Compounding

A percentage is a multiplication, and multiplications applied in sequence multiply their factors rather than adding their rates. A fall followed by a rise of the same percentage does not return the starting figure, because the rise is applied to a smaller base than the fall was. The order the two are applied in makes no difference whatsoever, which is the property that makes the effect arithmetic rather than psychological.

The consequence for an account is that a loss is harder to recover than it was to make, and the gap widens as the loss deepens. Recovering a given percentage fall requires a larger percentage gain, and the required gain rises faster than the fall that produced it. This is also why a run of losses decays rather than running in a straight line when risk is set as a fraction of current equity: each step down is smaller in money than the one before it, because the fraction is recomputed against a smaller base.

The same arithmetic runs in the favourable direction and is the reason a sequence of modest gains on a rising base outruns the sum of its rates. Nothing about it makes an outcome more likely, and it does not turn a method with a negative expected result into a positive one: compounding scales whatever the underlying process delivers, in both directions, and on a leveraged position the losses it scales are not limited to the amount deposited.