New papers on stop-losses and other order types
Posted: December 8th, 2025, 10:00 am
Hi guys, I'm a new forum member and am shamelessly plugging some research I have done on trade execution. I did it over the past year after becoming interested in Ed Thorp's work on gambling.
My first paper is on stop-losses, and it is supposedly going to appear in the May edition of Wilmott. It focuses on the mean of the P&L distribution. I have another paper on risk (variance of returns, etc.) and another paper on the strategy with a stop-loss + take-profit order + a half dozen other things in various stages of completion.
TLDR: Using too tight or wide a stop can kill your P&L, but there is an optimal value if you are prepared to take a view on the evolution of the drift. The optimal value can have a better return and lower variance than not using a stop-loss. You can find the optimal value as the solution of a simple 1D optimisation problem and in a simplified situation (having both a stop-loss and take-profit order) there's an explicit formula, a bit like the Kelly-criterion, except that it's for a stop-loss. Similar results are available in the buy-stop order, meaning that there are real situations where it can be advantageous to `time' the market. In essence, the delay in your share purchase increases the chance that you're on a positive drift path.
My first paper is on stop-losses, and it is supposedly going to appear in the May edition of Wilmott. It focuses on the mean of the P&L distribution. I have another paper on risk (variance of returns, etc.) and another paper on the strategy with a stop-loss + take-profit order + a half dozen other things in various stages of completion.
TLDR: Using too tight or wide a stop can kill your P&L, but there is an optimal value if you are prepared to take a view on the evolution of the drift. The optimal value can have a better return and lower variance than not using a stop-loss. You can find the optimal value as the solution of a simple 1D optimisation problem and in a simplified situation (having both a stop-loss and take-profit order) there's an explicit formula, a bit like the Kelly-criterion, except that it's for a stop-loss. Similar results are available in the buy-stop order, meaning that there are real situations where it can be advantageous to `time' the market. In essence, the delay in your share purchase increases the chance that you're on a positive drift path.