Comparing Risk-Managed and Unmanaged Strategy Simulations
Summary
The document considers how to study what happens after a stop, target, or other risk rule closes a position. For a formal strategy, it proposes running two simulations in parallel with identical entry logic: one uses bracket exits and the other does not. Comparing the resulting trades and equity paths can show how exit rules affect simulated performance, provided risk limits do not alter whether the system can place its entries.
A separate suggestion is to replay historical trades under alternative trailing stops when the original strategy cannot be formalized. The discussion notes that a simulation still needs an exit mechanism, and that examining hypothetical post-closure price paths may be simpler than running a full strategy simulation. These are backtesting approaches, not guidance to reactivate a live strategy after a loss or profit threshold. No performance evidence or method for selecting optimal stop values is supplied.
Key ideas
- Run otherwise identical strategy simulations with and without stop or target exits to compare their effects.
- Keep entry logic aligned, while recognizing that risk constraints can change which orders are allowed.
- Historical trades can be replayed with alternative trailing stops when the strategy is not fully formalized.
- The simulation still needs a rule for closing positions, and the document gives no evidence for an optimal setting.
Tags
Full text
# How to reactivate a risk mangement rule in an automated process # How to reactivate a risk mangement rule in an automated process If some conditions are met (stop loss, trailing stop, take profit...) we will close ours positions (sell/buy) to avoid having more loss or to ensure profit. In an automatic trading system, it is easy to set-up these rules. But is there an automated strategy / rule to re-activate our strategies? For example: - simulate our P&L if we have no risk management rule; and then reactivate ours positions if the simulated drawdown or run up reaches a certain value - reactivate after two days - reactivate at the end of the current month? Or is this part of an "automated" trading in fact manually done? ## Answer by Matt Wolf (score 2, accepted) https://quant.stackexchange.com/a/4860 I strongly assume you are talking about a simulated environment, only. Risk management rules are there for a purpose, adhere to them or throw them overboard entirely at your own peril. Having said that if you want to purely simulate how P&L, equity curve, and draw-down among others would have behaved had you not squared the positions due to targets or stops reached then you could do the following: Simply run identical strategies in parallel, same strategy logic inside except that one generates orders with attached targets and stops and the other generates orders without the bracket order. In that sense you will generate the same trade entries (given you are not subjecting your system to constraints that would prevent the system from generating orders because certain risk limits would otherwise be breached) but the exits will be different. However, in the end you still need to set a mechanism of how to close positions. If you are merely after investigating how P&L had traversed n-time units post the trade closure, given the trade was not closed, then there are way easier things to run for example in R. ## Answer by IgorS (score 0) https://quant.stackexchange.com/a/7622 I agree with @Freddy, that we are talking about backtesting trading strategy in simulation environment. Somehow you opened those positions - that means you have some trading strategy, the strategy which gives you enter (and probably exit) signals. There are two cases: - You CAN formalize and automate this strategy. - You CAN'T do it for any reasons. (1) What you are doing in the first case: is getting some trading simulation software (like Amibroker) or develop some backtesting utility yourself. And simply overlie your trading strategy signals with different risk management indicators (trailing stops or some bands) and get optimal stop values. (2) Use a set of historic trades which you already did as an input for simulation tool. And again run it with trailing stops.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.