Advanced Risk: Drawdown, Expectancy, Kelly (and Why Not Full Kelly)

Advanced · Advanced Path, lesson 7 of 12 · Trading · 7 min read · Published 19 Oct 2026 · Updated 19 Oct 2026

The math behind long-term risk: how drawdowns compound, how to measure a strategy's expectancy, what the Kelly formula says, and why most traders risk far less than it suggests.

Most risk lessons stop at one trade: where the stop goes and how big the position is. This one zooms out to hundreds of trades. Three ideas decide whether an account survives that long: drawdown, the fall from a peak; expectancy, what an average trade is worth; and the Kelly criterion, a formula that suggests how much of your account to risk per trade.

Kelly is famous, and often misused. It is built to maximize long-run growth under ideal conditions, not to keep you safe, and the gap between those two goals is why most traders who know the formula risk far less than it says. Every figure below is an illustrative example, not a target.

Open a trade on Traders Portal

Drawdown: why losses are harder to undo

A drawdown is the fall in your account value from its highest point, in USDT or as a percentage. Because each gain and loss works on a changed balance, the gain needed to recover is always bigger than the loss: gain needed = drawdown ÷ (1 − drawdown).

Losing streaks are where drawdowns come from, and they are more common than intuition suggests. With a 45% win rate, the chance that any given run of five trades is all losses is 0.55 multiplied by itself five times, about 5%. Over a few hundred trades, streaks of seven or eight losses are entirely normal.

On Traders Portal, the Portfolio page's Overview shows your maximum drawdown over a chosen period, and in Account → Alerts you can add an Equity drawdown alert that fires when equity falls a set percentage below its high for the day (UTC).

Expectancy: what an average trade is worth

Expectancy is the average result per trade over many trades:

Expectancy = (win rate × average win) − (loss rate × average loss)

It is often expressed in R, where 1R is the amount you risk on a trade. Measuring in R lets you compare results across different position sizes.

Use net figures. The Net PnL column in Closed Positions already includes fees and swap, and the Portfolio Overview lists your win ratio, average profit and average loss for a period. Fifty trades is a small sample, though: a couple of lucky winners can make a weak strategy look strong. Risk/reward and win rate explains how the two numbers trade off.

The Kelly criterion

The Kelly formula takes your win rate (W) and your payoff ratio (R, the average win divided by the average loss) and returns the share of equity to risk per trade:

Kelly % = W − (1 − W) ÷ R

With the numbers above, W = 0.40 and R = 2.25, so Kelly = 0.40 − 0.60 ÷ 2.25 = 0.40 − 0.267 = 13.3%. On a 5,000 USDT account, that means risking about 667 USDT per trade: the amount lost if the stop is hit, not the position value and not the margin.

Under idealized conditions (fixed, known odds and independent trades), risking exactly this share maximizes the long-run growth rate of the account. That is all it promises. It says nothing about how bumpy the ride is, and it assumes your inputs are exactly right.

Why traders do not use full Kelly

• Your inputs are estimates. Suppose your true win rate is 35%, not 40%. True Kelly is then 0.35 − 0.65 ÷ 2.25 = 6.1%. Risking 13.3% is more than twice that, and above twice Kelly the account tends to shrink over time even though each trade still has positive expectancy (0.35 × 2.25 − 0.65 = +0.14R).
• The drawdowns are severe. Eight straight losses at 13.3% each leave 0.867 to the power of 8 = 0.318 of the account: a 68% drawdown. In the idealized math, a full-Kelly bettor has roughly a 50% chance of seeing the account halve at some point.
• Trading is not a coin toss. Stops can slip in gaps, related trades lose together, edges fade as markets change, and leveraged positions are closed automatically at the stop-out level set for your account.
• People are not machines. Few traders keep following their rules through a 50% drawdown.

Fractional Kelly risks a set share of the Kelly figure. In this example, half Kelly keeps about three-quarters of the theoretical growth rate while cutting the swings sharply, and in the same idealized math its chance of ever halving the account falls to about one in eight. Many traders go further and use fixed-fraction risk, such as 0.5% to 2% of equity per trade, and use Kelly only as a ceiling and a check: if your honest numbers give a Kelly figure at or below zero, there is no edge to size, and no position-sizing rule will fix that.

Putting it to work on Traders Portal

1. Collect honest numbers

   Note at least 50 to 100 trades from Closed Positions, using Net PnL. Demo results are a start; live results are better.

2. Work out expectancy and Kelly

   Calculate win rate, average win, average loss, expectancy and the Kelly figure. Use cautious numbers: round the win rate down and the average loss up.

3. Pick a fixed fraction well below Kelly

   Choose a risk per trade, for example 1% of equity, and turn it into USDT. In the order panel, switch on TP / SL and check that the Loss (USDT) figure at your stop matches it.

4. Write drawdown rules

   Decide in advance what happens at, say, a 10% drawdown: halve your risk, pause, or review. Back it up with an Equity drawdown alert.

Conclusion

Drawdowns compound, so the size of each loss matters more than it seems. Expectancy tells you whether there is an edge worth sizing at all. Kelly shows the theoretical maximum-growth bet, and its sensitivity to bad estimates and its deep drawdowns are exactly why practical traders use a fraction of it or a small fixed percentage instead. Measure honestly, risk a small share per trade, and connect your sizing to the position-sizing habits you already use.

Topics: Risk Management