GTD, Trading Systems and Trust

I'm a fan of David Allen's "Getting Things Done" system. I don't actually follow it very well, but at least I like the theory. Yesterday I was listening to some podcasts from the 43folders website. David was talking about patching leaks in the GTD system. The whole point of GTD is to get everything out of your head and into folders or lists or whatever you're comfortable with. You process all the stuff in your life into a manageable system. He said many people will do most of it, but for whatever reason they won't completely trust the system. They're afraid they'll miss something so they leave a piece in their inbox rather than process it. Then they'll do the same for one more thing and one more until the system falls apart. It's not because the system is broken, they just didn't follow it.

So what does this have to do with trading? I was trading the emini Dow futures (YM) intraday today. As I explained in an earlier post, I don't have these trades automated. I have rules for them, but there's a good amount of discretion there as well. I only took one trade today even though there were at least three setups that I could see. The one I took didn't work out so great, the two I didn't take did. All three setups were pretty good, but "something" told me not to take the two that did well. It may have been the TICK readings or the advance/decline line...I couldn't really tell you for sure. It doesn't really matter. The setups were there, I just didn't trust the system. One trader I know used to say that the most unreliable part of any system is the pink blob that drives the mouse.

So how does one trust a system? Do you trust it because some author said such and such pattern usually works? Or someone said some indicator predicts price movement (just a hint..if anyone says something predicts price, run away)? I do believe there are some really good discretionary traders out there - guys that just have a really good feel for market movement. Unfortunately, I'm not one of them. I need something that has concretely defined rules that I can backtest. Then I can see what percentage of trades were winners, how big the average winning and losing trades were. And whether or not I can expect to make money over time.

I've seen some traders debate the usefulness of backtesting. Their argument is they aren't concerened with what happened in the past because they do not believe history will repeat itself. In the "Way of the Turtle", Curtis Faith addresses this very issue. To those traders Curtis asks these questions: "What is the alternative? How do you arrive at any strategy without knowledge of the past? How do you determine when to buy or sell? Do you guess?" I am firmly convinced that some form of testing is essential to the success of any system. It does not have to be completely computerized, although if it isn't, be prepared to spend quite a bit of time in the process. Computerized testing also has the advantage of not introducing bias into the results. If you do it manually, be sure you aren't cherry picking trades.

If you know what 'normal' is for a system, it's much easier to trust it. If you've had 3 losing trades in a row but you know that in the past there have been as many as 10 losing trades in a row yet overall the system is profitable, it's a lot easier to take that next trade.

All that was a lengthy way to say I really need to code up my intraday system.