I have been contemplating the value of backtesting in in this field of trading. As options traders, we have the ability to structure trades and adjust to whatever the spot is doing, unlike those trading spot. Hence, I feel only a working knowledge of the underlying is required. I have been getting into backtesting in FX, futures, and a little bit of equities, but barring using options as a sub for spot in directional trading, the complexity of options and myriad of possibilities makes for backtesting in non-directional trading pointless, IMO. I honestly don't know of any options backtesting software. I'm not asking for software with those capabilities. But rather since we can adapt our positions to whatever spot is doing, and options being multi dimensonal, defining a set of variables in backtesting seems moot. I understand why spot traders do backtesting, but for us, seems to be time better used. Note that I mean backtesting as in the general backtesting sense, where script junkies set insane parameters to optimize or whatever, not backtesting in the sense where you would test and see how a position would perform under differing market conditions. Your thoughts are appreciated. Please point out any ignorant parts of my post.
Supposedly OptionVue can do backtesting, but I tried the demo for a while and didn't have much luck. I've seen systems based entirely on volatility. For example, there was a system in TASC magazine a while back (maybe it was ActiveTrader?) which would sell premium if the IV exceeded a moving average of the IV by 100%. The author had 3 or 4 recent examples that worked really well. It seemed kind of interesting to be able to test such ideas.
FA, I understand your point that systems can be implemented in options, take that IV example for instance, but directionally, I can replicate the underlying using options and then some. Makes no sense to try and learn to trade the spot (unless intraday), if options can be better used. Seems to me if one learns the nuances of options, spot is pointless, unless used in a combo. And if so, backtesting is pointless, one only needs a working knowledge, eg look at chart.
Cyclicality of vol is the only method which lends itself to backtests. Beyond that, you're much better off simply testing direction or neutral price systems on spot and applying the best greek scenario to fit your predicted outcome. Numerical methods don't work when backtesting convexity -- test stat and implied vols, but not the premium.
Possibly a more interesting backtesting scenario is to refine option models to better fit the actual data. For example, there's a clear inverse correlation between IV on the S&P and the futures price. Creating a more accurate pricing model would seem valuable. On the other hand, I've seen quite a few academic papers, particularly on Levy models to deal with leptokurtosis, which try to create a model which best matches the prices used by the market makers. It seems a better approach is to find a model which more accurately prices the option and finds errors in the market makers' models.
I wouldn't say it's pointless - just dangerous - and if you see this point then that's the point. I think it's much the same with stock backtesting - nothing quite like real-time experience. I would suggest observing the actual option price charts and chains at various strikes from 90 days to expiration. Besides, by simply viewing hundreds of stock/option charts showing the intraday, daily, weekly, monthly, yearly and even 10 year history of HV and IV, I have seen phenomena which I would have NEVER thought of in backtesting. It has helped me to understand risk and the importance of knowing your risks (greeks) and your risk management gameplan. Now, we could of course just test for the best results, but I think this leads to ignorance as well. I would have seen the same results from visual backtesting as well as all the permutations of false-positives. There is NO substitute for watching the Greeks, option prices, news effects on IV, day-to-day gap effects, adjusting strategies etc. Best of luck.
to the best of my ability I test before putting money towards a system. with options backtesting, you almost always have to make assumptions. then you try to fill in the gaps with experience.
This is how I see the value of backtesting. Advantages - 1)gives you the confidence to pull the trigger. 2)reinforces a systematic approach to trading Disadvantages - 1)makes you think you've got a systen that works and can thus lull you into a false sense of confidence. 2)doesn't really work well with options for reasons given by some of the other posters. Overall, I've stopped backtesting because I feel it is time consuming and doesn't really offer me much benefit, i.e. poor reward for effort employed daddy's boy