Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
You've given a thorough and helpful exposition of your approach to system testing and management. There's really not much more you can to do help those willing and able to learn what you've generously decided to share. It was all very clear in theory, and the real life example you carried out made it clear in practice.
Personally I think this thread should be a sticky. This is definitely one of the best threads on the forum. If it isn't made a sticky, I think you should just bump the thread every now and again so that people who haven't seen it get a chance to find it.
Neither real traders nor promising traders-to-be care whether this system succeeded publicly or not: it should have zero impact on what they think of the thread or of you as a trader. And you strike me as someone who has real human connection off of message boards, so there is no reason to care whether the thread is "popular." I can't see a reason to continue the thread with a profitable system unless you feel you have something to prove - which you don't.
If you do decide to share anything else on this forum, though please post an announcement to this thread so that I can come check it out.
Food for thought: Whenever you add a contract it is by definition after a winning trade. In most cases, it's after a bit of a run. When you choose a point at which to "start", is it also after a run? My guess is that it is given your description of the process (e.g. waiting for good performance in the incubator, replacing strategies periodically with ones that are "more promising", etc.). All of this conspires to have you doing "starts" and "adds" triggered by runs. Have you considered the impact of this on the results immediately afterward?
As a further aside: Live trades are not actually randomly pulled from a bag and they are not as independent as we'd like to think. There are good and bad periods. (If you need an example, just look at EC vol recently.) In a sense, the market "knows" that your style has been receiving above average performance lately and that other styles have been seeing below average performance. Given the mean-reverting nature of "average" performance, the market may actually be "out to get you".
This is a really interesting take on the old (and generally invalid) idea that the market is "out to get you".
I would normally have said that the market doesn't know you and is not after you, but in the sense that performance tends to revert to the mean (whatever the mechanism), perhaps there is something to it, in a different way.
I'd be interested in @kevinkdog 's take on this. Do you see enough of this kind of mean reversion in your testing, and (a) does it matter, and (b) what can one do about it?
Very good point, especially the part about changing to strategies that look more promising. I see that a lot on some investor signal sites, where everyone piles into the latest and greatest looking system. They constantly jump from system to system, usually at the wrong time.
You are also correct that contracts are added after good runs, and many times are right before a bad run. That's happened to this system twice - a run up, followed by adding contracts, then a run down. Both times this happened, it would have been better to just stick with one contract.
So, maybe a legitimate question is: Do you include some position sizing rules to add contracts after a down run? Nothing as extreme as Martingale, but maybe something to take advantage of the normal ebbs and flows in a system. On the downside, it could be like trying to catch a falling knife.
I have definitely experienced that. Maybe there should be 2 bags: 1 with all the "high volatility" trades, and 1 with the "low volatility" trades. It might turn out that the under-performance this system has been seeing is really not as bad as it looks, given the current market volatility. The trick may be defining these "high" and "low" areas. It might not be hard to after the fact, but it could be tough in real time.
Thanks for the comments. Very interesting insights.
I definitely see mean reverting behavior in systems I trade. The big question I always have is "what is the actual mean the performance should be reverting around?"
Here is an example with a system I trade live. You can see that the actual performance does revert to a mean line, and the mean line is upward sloping (obviously that is important).
Just to be clear, the mean line I show is NOT a regression line. Before I started trading this system, my historical testing said I should make $XX per month. That is what I consider the mean line.
Here is another I trade live. I am leery of increasing size on it right now, because the actual performance is way above the mean line. If I had to guess, I would think the next 6 months will bring flat or down performance, and get closer to the mean line. If that is true, there is no sense in increasing size right now.
My two cents... I would disagree with adding hysteresis to your money management protocol. No matter what money management method you choose, there's going to be a point where you go from 1 contract to 2. If you delay this transition and catch a run, you are leaving a LOT of growth on the table. (Go ahead, experiment with the long-term impact of deferred compounding early on.) If you delay going back to 1 contract if you get unlucky, you've violated your initial risk parameters. This is a key disadvantage of this kind of money management, but I think it's worth it.
In essence, you're trading the pain near the contract add/remove thresholds (until you get past 7 or so) for maximum compounding. Repeat enough times and I think you'll find you're better off just accepting the chop at the thresholds as part of the process.
You should be congratulated, by the way. This thread is absolute proof that you have the discipline and confidence that so many lack, and both are necessary if one is to be a successful trader over the long term. Well done!
Regarding your two charts, be careful. There are two "averages" at play. The first average is the actual "realized average" for that set of trades. It's a straight line connecting the first and last dot. The other average is your "expected average". They are two different things with entirely different interpretations.
In the latter graph, you are observing that the "realized average" is different from your "expected average". This can occur for one of three reasons... pure chance (you can compute the odds of this using MC), your expected average is too low/high, or something has changed in the market. Once pure chance is ruled out, I default to "my expectations were formed incorrectly" and examine my development process for leaks. I have control over my process but do not have control over the markets, so time spent worrying about that is not productive.
I wish I had better answers for you regarding how to take advantage of the mean reversion idea, but even after contemplating this for quite some time many open questions remain for me. Ultimately, about the only conclusion I've drawn is that unless you know the mean with some degree of certainty there isn't much action you can confidently take.