Why the intuition fails
Averaging down feels safe: add to a long as price falls, the average entry drops, and liquidation moves further away.
That intuition holds under a single maintenance rate. It fails because the rate is not single — it steps up with position size. Adding does two opposing things at once:
Which one wins depends on how close you already were to a tier boundary.
- The average entry improves, which pushes liquidation away
- The position's notional grows and can cross into a higher maintenance tier, which pulls liquidation closer
When it gets worse
The trap is sprung when the position already sits near the top of its tier.
On Binance BTCUSDT, the second tier caps at 600,000 USDT of notional at a 0.5% maintenance rate. Cross it and the third tier applies 1% — **the rate doubles in one step**.
A 550,000 position that adds 100,000 lands in the third tier. The gain from a better average entry very plausibly does not cover a maintenance rate going from 0.5% to 1%.
Every addition near a boundary is worth computing first.
How to check before adding
Compute the post-addition position and compare it to where you are now. Three numbers:
The calculator applies each venue's published ladder, so entering the post-addition parameters shows which way the liquidation price moves:
The per-tier tables also show how much room remains before the next boundary.
- Total size after adding
- Average entry after adding
- Total margin after adding, including whatever you post
- Run it before you add
Venues place the boundaries differently
The same addition can produce different outcomes at different venues, because both the boundaries and the unit they are measured in differ.
Binance and Bybit band by notional value; OKX bands by contract count. On OKX a rising price does not push you into a higher tier. On the other two it does.
So "adding was fine on one venue and made things worse on another" is entirely possible, and nobody miscalculated.
It can happen without adding at all
On a notional-banded venue, **a one-directional price rise can push notional into a higher tier by itself**.
A long that you never touched can face a stricter maintenance requirement after a large move up. That makes tier position something to re-check while a position is open, not only when adding to it.
What to do about it
There is no general answer, but there are facts worth establishing first.
The first three are calculable. The fourth is a judgement.
- How much room sits between your current notional and the next boundary
- Whether adding crosses it, and what the maintenance rate becomes if it does
- If it crosses, how much margin restores at least the liquidation distance you had
- Whether the alternative is better: reducing instead of adding, or opening separately where the ladder is more permissive
A note on what this is
This explains a mechanism; it is not a recommendation to average down. Averaging down enlarges a losing position, and this page only explains why it sometimes fails to improve even the margin safety it appears to buy.
Leveraged trading can lose your entire deposit. The liquidation that happens is the one the exchange computes.