Why our fee rises with leverage — and why we show you the curve

What was wrong with a flat fee

The desk originally took a flat 5% of net profit whatever boost you used. It was simple, it was disclosed, and it was wrong in both directions at once.

Consider what the house is actually selling. At ×2, a player is making a slow directional call and mostly wins or loses on their own reading; the house carries very little variance. At ×1000, the house is carrying a position that can swing a hundred times the stake in a minute, and it is doing so on every open ticket simultaneously. Those two products cost wildly different amounts to provide. Charging them the same rate means the ×2 player is paying for the ×1000 player’s risk.

That is not a fairness argument dressed up as economics — it is the same logic that makes an insurer price a scooter differently from a superbike. The premium should track the exposure.

The curve, and why it is a curve

The obvious fix is tiers: 3% up to ×10, 5% up to ×100, and so on. We did not do that, for one specific reason — a tier boundary is a thing to game. If the rate jumps at ×100, then ×99 and ×100 are meaningfully different products, and a player learns to sit just under the line for reasons that have nothing to do with their view of the market.

So the rate is interpolated continuously, on a log scale between a floor at ×1 and a ceiling at ×1000. Every rung costs slightly more than the one below it and there is no step to hide behind.

A log scale rather than a straight line because leverage itself is felt logarithmically: the step from ×2 to ×5 changes the character of a position far more than the step from ×500 to ×503. Pricing on the same scale keeps the fee proportionate to the thing it is pricing.

The share comes out of a win. The entry fee comes out of everything

There are two charges and they behave differently. The entry fee is fixed when you open and taken from your result at close, whether that result is a win or a loss — it is the price of the position, not a cut of its outcome. It rises with the boost for the same reason the share does, and it is the larger of the two at the top of the curve.

One thing it does not do is push you past your stake. A liquidation costs exactly what you put in and nothing more; the entry fee is absorbed by that floor rather than charged on top of it.

The share is the part that only ever comes out of a win. Order matters: funding and the entry fee are subtracted first, and the share is taken from whatever is left. If they exceed your price gain there is no positive net profit and therefore no share, even though the price went your way.

1. gross = price P&L − funding − entry fee
2. share = gross × your rate, but only if gross is positive
3. net = gross − share, floored at −stake

The part we think matters most

A variable fee is only defensible if it is visible. The moment a rate depends on something, an undisclosed rate becomes a place to hide — and the reference this model borrows from keeps its exact cut hidden, scaling it by leverage, size and hold time behind the scenes.

So we scale by leverage alone, which is the one input a player chooses deliberately and can see on screen, and we print the resulting percentage in the ticket for the rung currently selected. Move the boost and the number moves with it, before you have committed anything.

Keeping the quoted number honest is an engineering problem as much as an editorial one. If the ticket computed its own estimate and the server computed the charge, the two would eventually disagree — different rounding, different constant, a config change that reached one side first. So the server sends the floor and ceiling, and both sides resolve the rate through the identical function. What you are shown is what you are charged, by construction rather than by promise.