Liquidation days: what they mean for a follower

A liquidation day is a day when the account lost 90% or more. It is not a bad day. It is a different kind of event: the account effectively stopped existing and was either restarted with a new deposit or never recovered. On this site it is the most heavily penalised component of the rating.

How we count them

We take the public daily series Bybit exposes (yield-trend, 180 daily points) and count the days where the daily ROI is −90% or worse. No guessing, no interpolation - either the series shows it or it does not.

One caveat: where our daily series shows −99% while the exchange reports −20% for the same 90 days, we trust the exchange and do not count the day. That pattern is a data artefact (a deposit or reset inside the window), not a wipeout, and counting it would punish an account unfairly.

Why it matters more than return

Two accounts can both show +150% ROI over 180 days:

• Account A: a steady grind with a 25% drawdown.
• Account B: +40%, then one −93% day (deposit), then +200%.

The rating gives A about 25 points more. That gap is not a modelling artefact - it reflects what happens to the person copying.

What it does to a follower

Copy trading opens your positions proportionally, but it does not copy:

• your deposit (so a wipeout wipes your money, not a portion of theirs),
• your leverage limits or per-symbol margin caps,
• your stop-loss settings,
• your willingness to hold through the recovery.

There is one more mechanical difference that matters: Bybit's auto-unsubscribe rules. Twenty failed copies in a row, or 100 within 24 hours, and the platform unsubscribes you - usually right in the middle of the recovery the trader needed. So the follower tends to keep the loss and miss the return.

The thresholds

Liquidation days in 180 days	Penalty	What it usually is
0	0	Normal for a careful account.
1	−18	A single catastrophic event: usually a grid blowing up once.
2	−25	A system that keeps failing, not a one-off.
3+	−30	The rating exists mainly to warn you here.

Compare that with the return bonus: at 400% ROI you can earn at most +20, and at most +10 if the exchange win rate over 180 days is ≥ 90% (a grid). A liquidation day costs more than the best possible return bonus. That is deliberate: we do not want a wipeout to be buyable.

Before you copy anyone with a liquidation day in history

1. When was it? A wipeout 8 months ago, followed by a steady recovery, is a different object from one three weeks ago.
2. What happened to the deposit? If the account got a fresh deposit and the equity curve restarted, the old history is not comparable. Our curve is built from the daily series, so a visible discontinuity is worth noticing.
3. What is the account doing now? Check the current drawdown column: if the trader is again 30%+ below the peak, you are joining at the wrong point of a cycle that has already failed once.
4. What do followers' PnL say? A trader with a long recovery and negative 90-day follower PnL has already demonstrated what following them feels like.
5. Size accordingly, or don't. Our low-drawdown screen keeps accounts with zero liquidation days - see /low-drawdown/.

How liquidation days interact with other components

A wipeout is almost never an isolated fact. It drags three other metrics with it:

• Max drawdown goes to 90%+ and costs up to 20 points.
• Daily volatility spikes, costing up to 15 points.
• Worst day hits the −90% node and costs up to 12 points.

The worst single account in the database can lose 55+ points across those four lines. This is intentional: the formula is built so that surviving is worth more than earning.

The honest caveat

A liquidation day is a historical fact about someone else's account. It is a strong prior for risk, not a prediction about your account. Copying a trader who wiped out once, at small size, with your own stop-loss, may be a rational trade. Copying at full size, with no limits, is not. Use the number to size the position

• not to make the decision for you.