Risk of ruin looks different when costs eat half the edge
Posted: Sat Oct 31, 2026 7:27 am
Textbook risk of ruin formulas assume the edge is known and stable. For scalpers, costs shrink the edge so much that the numbers can flip.
Suppose a strategy wins 55 percent of trades with equal win and loss sizes of 3 pips. The gross edge is 0.3 pips per trade. With 0.8 pips of cost per trade, the net expectancy is negative, and ruin is not a risk but a certainty.
Even small changes in costs shift the picture. Reducing average cost from 0.8 to 0.4 pips turns a losing system into a marginal one. Reducing it further makes it viable.
That sensitivity means risk estimates have to use net numbers, measured from live trading, not gross results from a test. And the estimates need updating whenever spreads or commission change.
A margin of safety is wise. Assuming the true edge is half the measured one gives a more conservative position size.
What net expectancy is measured, and how sensitive is it to a rise in costs?
Suppose a strategy wins 55 percent of trades with equal win and loss sizes of 3 pips. The gross edge is 0.3 pips per trade. With 0.8 pips of cost per trade, the net expectancy is negative, and ruin is not a risk but a certainty.
Even small changes in costs shift the picture. Reducing average cost from 0.8 to 0.4 pips turns a losing system into a marginal one. Reducing it further makes it viable.
That sensitivity means risk estimates have to use net numbers, measured from live trading, not gross results from a test. And the estimates need updating whenever spreads or commission change.
A margin of safety is wise. Assuming the true edge is half the measured one gives a more conservative position size.
What net expectancy is measured, and how sensitive is it to a rise in costs?