Most off the shelf market making software will allow you to simulate your trading prior to going into production mode. However, this is a matter of buyer beware as most people selling the software will say things like, "Wellllllll... everyone tells us our simulator is very accurate to actual results." If you want you can believe them, but back testing an mm strategy is an exercise in futility as it is impossible to know what kind of fills you are actually going to get.
You can assume various kind of fills and then simulate these scenarios. Besides that, your completely right... Backtesting an mm strategy is more about making sure that it operates correctly.
if you assume worst case fills (buying the offer/selling the bid or you are always last in the queue) and you make or barely lose money then its probably not a terrible strategy.
I don't use off the shelf products - I like to code all my own so that I fully understand what it's doing. For price-taking model, I am pretty confident that life and my backtest match. But I am now wondering whether there are some ways to add fill logic (for FX) to a market-making model that would at least give me a ball-park idea of whether the strategy might produce positive returns. When I say market-making - I am not trying to provide liquidity - I am trying to enter and exit my positions without crossing the spread.
If you write your own stuff I would say sim-trading may be worth a shot. Back in '04-'05 I had experience with a few different banks allowing one prop firm in particular to sim-trade their FX feeds with the firm's custom software. Of course they knew there was sufficient capital behind them should simulation appear profitable. The model was actually price taking so simulation proved to be accurate in this case. Anyhow, I am not sure if banks allow this anymore or not haven't really heard anything about it for a few years. I like the thought rosy had. Should just do it that way.
Rosy2, I am curious. Was your observation just a thought - or have you tried and found it to be workable?
Does FX give you the size of the bid and ask? My understanding was that FX had a more opaque operation where everything basically depended on what dealers you were going through. Without an idea of the size at the bid/ask there would be no way to tell where you were in the queue. With a stock on the other hand, if you were going to enter an offer at the ask, you could mark the size when you entered your order and count the trades at that price to know if you were to be executed.
i dont know anything about spot FX; i always thought people paid the bid/ask spread. But I do what I wrote for exchange traded stuff