Athenum Research · Version 1 · 25 September 2026
How often does 10x, 20x or 100x leverage get liquidated?
A leverage calculator shows where your liquidation price sits. It does not show how often the market actually got there. We measured it: at every hourly close for twelve months, we opened a hypothetical long and short on Binance's BTC and ETH perpetuals and checked whether the mark price reached the liquidation price within the next 4 hours, 24 hours, 7 days or 30 days.
38.8%
of 20x BTC longs reached their liquidation price within 7 days
3,399 of 8,760 hourly entries, 25 August 2025 to 25 August 2026 (UTC). Over the same 7 days: 12.8% at 10x and 73.7% at 50x.
Binance USD-M BTCUSDT and ETHUSDT perpetuals · entries at the 8,760 hourly closes from 25 August 2025, 01:00 UTC to 25 August 2026, 00:00 UTC · mark-price candles through 23 September 2026, so every 30-day window is complete · isolated margin, 0.4% maintenance margin, no fees or funding.
Bitcoin (BTCUSDT): share of entries that reached liquidation
| Leverage | Distance to liquidation | Long, 24 h | Short, 24 h | Long, 7 days | Short, 7 days | Long, 30 days | Short, 30 days |
|---|---|---|---|---|---|---|---|
| 2x | 49.80% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| 3x | 33.07% | 0.0% | 0.0% | 0.0% | 0.0% | 4.3% | 0.0% |
| 5x | 19.68% | 0.0% | 0.0% | 2.2% | 1.5% | 22.9% | 9.2% |
| 10x | 9.64% | 0.8% | 0.3% | 12.8% | 7.7% | 40.5% | 37.5% |
| 20x | 4.62% | 6.2% | 5.2% | 38.8% | 36.1% | 63.3% | 67.7% |
| 25x | 3.61% | 11.2% | 9.3% | 49.7% | 46.1% | 70.9% | 73.4% |
| 50x | 1.61% | 39.4% | 37.7% | 73.7% | 74.0% | 88.6% | 85.5% |
| 75x | 0.94% | 58.0% | 58.8% | 83.6% | 83.4% | 93.9% | 90.7% |
| 100x | 0.60% | 71.3% | 72.4% | 89.4% | 88.9% | 96.2% | 93.6% |
| 125x | 0.40% | 80.7% | 81.6% | 93.2% | 92.4% | 97.8% | 95.7% |
Ethereum (ETHUSDT): share of entries that reached liquidation
| Leverage | Distance to liquidation | Long, 24 h | Short, 24 h | Long, 7 days | Short, 7 days | Long, 30 days | Short, 30 days |
|---|---|---|---|---|---|---|---|
| 2x | 49.80% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| 3x | 33.07% | 0.0% | 0.0% | 0.5% | 1.2% | 11.5% | 7.2% |
| 5x | 19.68% | 0.2% | 0.2% | 7.4% | 1.9% | 32.2% | 19.0% |
| 10x | 9.64% | 1.9% | 1.7% | 24.7% | 23.4% | 58.2% | 54.1% |
| 20x | 4.62% | 15.4% | 12.0% | 50.6% | 49.2% | 74.1% | 74.5% |
| 25x | 3.61% | 23.6% | 19.1% | 59.1% | 58.3% | 78.1% | 79.4% |
| 50x | 1.61% | 51.4% | 50.6% | 79.4% | 79.3% | 90.8% | 88.7% |
| 75x | 0.94% | 68.9% | 68.9% | 88.0% | 87.7% | 95.2% | 93.4% |
| 100x | 0.60% | 78.9% | 79.4% | 92.9% | 92.2% | 97.4% | 95.6% |
| 125x | 0.40% | 85.7% | 86.0% | 95.2% | 94.6% | 98.3% | 96.9% |
Each cell divides the entries that reached liquidation by all 8,760 entries. Distance is the adverse price move from entry to the liquidation price. The CSV adds 4-hour windows and exact counts.
How far price moved against a fresh position
The same data, read the other way: within 24 hours of entry, the median worst move against a long was 1.21% for BTC and 1.67% for ETH. In one of ten entries it exceeded 3.80% (BTC) and 5.56% (ETH). A 20x position sits 4.62% from its liquidation price.
What changes the odds
Two things dominate: the distance to the liquidation price, set by leverage and maintenance margin, and the time a position stays open. At 10x, 0.8% of BTC long entries reached liquidation within 24 hours, but 12.8% within 7 days. ETH moved more than BTC in this period, so the same leverage was liquidated more often. Fees, funding and larger positions in higher maintenance tiers move the liquidation price closer, so real outcomes can be worse than this table.
Download, verify and cite
- CSV: touch counts by asset, side, leverage and window
- CSV: adverse-move quantiles
- JSON: results, source requests and SHA-256 hashes
- JavaScript: reproduction script
- JavaScript: pure calculation module
- Python: independent recalculation
Suggested citation: Athenum Analytics (2026), How often does 10x, 20x or 100x leverage get liquidated?, version 1, 25 September 2026. Source candles: Binance. https://www.athenum.xyz/research/leverage-liquidation-odds/
How to reproduce the result
Put reproduce.mjs and analysis.mjs in one folder and run the command below with Node 22 and internet access. The script downloads Binance last-price and mark-price hourly candles, checks that every hour is present exactly once and stops on missing, duplicate or invalid candles. No values are filled in. It writes both CSV files and summary.json.
node reproduce.mjs ./resultsE is the last-price close of an hourly candle (Binance labels candles by their open time; the close is one hour later). Long liquidation price = E × (1 − 1/L) / (1 − MMR); short liquidation price = E × (1 + 1/L) / (1 + MMR), with leverage L and maintenance margin rate MMR = 0.4%. A long counts as liquidated within h hours if one of the next h hourly mark-price lows is at or below its liquidation price; a short if one of the next h mark-price highs is at or above it.
A separate Python implementation using only the standard library downloaded the same candles and recalculated every count and quantile. All 256 values matched.
python3 verify.py ./results/summary.jsonLimits
This describes one venue and twelve months; it is not a forecast, a trading signal or advice. Trading fees, funding payments, higher maintenance tiers, added margin, partial liquidation and liquidation fees are excluded. Binance liquidates on the mark price; we use its published hourly highs and lows. Entries overlap in time, so the 8,760 observations per asset are not independent experiments. Other periods and other exchanges can produce different frequencies.
The download contains our calculated statistics and code, not the underlying price feed. You may cite, share and reproduce these Athenum materials with attribution to Athenum Analytics and Binance as the data source. Obtain source prices from Binance under its terms; no licence to Binance raw data is granted here.