This is a great insight. It is true... with very few exceptions. Automation only ** increases productivity **... So one can take an ** existing profitable manual strategy **... And ratchet up volume and profits. But if you do not have a solid, profitable strategy to begin with... Automation will get you no where. "automated groups spend more time on homework than manual groups" The concept of a highly intelligent and profitable Bot is largely a myth. One can view a good ATS... As a video game MANNED by a Pro Trader. Then you get all the benefits of Human Expertise leveraged by Automation. Automation leverages Human Expertise.
I think the main reason that an otherwise profitable strategy fails once it is automated, is because of backtesting. In my experience, the 'tweaking' for a very specific set of data doesn't allow for future performance. Although the backtesting results will look fantastic, it is not a true representation of what the system will do in the future. If you can create a system that requires no (or VERY few) 'tweaks,' in the backtesting phase, you may be onto something! Also, another good test is to run it on other markets. If you design the system on the S&P but it demonstrates comparable performance in other markets, it shows universality, and suggests that the system may have value for future trading.
The topic of this thread is better to be @ Reasons for failure in mechanical trading systems@ Automation is just a way of executing your trade rules, except the case of high-frequency machine scalping, how can an ATS be profitable, if its strategy cannot prove success in manual trading ? I found most of the reasons discussed in this thread is about the strategy part of the failure of an ATS. But by definition, we may expect to see more discussion about the automation part of the failure of an ATS.
Don, With all due respect you must not know me. If you did you would know that I am the proof that your statement is untrue. You trade stocks mostly, I assume. I try and never assume anything but I will assume that. I trade futures, you know why because if you can beat the markets with a big honking mechanical stick. You will not waste your time with trading stocks. My opinion is stocks are for investment not trading. Your a broker, Im a trader. Its like a car salesman and a car buyer. Maybe your a broker because you haven't opened your mind that anyone can develope and use mechanical systems with greater success than any human mind could ever dream of. Honestly I would like to help you some way to at least accept that the other side exist and can do as well as any human can do. I respect you, how do I get you to respect me and what I am capable of? I will do anything, I will call trades to you minutes BEFORE I place them. I will set up a web site and post the trades for all the world to see in REAL TIME! What if anything can I do to prove to you? Name it! Sincerely and Respectfully, Mark Brown
There is no need to "prove" anything to me or anyone else. You may be one of the exceptions who make a killing with an automated program of some kind, and I applaud you for it. All I'm doing is expressing the facts about a 100 or so traders who attempted automation and, although working for a while, tended to end up in trouble (it happens to the best of us at times). We have traders who use the TICK3R's auto scalper program and make some money (as an adjunct to their normal trading)...but they seem to be tweaking it all the time, and even manually intervening a lot. We've had groups of traders who babysit auto programs as well (they seem to work for a while, same deal, need to be constantly updated and adapted). You need to remember that my brother and I traded long before we were "brokers" - and still do every day (we "lead by example" for our traders whenever possible). We traded on the CME floor in the 80's, and we still trade 1000's of emini's daily at times. Our CME membership is in Bob's name, he gets the ultra cheap commish, I don't so I trade smaller size of the e's. As far as making money, I do fine, and this isn't bad for my brother: http://www.stocktrading.com/trader10030.html That was 2004, and 2005 was still very good, but not that good, and 2006 was better than 2004. This is personal trading, not business profits or anything. I assume you're in Chicago or NY (?) - let's chat when I come to town. All the best, Don
I've been saying this over and over again... The failure in any business comes from incompetency of the skills required and the experience required to succeed as a business. I can say that 99% of the existing system traders don't have the required skills to trade a system... First and foremost, you can't expect to consistently profitable with just a great system just by owning a Tradestation, Wealth-Lab or other 3rd party software. It's not a matter of coming up with ideas, but the ability to test extensively. I own TS2000i and use it occassionally but I don't use it to run extensive tests. Time consuming tests are done by C# for me. 3rd party backtester has it's value but it hold a very small weight in my development routine. Second, risk management. A while back, I thought that risk management was the most important factor on system development. But currently, I'm back to my basics to developing a great system, as most significant. It's really my second time around all the studying with the question of "what is curve-fit/robust" remaining. I don't mean it in a "position sizing" sense but in all the Distibution -> VaR -> GARCH -> Robustness kinda sense. You have to know what environment you're dealing with... from that you get skeptical about the whole robustness of quantitative analysis... Went off a bit but conclusively... RISK > REWARD > EXPOSURE > MANAGEMENT Third, execution. This has to be blamed to all the book writers and all it's followers about systematic trading. This is one simple example of why they're incompetent developers. Most system traders are failing discretionary traders and they couldn't grasp the significance of trade execution. The ability and discipline to scratch a trade, taking a loss, or riding a profit is very important. Most seem to stop the system development with finding a "trend/move" or in simple terms "signals" but don't get into how the model should be executed. Of course, having a "signal" is an integral part of a system. My point is people don't step into the skills, as much as they should. Finally, the misconception of a discretionary trader vs. systematic trader. My thought is if you can't make it as a discretionary trader, you won't make it as a systematic trader. The time span of how long it takes doesn't matter. Changing ways doesn't make you better or worse. A bad chart reader stays a bad chart reader. (Nothing wrong but....) Buying a system is like perscribing to a market analysis newsletter. A non-disciplined trader stays a non-disciplined trader. Executing a trade by hand is the same as auto-executing a trade. ... Anyways... I've jumped around thoughts but even the great discretionary traders don't make money all the time. And... in terms of Bright Trading and the surviving props., there needs to be a line drawn between traders/desk and there are lines betwen models/systems. I don't see a lot of traders in prop. these days. And quants are not system developers.
Mr. