Very cool - I've been a developer for a little over 30years myself (I was a video game programmer in a former life). Some of the real time techniques used in most games actually translate over to trading bots surprisingly well (thank god I don't have to code the stuff in 6502 asm). I certainly would appreciate any sage advice you'd like to share - feel free to PM me anytime. I'm a bit of a voracious reader, so if there's any trading or optimization books you've found especially helpful along the way, I'd love to check them out (so much of the trading literature out there is total dreck). Thanks again!
Sure - its simply holding your position after entry for some fixed amount of time (seconds, minutes, hours, days, etc). Your position gets stopped out by a time rather than price threshold.
As requested, here are some comments on some of your conclusions that I disagree with. Some of my disagreements are because in many cases how you frame the problem determines the answer. For example, most of my research is futures focused and I rarely exit a trade before the close unless I am stopped out. Just so you know where I am coming from. I think if you were either looking at shorter timeframes or looking at stocks, you might come up with different answers (or more likely different degrees of the same answers). I can pretty much guarantee you that this isn't true. All it takes is one example for that logic to end up on very shaky ground, and I have come across multiple examples of money management schemes that have turned a winning system into a losing system. This is something that I think depends on the system. Some of my profitable systems behave in accordance with what you are saying, but some don't. I have a system with a pretty tight stop and no profit target (it doesn't exit a profitable trade until it hits a signal going the other direction) that makes a bunch more money with that money management setup than it does when I remove the stop or add a price target or set a time stop or whatever else. I also have a system without a price stop but with a profit target and a time stop. If I explained the theory behind this system and why it works it would be obvious why I picked this particular scheme, and it in fact had by far the best results in my testing, as I thought it would be before I even started my testing. Just because of the characteristic of the non-random market behavior that it exploits. I think your best bet is to compile a library of money management schemes and try them out on all your systems. Study the results when compared to each other and really understand why one works better than the others for this particular system. I can tell you right now that out of the many, many ideas I have tested, I have never seen scaling out have a better expectation than exiting everything at once. I agree with Buy1Sell2 in that your goal in developing a single system isn't to smooth returns, it is to maximize expectation. Trade multiple systems concurrently to smooth your returns. I didn't comment on this earlier, but this is a point I also agree with in general. There are only so many hours in the day. Disclaimer: I am not an old hand at this and haven't been profitable that long in the grand scheme of things, so take my comments with a grain of salt.
Thanks for taking the time to share your insights - much appreciated. I currently bot-trade stocks intraday, but am currently backtesting strategies on NQ. Since I don't really want to go prop, the extra leverage of futures, along with the preferred tax treatment is very attractive to me. But regardless of the instrument, I prefer to be flat at the close and hold no overnights. I actually agree with you on this - I should have phrased that more clearly. I meant any of the *tested* exits (and subject to the market model I was using), rather than all possible exit techniques. There's certainly countless ways to screw up a perfectly good edge. Agreed - I can imagine there are systems where bracket orders make sense (ie: scalping systems). But in my experience at least, I've never come across a system that was enhanced by adding bracket orders. But the test I ran did seem to indicate that if you must use a stop, its generally best to not use a target and let your winners run. This makes sense intuitively as well. Just to make sure our terminology is in sync, when I think money management, I think position sizing, rather than stop losses based on % equity risk - is this what you're referring to? I've played around with position sizing a bit, but have always found that fixed fractional position sizing seems to work well. I currently don't martingale or average down, but I'm open to the idea (no doubt I'll get flamed by someone for THAT disclosure, lol). Anytime I've backtested stops, I've always found that getting rid of them improved my results. But there is obviously a very large universe of possible systems out there that I've yet to test. Can you share any examples of the MM techniques you're referring to (without disclosing anything proprietary, of course)? But I agree that if there is any free lunch, its in trading multiple uncorrelated instruments/systems and dialing in as much leverage as you can (subject to the amount of pain you're willing to take). The trick of course is finding two or more instruments that aren't correlated these days. Everyone seems to be hedging everything against everything else... Makes you wonder if someone should just create a world ETF that has everything in it. Hedge to your heart's content...
