Most of the retail trading community spends ninety per cent of its attention on entries. What indicator to use. What pattern to wait for. What oscillator setting works best.
This is the wrong allocation of effort.
The single most important determinant of a strategy's risk-adjusted performance is how much you put on per trade. Get sizing right and a mediocre entry signal will be profitable over a meaningful sample. Get sizing wrong and the world's best signal will eventually blow up the account.
Three sizing frameworks worth understanding
Fixed fractional. Risk a constant percentage of equity per trade. Simple, robust, almost impossible to mess up. The weakness is that in a high-volatility regime you size as if it were a normal-volatility regime, so your dollar risk per trade silently grows when conditions are most dangerous.
Volatility targeting. Adjust position size inversely to recent realised volatility, so dollar risk per trade is approximately constant across regimes. This is what most professional managed-futures programmes do. It scales down naturally in stress and scales up when conditions are quiet. The weakness is that it can become procyclical near regime transitions, when realised vol is still calm but the world is about to break, the vol-targeting rule keeps you fully sized.
Kelly fraction. Mathematically optimal but extremely fragile. The pure Kelly formula assumes you know your win probability and your win/loss payoff ratio precisely. You do not. Estimation error in either input produces wildly over-aggressive sizing. Most professional users run Kelly divided by four or by eight to leave margin for the inevitable estimation error.
What we do
At Bostock, vol-targeting is the default chassis. Position sizes are scaled inversely to a blend of five-day and twenty-day realised volatility, with a smoothing factor to prevent over-reaction to single-session moves.
A Kelly-fraction overlay sits on top, capped at a hard upper limit. The overlay raises size when the underlying signal is strong and recent vol has been low. The cap prevents over-leverage when both happen at the same time, which is a state in which sizing models are most likely to be wrong.
A regime-aware floor adjusts the target vol down when the cross-asset correlation indicator we wrote about elsewhere on this site moves above its threshold.
A mediocre signal sized well will outperform a brilliant signal sized badly. This is not a tradable secret; it is a discipline.
The practical implication
For anyone trading their own account, the practical implication is uncomfortable. Spend more time on your sizing rules than on your entry rules. The work is much less glamorous. The maths is genuinely harder. The payoff is much larger than another fifteen hours studying RSI settings.
The single most expensive mistake in retail trading is using the same dollar size on every trade. The cheapest improvement most traders could make tomorrow is to scale that size by the volatility of what they are trading.