Why does BlofinX use quarter Kelly rather than half or full?

Half and full Kelly both assume the win probability is known. BlofinX estimates it from its own closed trades, so the input carries real error. A quarter retains roughly 44% of full Kelly's theoretical growth rate at a quarter of its volatility, which is the right trade when the edge estimate itself is uncertain.

What is the maximum account risk per trade?

2% of capital, enforced as a hard cap in the sizing engine. If the Kelly calculation returns a larger number — and on a high-conviction setup it easily can — the cap overrides it. Institutional risk guidelines generally suggest 1% to 2% of portfolio equity regardless of model conviction.

What happens before the engine has enough trade history?

It does not size from Kelly. Below 30 closed outcomes the engine holds a fixed conservative size, because a win rate measured off a handful of trades cannot be distinguished from luck, and Kelly is extremely sensitive to that input.

Fractional Kelly Position Sizing for Crypto

Even with a high-accuracy signal generator, trader accounts can suffer catastrophic drawdowns if position sizing is arbitrary. The Kelly Criterion, developed by researcher John Kelly at Bell Labs, provides the exact mathematical fraction of bankroll to risk on a trade — provided you know your edge. That proviso is where most implementations quietly fail.

Position size calculatorPut your own win rate, payoff and stop distance in and see the quarter-Kelly size the engine would take — including when it refuses.

The Kelly Formula Breakdown

The standard Kelly formula is expressed as K% = W - [(1 - W) / R], where W is the win probability and R is the win/loss payoff ratio. If a trader has a 60% win rate and a 2:1 reward-to-risk ratio, Full Kelly recommends risking 40% of capital per trade. Almost nobody should act on that number.

Why Full Kelly is Dangerous in Volatile Crypto Markets

Full Kelly assumes precise knowledge of win probabilities. In real-world crypto trading, where return distributions have heavy tails, W is estimated from a noisy sample — and Kelly's penalty for overstating an edge is asymmetric. Stake double the true optimum and you do not get double the growth; you get ruin. Fractional Kelly scales the whole bet down to buy insurance against your own estimate being wrong. Growth falls off gently, volatility falls off linearly, and survival improves sharply.

Quarter-Kelly Allocation = 0.25 * (Win_Rate - (1 - Win_Rate) / Reward_Risk_Ratio)
Risk per trade = min(Quarter_Kelly_Pct, 2.0%)   // hard cap, whatever Kelly says
If closed_trades < 30: hold a fixed conservative size instead

How BlofinX Actually Sizes: A Quarter, Capped at 2%

Every position the engine opens is sized from its own realized results, not from a claimed win rate. It takes a quarter of full Kelly, then clamps the result to a hard 2% of capital at risk per trade — the cap wins whatever the formula returns. Two further rules matter more than the arithmetic: below 30 closed outcomes it refuses to estimate an edge at all and holds a fixed conservative size, and a non-positive edge sizes to zero rather than to a small position. Clamping a negative Kelly up to something positive is the classic bug that turns 'do not trade this' into 'trade this small'.

Summary

Quarter-Kelly sizing behind a hard 2% cap gives up some theoretical growth to buy a large reduction in drawdown — and, more importantly, it stays correct when the win rate it was fed turns out to be optimistic. That is the trade every surviving book makes.

Fractional Kelly Position Sizing: The Math BlofinX Sizes Trades With