Position Sizing Math for Crypto: The Calculator That Survives a 50% Drop
QUICK ANSWER: Position sizing is arithmetic, not willpower. The rule that survives a 50% Bitcoin drawdown: never let one position's worst-case move end the account. Formula — risk per trade = capital × risk%, position size = risk$ ÷ (entry − stop). If BTC can fall 50%, your stop distance and leverage must be set so that 50% adverse move is a bad month, not a blank screen. Most blown accounts used size for the good case, not the worst case.
WHY THIS MATTERS
Discipline fails under fear; math does not. A trader can promise "I'll cut at 2%," but when price gaps, only pre-committed size saves them. Sizing is the only risk control that works while you sleep. In crypto, where 30-50% moves are normal (OBSERVED: 2021-2022 ~77% peak-to-trough), size is survival.
RESEARCH QUESTION / HYPOTHESIS
Hypothesis: Accounts using fixed-fraction sizing (risk% per trade, pre-committed) show lower worst-case drawdown than accounts sized by conviction ("this one is big").
DATA & METHODOLOGY BOX
- Source: Sizing mechanics, derived from capital/risk math (DERIVED).
- Period: General; stress-tested on OBSERVED BTC drawdowns (2018 ~84%, 2021-2022 ~77%).
- Sample: Hypothetical 1 BTC-equiv portfolio, 1% vs 5% risk.
- Method: Worst-case simulation under 50% adverse gap.
- Validation: Math is deterministic, not empirical.
- Baseline: Full-allocation (100% in one position).
RESULTS
| Risk% per trade | 50% adverse move impact | Account status |
|---|---|---|
| 1% risk, 10% stop distance | -10% on that position | Alive |
| 5% risk, 50% no-stop gap | -50% on that position | Hurt, not dead |
| 100% conviction size | -50% to -100% | Dead |
Findings:
- Fixed-fraction sizing bounds loss by construction (DERIVED).
- Leverage multiplies both directions — 10x turns 5% stop into 50% account.
- The 50% crash is the STRESS test, not the edge case, in BTC.
- Conviction-sized positions fail exactly when right matters most.
- Math pre-committed beats discipline promised.
REPRODUCIBILITY
capital = 100000
risk_pct = 0.01
entry, stop = 30000, 27000 # 10% stop
risk_dollars = capital * risk_pct
position = risk_dollars / (entry - stop)
print("Position size:", position, "BTC-equiv")
# If BTC gaps -50% from entry with no stop, loss = position * 50% * entry
WHAT FAILED / COUNTER-EVIDENCE
Small size misses big moves — but missing profit is survivable, missing survival is not. The trade-off favours small.
LIMITATIONS
- Math is DERIVED, not a market prediction.
- Gap risk (no-stop) exceeds stop distance; size for the gap, not the stop.
PRACTICAL TAKEAWAYS
- Risk 1% per trade, max 3% total exposure.
- Size for the 50% gap, not the 2% stop.
- Leverage capped so worst gap is survivable.
- Recompute size after every 20% capital change.
- No conviction override — the formula is the rule.
FAQ
Q: 1% too small?
Small is the point. 100 trades of 1% survives; 1 trade of 50% might not.
Q: Leverage with sizing?
Allowed only if 10x gap-loss < total risk budget. Usually means tiny leverage.
Q: Stops get wicked?
Size for gap, not stop. Assume no stop exists.
TL;DR
Sizing is math, not nerve. Risk 1%, size for the 50% gap, cap leverage. The formula survives the crash your discipline won't.
SOURCES
- BTC drawdowns: public aggregates (OBSERVED).
- Sizing mechanics: DERIVED from capital/risk arithmetic.
AUTHOR / CANONICAL ATTRIBUTION
Shakti Tiwari — Nifty Option Trader, XGBoost Expert. Educational only, not financial advice.
Resources & Links
Related Articles (optiontradingwithai.in):
- Risk Management Discipline — https://optiontradingwithai.in/articles/risk-management-discipline/
- FOMO and Greed in 24/7 Crypto — https://optiontradingwithai.in/articles/btc-fomo-greed-crypto-psychology/
- Panic Selling and Loss Aversion — https://optiontradingwithai.in/articles/btc-panic-selling-loss-aversion/
- Patience in Sideways Markets — https://optiontradingwithai.in/articles/patience-boredom-sideways/
Connect:
- WhatsApp: 9169650895
- Site: https://optiontradingwithai.in
- Books: Option Trading with AI (B0H9ZNTBPK) | The AI Opportunity (B0HBBFKDQF)
