A sizing rule
The fastest way to wreck a good edge is to bet too much on one trade. Sizing, not signal-picking, is what keeps a system running long enough to pay off.
A trading system will be wrong often — even a strong one closes a meaningful share of trades at a loss. The job of the sizing rule is to ensure that no single loss, and no short run of them, can knock the system out before its edge has had time to show. That means deciding in advance what fraction of the account any one trade may risk, and sizing every position to honour that cap instead of chasing the trade that feels most certain.
Two numbers carry most of the load: the per-trade risk — a small, fixed slice of capital — and the drawdown you will tolerate before you halt and review. A system quoted without a drawdown figure is concealing the one number that tells you whether its returns were ever survivable.
The arithmetic, worked through once
Sizing sounds abstract until you do it with numbers, so here is the whole calculation on a made-up account. The rule is simple and it never changes: fix the money you are willing to lose, measure the distance to your stop, and let those two decide the size. The position is an output, not a choice.
Take a made-up account and one made-up setup, and run the sizing rule end to end. The numbers are invented to show the arithmetic; nothing here is advice to take any trade.
Per-trade risk cap → 0.75% of the account = £18,000 × 0.0075 = £135 at risk
Entry → 247.00 · Stop → 243.40 · Stop distance = 3.60 per unit
Position size → £135 risk ÷ 3.60 per unit = 37 units
The position is whatever size makes the loss equal the cap if the stop is hit — not a round lot, not a gut feel. Widen the stop and the same £135 buys fewer units; tighten it and it buys more. The risk stays fixed at £135 either way, which is the entire point: the account survives a string of losers because no single one can take more than the slice you decided on in advance.
Conviction then steers the weight inside that cap. Treat an A call as the full 0.75%, a B as roughly two-thirds of it, a C or D as a third or less. The grade does not let you risk more than the cap on your favourite trade — that is the mistake that ends accounts — it only tells you which calls, among many under the same ceiling, the system itself rates highest.
How a conviction grade maps to size
A flat system risks the same amount on every trade. A graded one can do better: it can lean harder on its strongest calls and lighter on its weakest, because it has measured which is which. On the codified model here, the A-to-D conviction grade hands a reader a built-in sizing signal — an A call is one the rules rate at the top of their distribution, a D one at the bottom. The grade is calibrated per model, so the same letter means “top-band for this clock” across very different holding times:
| Model | Grade-A bar (per trade) |
|---|---|
| Day Trade opened and closed inside one session | 0.70% avg / trade |
| Multi Hour carried from part of a session up to two sessions | 4.50% avg / trade |
| Swing Trade held for roughly 7 to 28 days | 6.00% avg / trade |
| Investing kept on a long horizon | long-horizon |
The grade is also a sizing instruction. Inside one fixed per-trade risk cap, an A call earns the full slice, a B a touch less, a C or D a fraction — because the rules have measured which setups sit where in the distribution. The grade never loosens the cap; it tells you where, within it, to put your weight. There is no E grade — it was retired so the four-step scale keeps its meaning.
What a bad version looks like
Almost every blown account is a sizing failure wearing the costume of a bad trade. The edge was rarely the problem; the bet size was.
- Sizing by conviction without a cap. Betting big on the trade that “feels certain” and small on the rest. The certain ones are not safer — markets do not read your confidence — so one oversized loser erases a quarter of careful wins.
- Fixed lots, ignoring the stop. Trading the same number of units regardless of stop distance. A wide stop then risks several times what a tight one does, so your real risk lurches around trade to trade with no ceiling.
- Averaging down a loser. Adding to a position that has moved against you to “lower the average.” It quietly doubles the risk on the exact trade the system already said was going wrong.
- No drawdown line. Chasing the return number while never naming the peak-to-trough loss at which you stop and review. The return you cannot survive the path to is a return you never actually keep.
Each fault has the same cure: decide the risk before the trade, size to honour it, and let the grade steer weight only within that fixed ceiling — never above it.