Sorry, perhaps my terminology is a bit loosey-goosey. I am not referring to position sizing. I have spent some time researching that too, of course, and it sounds like we are generally in the same place. I usually just risk the same percentage per trade and call it good. I have found scaling in, scaling out, or any other position sizing mechanism to be inferior when shooting for maximum expectation. And as I mentioned, I think the name of the game is maximum expectation. I can manipulate my Sharpe and whatever else all day long by fiddling with things, but keeping it fixed fractional and entering and exiting all at once is ultimately the right way to maximize expectation in my experience. I am not trying to sell anything to anyone with my big Sharpe. I don't completely ignore it, but it isn't high on my priority list. This is true in many cases, but I have 1 system that most certainly performs WAY better with a stop. Think of a non-random event where there is an uncertainty in the market and you are basically playing a breakout. A week from now the market will either be up a lot or down a lot. You want to be in a week from now if you were holding the right end of that, but if you were on the wrong end, you wanted to be out as soon as the direction became apparent (and ideally you would have flipped to a position in the new direction). My system I am talking about is not any kind of traditional breakout system, but this is basically what is happening. On the flip side, I have a couple of systems with no price stops, because stops really drag down the returns. I would say definitely keep it in your bag of tricks, because you will probably come across systems where a tight stop will make you the most money. (I should clarify when I say "no stop" I actually have a disaster stop in about 80-100 points below where the market is, but this has never been hit in testing or in real life.) Yeah, I actually trade all of my systems solely on the ES right now. I haven't had any luck generating anything good on other non-correlated instruments yet, but I still have plenty of work to do there. Great point about leverage too. I feel like I am reading something that I wrote here.
Its actually really great to hear another strategy trader confirm similar findings and experiences. I hear a lot about Kelly around here, but I've yet to find a way to integrate it into what I'm doing. Fixed Fractional always seems to be the best fit and definitely keeps things simple. Is Sharpe your fitness criteria of choice? I mainly use RAR or Monte Carlo runs that factor return and drawdown. Ahh, so its sort of like you're setting a trap for the event and the stop is obviously an integral part of the whole mechanism. I assume your reward/risk on such a trade is relatively large, enough to absorb the negative aspects of the stop? Yeah, when I make the switch to futures, I'll definitely have a disaster stop two or three sigmas out at all times. I've actually built that in as a requirement in my bot "API". I think I'll eventually end up in ES - gotta love the liquidity in that puppy. The tick files are freaking HUGE. I'm just more familiar with NQ, as I watch it during my stock trading and use it as a performance benchmark. NQ appears to be a bit more volatile than ES, and has a larger volatility/margin ratio. TF seems to have even better swings, but is thinner, plus I'm too cheap to pay IB's fees for its data feed (total BS). But its good to hear that someone's trading ES successfully - the general consensus around here is that its one tough nut to crack. Hey, great minds think alike, right?
I've probably spilled too many beans in this thread already, but I wanted to comment on this one last thing. Most of my testing revolves around the ES. When I started I did this on purpose. If I was going to spend thousands of hours researching systems, it didn't make sense to spend all that time in an instrument that couldn't support large size. I mean, if it only ends up that I ultimately make a couple hundred thousand dollars in my entire life of trading, then that's ok, but my hope (like everyone else's I'm sure) is to make a truckload. If your ultimate goal is to make the most money, you don't spend thousands of hours mastering Backgammon, you spend it mastering Poker, because that is the biggest game around. Yes it is more difficult, but I'm here for the big payoff in the end, and the payoff is biggest in the ES. It is a "Go big or go home" mentality, and I'm sure there are others that have a different view on things and want to be the small fish playing in the ponds that the big boys stay out of, but I have made good progress and I don't have any regrets so far. With that, I think I have taken your thread far enough off course. I am really surprised at how civil the discussion has stayed in here. Usually in a thread where anything bad is said about scaling out a bunch of people come in and defend it.
Yeah, I have to admit I was concerned that perhaps I've been a bit too candid (the views/posts ratio of this thread is pretty high, so there are apparently lots of folks looking over our shoulders). But its good Karma to give something back from time to time. My hope is that some other traders will chime in here and there with some tidbits of wisdom from which we all can benefit. We're all competitors, but not necessarily direct ones. I agree - if you're going to do it, you may as well do it big. I have no doubt I'll be in ES before the year is out, but I'm just more familiar with NQ right now, so its a comfort zone thing. Yeah, the signal/noise ratio of this thread has remained unusually high so far. Given that this is ET, I figured the first response would have been something along the lines of "Thanks for the test results, Captain Obvious", but its been pretty civil so far. Anyway, thanks for all the sage comments - its been a pleasure dialogging with you.