Thanks for the insight plum. Since I dont have historical data on A/D, etc, im trying to maximize the gains on a purely mechanical system only using candle data as the first step. Then look back and use a discretionary approach to occasionally pull the plug. But if I can somehow tweak the system to pull the plug itself without giving up radical profits in other areas, that would be the best of both worlds. Still have some analysis to do. The pre-market may give me the clue I need. 4/20/2000 took a hard dive right at the opening 30 mins before stocks start trading. Will dig deeper in this area. peace axeman
Folks forget that it's the markets job to keep the most people off balance the majority of the time. If you're confused, gun shy, and afraid to ride the obvious trend, then Mr. Market is once again doing his thing. When it gets easy, (for about 5 minutes usually) then Mr. Market will find a way to shake the fleas off his back. And knowing this doesn't make it that much easier.
(1) Does the system still have a profitible year? (2) Is the position so large that a day like that would wipe out the account? Sounds like more a of money management issue. Position should be small enough to limit catastophic failure to 20% of total equity, something along those lines. On the indicator side, if you must tweak it, (1) did the system make an unusually large number of trades that day? (2) did the system have an unusually long string of losses? Try a statistical approach. In other words get a quantitative measure of the system's normal behavior. Turn it off when it deviates. Just something to try.
If the problem is that every time it tries to trade it loses, try adding a line to the code that after say 3 (or whatever number you decide) consecutive losers in 1 day it stops trading or something like that? or a max intraday drawdown or something big enough to work inside and small enough to save you? There may be other days that get saved as well, but they might not be so obvious. Just a thought... Natalie
having a system that only goes long is like driving a car that only goes right... i would find a new system to begin with then trying to revise one that can only be 50% successful, and thats if its right every time...
I have found that the dynamics of going long and short can be very different. Therefore I model them seperately. Taking a profitable long only system, and then reversing the logic, does not always make the short version profitable. And I never said that I didnt have a system very similar to this one based on the short side, that only goes short peace axeman
i would some how implement a short strategy to compliment your long strategy... just to better your odds...