Why Inconsistent Bet Sizing Quietly Kills Good Systems

Most traders treat position sizing as an afterthought to entries and exits. They obsess over the signal and improvise the quantity. But expectancy is a per-trade average, and an average only compounds cleanly when each trade is weighted roughly the same. The moment your sizes scatter — big here, tiny there, huge on the one you "felt good about" — your account stops behaving like your system.

Here is the uncomfortable truth: a positive edge is a statement about the average trade. If your sizing is inconsistent, you are no longer trading the average. You are trading a handful of oversized bets, and the rest is noise. When those oversized bets happen to coincide with losses — and they usually do, for reasons we'll get to — even a strong system can post a losing year.

A profitable system needs a repeatable bet size

Think of expectancy as the house edge in a casino. The edge is real, but the house only collects it by taking thousands of bets of similar size. If a casino let one gambler wager the equivalent of a thousand normal hands on a single spin, a positive-edge game could still bankrupt the table on a bad night. Variance dominates when size is uneven.

Your account is the table. A repeatable bet size is what lets your edge actually express itself over a sample. Without it, a good month and a blow-up are the same strategy — the only difference was how much you happened to have on.

Three things to measure in your own sizing

You don't need a model to audit this. You need your closed trades and a willingness to look at quantity, not just outcome.

1. Size consistency. How much does your trade size vary around your own baseline? The clean way to express this is the coefficient of variation — the standard deviation of your size divided by its average. A low number means a repeatable bet. A high number means you don't really have a position size; you have a mood.
2. Oversizing. Which trades sit far above your normal size? These are the ones that do outsized damage when they go wrong. A single trade at four times your baseline can erase the careful work of dozens of normal ones. Count them, and look at how they resolved.
3. The dangerous one — size versus outcome. Is there a correlation between how big you went and how the trade turned out? Specifically: do you size up into losing trades, or after a loss? This is the pattern that quietly converts a winning system into a losing account.

Why averaging down and revenge-sizing destroy expectancy

The third measure matters most because of a behavioural trap nearly every trader falls into. There are two flavours, and both push your biggest size onto your worst trades.

The first is averaging down: a position moves against you, so you add to it to "improve your average." The trade was already losing, and now you've made it your largest. If it keeps going against you — which is precisely what losing trades do more often than winners — your single biggest loss arrives on a position you deliberately inflated.

The second is revenge sizing: you take a loss, feel the need to win it back, and double the size on the next trade to recover faster. Now your largest bet follows your worst emotional state. Even if that next trade is a coin flip, you have systematically loaded risk onto the moment you were least able to judge it.

The market doesn't punish you for being wrong. It punishes you for being biggest when you're wrong.

Both habits create a negative correlation between size and outcome — and that correlation is pure poison to expectancy. Your wins come in small, your losses come in large, and the per-trade math that made your system profitable is silently inverted. As we cover in our piece on whether you actually have an edge, expectancy is the number that decides everything; sizing is the lever that can quietly flip its sign.

The fix is boring, which is why it works

Sizing discipline isn't a clever technique. It's a rule you decide once, away from the heat of a live position, and then refuse to override.

• Use a fixed unit or fixed-fractional rule. Either trade the same quantity every time, or risk the same fixed percentage of your account on every trade so size scales with equity but stays proportionally constant. Both collapse your coefficient of variation toward where it belongs.
• Never increase size to recover a loss. No averaging down on a losing position, no doubling up after a red day. The size you take is set by your rule, not by your last result or your current feelings.
• Treat oversized trades as exceptions you have to justify. If a setup truly warrants more, that should be a defined, rare tier — not an impulse. If you can't name the rule that permits it, you're back to trading a mood.

The payoff is that your edge finally gets to do its job. With a repeatable bet, a good system produces a smoother equity curve, smaller drawdowns, and — over a sample — the returns the math always promised. Once your sizing is steady, it's worth checking how much risk you're actually being paid for; return-per-risk tells you whether the smoothing is translating into a better deal.

Fix your entries if they need fixing. But before that, make every bet the same size. A mediocre system sized consistently will quietly outlast a brilliant one sized on instinct.