Adrian Reid explains why tweaking indicators won't fix a broken trading system. When a system hits a new maximum drawdown, like during the COVID crash, you need structural changes or new rules to handle conditions your backtest never saw, not just parameter adjustments.
I generally don't tend to, you know, tweak a moving average or tweak an RSI or something like that to fix a system. If I re optimize a system because I just do that periodically to just check, okay, is the system still set up in any sync with the market, then that's kind of different. But the response to the system breaking, achieving a new maximum drawdown.
Is typically more than just re optimizing. It's something structural. I tend to approach it that way anyway. It's kind of hard to optimize your way out of a broken system. But what you can do is say, ah, this set of market conditions has never been seen before. It wasn't in the backtest. And clearly there's a weakness in this system. If that sort of behavior happens in the market, that would be like an extra rule as opposed to a re optimization. I think that's okay. Like the COVID drop, which was the market was at an all time high and this fell into a bear market which hasn't happened before.
Without some, you know, previously. It's sort of, you know, the market has some dips and rallies on its way down.
But that Covid crash sort of illustrated a new type of market behavior which caused some pain for some systems which could lead rise to different rules or new rules. I think that's a good way to do it as well. But it's really a, you've got to put your Sherlock Holmes hat on and investigate what broke it, what changed, and then think, okay, is it optimization, fine tuning? In which case this system maybe isn't robust, or is it new market behavior which the system wasn't prepared for, in which case, can I add some rules or adjust the structure of the system to survive those rules, those market behaviors? Or is it something that you just don't understand and the system's not working anymore and you can't see any real reason for it. And that's the most dangerous one because there's some probably some underlying market behavior or dynamics that have shifted that have just caused your system to fall apart. And in which case if you can't find a rational reason, usually it's time to shut off the system and look for something else.
