Two traders can run identical entry rules and end the year with completely different outcomes, because sizing decides how much each decision is worth. It is the part of trading with the largest effect on results and the least glamour.
These concepts all answer versions of one question: how much of your capital should be exposed, to what, and at the same time as what else.
Deciding how much capital any single trade may risk. Fixed-fractional sizing, where each trade risks a set percentage of the account, is the most common approach because it scales down automatically after losses. The specific method matters far less than having one at all.
Dividing capital across strategies rather than trades. The mistake is allocating by recent performance, which systematically overweights whatever just had a good run and is most likely to mean-revert next.
Measuring how strategies or assets move together. It is the quantity that determines whether diversification is real. Crypto correlations also rise sharply during stress, so a portfolio that looks diversified in calm conditions can behave as one position exactly when that matters.
Running strategies that make money in different conditions, such as a trend strategy alongside a range strategy. Done properly, one is usually working while the other struggles, which smooths the equity curve without requiring either to be better.
Choosing allocation weights to maximise return for a given level of risk. Classical mean-variance methods are notoriously sensitive to their inputs, and small errors in estimating expected returns produce wildly different weights, so simpler and more robust weightings often outperform in practice.
Periodically returning allocations to target after winners and losers drift them apart. It enforces selling strength and buying weakness mechanically. It also generates fees and taxable events, so rebalancing too often costs more than the drift it corrects.
Controlling the depth and duration of losing periods rather than only the return. Deep drawdowns are dangerous less for the arithmetic than the behaviour: almost everyone abandons a strategy near the bottom of one, converting a temporary decline into a permanent loss.
Sizing rules only help if they are applied when it is uncomfortable to apply them. Putting them on the execution path removes the moment where a human overrides the plan.
Your keys, your account, your rules. We only send the orders.
