperforms fixed sizing -- but demonstrating that predictive value requires a substantial verified trade history.
7. Which Approach to Use: A Decision Framework
| Trader Profile | Recommended Approach |
|---|---|
| Fewer than 100 live trades | Fixed percentage, 1% per trade. No variation. |
| 100-200 live trades, no strategy refinement yet | Fixed percentage, 1-1.5% per trade. Review at 200 trades. |
| 200+ live trades, verified positive EV | Fixed percentage with optional condition-based variation if win rate data supports it. |
| 200+ live trades, verified setup-quality differences | Structured variable sizing with defined tiers, each backed by documented win rate data. |
| Any trader using Martingale or loss-chasing progressions | Switch to fixed percentage immediately. |
The decision framework is sequential: you earn the right to use more complex approaches by generating the data that shows they are justified. Jumping to variable sizing without that data is adding complexity without adding edge.
8. The Long-Run Compounding Effect of Fixed Percentage Sizing
The compounding advantage of fixed percentage sizing over fixed dollar sizing becomes more pronounced over larger trade samples. At 200 trades, the difference is small. At 1,000 trades, it is meaningful.
Consider a trader with a 58% win rate and 80% payout operating over 1,000 trades:
Fixed dollar ($10 per trade, starting from $1,000):
- Each winning trade adds $8
- Each losing trade subtracts $10
- 580 wins: +$4,640; 420 losses: -$4,200
- Net gain: +$440 (44% total return on initial $1,000)
- The dollar stake never changes regardless of account growth
Fixed percentage (1% per trade, starting from $1,000):
- Each win/loss percentage is calculated on the current balance
- The stake grows as the balance grows, shrinks as it shrinks
- Net gain over 1,000 trades at this edge: approximately $480-$520 (48-52% total return)
- The compounding benefit is modest at 1% stakes but real
The compounding difference is not dramatic at low risk percentages — this is by design. The benefit of fixed percentage is primarily the automatic drawdown protection during losing periods, not explosive compounding during winning ones. The modest compounding benefit is a secondary advantage, not the main argument for using it.
At higher risk percentages (2-3%), the compounding difference between fixed dollar and fixed percentage becomes more significant because the absolute dollar movements are larger, which means the reinvestment on winning trades makes a more meaningful difference. But higher risk percentages also amplify drawdowns, which is why the compounding math should not be used as an argument for raising your risk percentage without a verified edge that justifies it.
9. Documenting Your Sizing System in a Trade Log
Any variable sizing system — whether condition-based or quality-based — requires its own column in your trade log. Without a documented pre-trade record of which tier or condition triggered a different stake, you cannot evaluate whether the variable rule is actually working.
The minimum log columns for variable sizing:
| Column | Example Entry |
|---|---|
| Date/Time | 2026-09-27 10:15 |
| Asset | EUR/USD |
| Setup Tier | Tier 1 (high confluence) |
| Stake | $15 (1.5% at $1,000 balance) |
| Standard Stake | $10 (1% baseline) |
| Direction | Call |
| Expiry | 5 min |
| Result | Win |
| P&L | +$12 |
The "Setup Tier" and "Standard Stake" columns are the additions specific to variable sizing. After 100+ trades, you can group by Setup Tier and calculate win rate per tier independently. If Tier 1 shows a higher win rate than the overall average, the tier definition has predictive value. If it shows the same or lower win rate, the tier is capturing false confidence rather than genuine setup quality.
This data collection step is not optional if you are using variable sizing. Running a variable sizing system without measuring whether the variables actually predict outcomes is trusting a rule you have not tested — which offers no advantage over fixed percentage sizing, and adds the behavioral overhead of making sizing decisions on every trade.
Frequently Asked Questions
Is it better to use a fixed dollar or fixed percentage per trade?
Fixed percentage is better for most traders in most situations. The automatic adjustment to drawdowns and growth makes it more dependable across varying balance levels. Fixed dollar only makes sense if your account balance is so stable that the implied risk percentage does not meaningfully change -- a situation that rarely applies to active traders.
Can I use different position sizes for different asset types?
Yes, and it can make mathematical sense -- but only if your trade log shows a meaningfully different win rate for those assets. If EUR/USD historically shows a 62% win rate for you and GBP/JPY shows 52%, trading EUR/USD at 1.5% and GBP/JPY at 0.75% is evidence-based variable sizing. Making the same adjustment based on gut feel about which asset "feels" more reliable is not.
What is the difference between variable sizing and the Martingale system?
Martingale is a specific type of variable sizing that increases stakes after losses, driven by loss recovery. Its defining property is that stakes grow without bound during losing streaks, eventually exceeding account balance. Variable sizing as described in this article adjusts stakes based on assessed setup quality or market conditions -- not based on recent losses. The critical distinction: variable sizing should never produce a situation where you need a specific trade to win to prevent ruin.
How do I know if my setup quality tiers are actually predictive?
The only method is to record your confidence rating for each trade before entry, then track win rates by tier over 200+ trades. If your "tier 1" setups show a genuinely higher win rate than your "tier 2" setups with statistical significance, the distinction is real. If the win rates are within 2-3 percentage points of each other, the tiers are not reliably predictive and variable sizing will not help.
Should I increase my position size after a big winning day?
Not mechanically. A big winning day tells you that your strategy worked well during that session's conditions -- it does not predict tomorrow's conditions. The appropriate response to a profitable session is to note it in your trade journal and continue at your standard risk percentage the following day. If you have a verified 200+ trade track record showing consistent positive EV, gradual increases in your base risk percentage (from 1% to 1.25%, from 1.25% to 1.5%) are reasonable -- but these should be permanent baseline adjustments, not session-by-session stake inflation based on yesterday's results.