CryptoScoopDaily
Home

Tools

Trading calculators Position size calculatorLiquidation price calculatorFutures calculator Profit & DCA Crypto profit calculatorDCA calculatorBitcoin DCA calculator Send crypto Send crypto checkCheapest way to sendWrong network checker Research tables RWA trackerStablecoin dashboardAirdrop checker All calculators

Guides

How to send crypto safelySent to the wrong networkDo you pay tax on crypto?How to file crypto taxesHow much to risk per tradeWhat is DePIN?All guides About us Contact us

Kelly Criterion Calculator for Crypto Position Size

If you know roughly how often a setup wins, and how big wins are versus losses, what fraction of the account is the ceiling — not the order?

How often a trade like this wins. 55 means a bit more than half.
What you usually make on a winning trade.
What you usually lose on a losing trade.

Guide

How to use the Kelly criterion for crypto position size

It is a formula for the fraction of a bankroll to bet when you estimate how often you win and how big wins are versus losses. Used raw, it maximizes long-run growth and also maximizes how hard a bad streak can hit you. Most people use half or quarter Kelly.

Kelly's fraction is win rate minus the chance of losing, divided by the payoff ratio. If that number is negative, Kelly says bet nothing. A 55% win rate with a 2-to-1 payoff (average win $200, average loss $100) gives full Kelly of 32.5% of the account per independent bet — a number almost nobody should trade. Half Kelly is 16.3%. Quarter is 8.1%. Crypto regimes change, so last month's 58% win rate is not a known fact. Treat Kelly as a ceiling on the position-size calculator, not as an order ticket. If the edge is not real, the formula will still output a fraction, and that is how people blow up. This page does not pull your journal. You type the numbers.

Example

Kelly criterion example: 55% win rate, 2-to-1 payoff

Match this on the sliders: 55% win rate, $200 average win, $100 average loss.

Win rate
55%
Payoff
2.0
Full Kelly
32.5%
Half Kelly
16.3%
Quarter
8.1%
Use as
A ceiling

32.5% of the account on one idea is a career-risking bet for most people. Take quarter Kelly to the position-size calculator and turn it into a stop, not into a market order.

You can do these on a napkin.
Step In words Example
1. Payoff Average win ÷ average loss $200 ÷ $100 = 2
2. Full Kelly Win% − (lose% ÷ payoff) 0.55 − (0.45 ÷ 2) = 32.5%
3. Quarter Kelly Full Kelly ÷ 4 8.1% of the account

Limits

What this Kelly calculator cannot do

  • It will not prove you have an edge.
  • It will not place a trade or a stop.
  • It will not pull a live win rate from an exchange.
  • It will not save you if last month's 58% was luck.

If Kelly printed a fraction you can live with, open the position-size calculator and turn that fraction into coins and a stop.

Keep going

Same toolkit, a different question — pick the next calculator.

FAQ

Questions people ask

A formula for the fraction of a bankroll to bet when you estimate how often you win and how big wins are versus losses. Used raw, it grows the pile fast and also makes a bad streak brutal. Most people use half or quarter Kelly.

Take your win rate, subtract the chance of losing divided by the payoff ratio. If that number is negative, Kelly says bet nothing. A 55% win rate with wins twice as large as losses gives full Kelly of 32.5% — a number almost nobody should trade.

Because win rate and payoff are guesses, not known constants. Full Kelly is extremely jumpy. Half or quarter Kelly is the usual practical cap for people picking their own trades.

Only as a ceiling, and only if the win rate comes from a large, boring sample. Crypto regimes change. Treating last month's 58% as a known fact is how people blow up.

It is a second opinion on how large a fraction to risk. Turn that fraction into coins and a stop on the position-size calculator. Do not send Kelly's percent straight to the order ticket.

Then the numbers you typed do not have an edge. Either the win rate is too low, the payoff is too thin, or both. The formula will still output a fraction if you force bad inputs — that is on you.

No. Last month is one regime. If the sample is small or the market changed, Kelly will still print a confident number. That number is not a promise.

No. You type a win rate and average win and loss from your own journal. A widget cannot know your edge.