Instead of "sequential" consider "simultaneous". As follows. Lots of mechanical systems have stops that are a function of recent market activity. For example, a stop 2*ATR below the entry, or a "Chandelier" 4*ATR below the highwater mark, or a stop at the middle Bollinger band when entering at the upper Bollinger band. (As opposed to systems whose stops are a fixed distance away like "a $1500 protective stop" or "a stop 3 points below entry"). If your collection of systems has stops whose distance varies as market conditions change, then perhaps you can run all of your systems, collect all of the entry signals they generate, and only take the 5 signals (or the 2 signals, or ...) whose risk today is lowest, i.e., the smallest dollar-distance-to-the-stop. Now you can have a very large collection of systems and run them on a very large portfolio of tradeable instruments, generating a very large pool of possible entry signals. BUT you only need an account big enough to trade the 5 lowest risk signals (or the 2 lowest, or ...) Naturally you will want to test this idea to see how well it performed on past history. It will require testing software that runs multiple systems on multiple portfolios, at the same time. My trading friends and I have tested it on a collection of 3 systems plus 50 futures markets and in my opinion it worked quite well. Additionally there is a systems vendor who is marketing a product-for-sale named "relativity" that does the same thing ( http://tinyurl.com/3d7bec ). Read his flowery advert for persuasive arguments that this is a jolly brilliant way to trade. But remember, you can use the ideas he discusses (praises!) without buying his product.
I find that an interesting idea, and hadn't thought about it. The only problem I have with it (for my specific purpose) is that it introduces a rule, which creates additional uncertainty. I.e. the rule might show great performance in a backtest, which might not continue in the future.
If you are limited on funds, you should put together a portfolio of the best performers on as many markets as possible. When you increase, add another market if possible. Try to avoid trading a majority of your money in any given market.
What about trading the system that is currently in the largest drawdown? Could reeval every 3, 6, or 12 months. Can you test that?
Doesn't this just mean that you will be stopped out more often? That wouldn't in and of itself be a bad or good thing, but you'd need to test it to see what the results were.
This can be a dangerous method because you never know whether a system is drawndown or broken. You must have alot of confidence in the systems FUTURE performance.
I agree. As soon as you shuffle through systems based on some rule, you get in trouble if the rule stops working in the future. That's why I suggested random shuffling.
I'm not sure about random. I've never really tested the idea. But why not trade the best system? If you have to pick, pick the one(s) that you have the most confidence in. Out of curiosity, do you have one system that trades in multiple markets, or multiple systems in one or more markets?
I think your confusion stems from the fact that you still haven't arrived at the best system that - 1. Matches your Psychology 2. Your return expectations. If I were you, I would go back to drawing board and continue my research. Because as soon as you have created THE SYSTEM for yourself, all questions would disappear and you would simply start applying it.
That is consistent with JimmyJam's advice to focus on a single system and improve it. I find that a strong argument.