Max drawdown: how to read it and why the exchange number is too kind

Drawdown is the only metric on this site that predicts whether money survives. It is also the one most often quoted wrong. This guide explains what it is, why the exchange's number is systematically softer than the truth, and where the line sits.

What drawdown is

Drawdown is how far the account fell from its own peak:

drawdown(t) = 1 − capital(t) / peak(t)

• −20% means: the account was at its high, then fell 20% below that high.
• It says nothing about when it happened. A 20% dip that recovered in two days and one that lasted four months are the same number on a chart, and very different experiences for a follower.

Maximum drawdown is the worst such value in the window. Our leaderboard computes it over the compounded equity curve of the last 180 days.

Why the exchange figure is kinder than the truth

Bybit's drawdown field for longer periods can exceed 100% and is an additive percentage sum, not a real depth of loss. Two effects make it systematically smaller than what a follower experiences:

1. Deposits and withdrawals. If a trader withdraws profit mid-window, the visible fall is diluted by the capital that left the account.
2. The window start. A 180-day drawdown measured from a lucky local peak can look better than the drawdown a follower entering at the top actually saw.

We therefore rebuild the curve ourselves from the daily series (yield-trend, 180 daily points), compound it, and measure the fall from the running peak. Where our series and the exchange disagree wildly (our curve shows −99%, the exchange reports −20% for the same 90 days), we trust the exchange - that pattern is a data artefact, not a wipeout.

The thresholds that matter for a follower

Max drawdown	What it usually means
0–10%	Very rare. Usually a small account or a period without real risk.
10–20%	Normal for a careful swing trader. Copying is usually survivable.
20–40%	Typical for a high-return account. Half of it would be painful.
40–60%	You need the trader's recovery to be right. Most followers quit here.
60–90%	One bad month from a wipeout.
90%+	A liquidation already happened, or is one volatility spike away.

Our low-drawdown screen keeps accounts under 25% on our own measure, with no liquidation days at all - see /low-drawdown/.

Why "50% drawdown and back" is a red flag

A trader who went through −50% and recovered did it with a capital base that halved. If you copied at the top, your own account took the same path - and your position size limits, your stop-loss settings and your willingness to sit through it are not the trader's. Recovery stories are also the most common shape of a martingale grid: many small wins, then one loss large enough to erase them.

The rating penalises drawdown with a linear scale: 15% costs nothing, 30% costs 6 points, 50% costs 11, 100% costs 20. Volatility and worst-day penalties are on top of that, because a deep drawdown that arrives gradually and one that arrives in a single day are not the same risk.

How to use this number before you copy

1. Open the trader's page and look at the monthly returns strip. A deep drawdown spread over months is a different animal from a single -40% month.
2. Compare our drawdown with the exchange figure. A large gap means deposits and withdrawals, and the exchange number describes a person you cannot copy.
3. Check the liquidation days column. Zero is the single cheapest filter available.
4. If you copy anyway, size the position so that the same percentage drop does not threaten your account, and set a copy stop-loss (PCSL) so you are not present for the whole recovery.

Past performance does not predict the future. Drawdown is how you compare accounts on the same basis - not a promise that your money survives.