Patrick, thanks for taking the time to check it out with your contact.
His response, though conceptually valid, misses the point IMO, in that it assumes that I'd treat the model/system as infallible. Although I have complete faith in the system since it has performed well over more than a decade, I still recognize that it can go wrong and therefore I use money management stops.
The reality is that over the past year and so far this year, I have been stopped out 4 times in the case of the spoo and in each instance, if I had not used a money management stop, the trade would have ended up profitable or at the least not losing as much money as resulted through the money management stop. Given that I still insist on a money management stop, the chances of a catastrophic loss is negligible. As Fred pointed out, LTCM was leveraged to the core and as best as I know money management stops were not a factor -- the assumption by the operators was that the models could not fail -- and that is an assumption that I would never make no matter how well the model performed.
>>The bottom line here is that the past never EXACTLY repeats itself, but even a perfect retrospective model would be a slave to the past<<
I agree but as I stated that is why the use of a money management stop is an imperative.
>>the human decision-making capability inherent in any skilled trader would not be enhanced by arbitrary constraints, e.g., in this discussion, I will trade X contracts, no more and no less<<
Again, I agree -- but a disciplined trader is less likely to stray in this regard. If one equates the number of contracts to be traded with the observation of system signals, I cannot recall the last time that I did not take a signal when the system generated one -- other than if I was on vacation or a similar circumstance. So I have no doubt that once I have a validated model, I'd follow the rules and trade the number of contracts that are called for per the model.
But I do appreciate your taking the time to try and sound out your friend -- even though his response was that of the typical academician. Not really a surprise in that I asked my brother -- who is also a professor of mathematics in the UK -- and he reacted on the same lines ie questioning the validity of any such model. However, I have asked him to humor me and give me a response despite his belief that a model based on past behaviour will invariably fail. It remains to be seen whether he will do so. I suspect that he will at some point when he is bored -- and if I bug him a couple of times!! |
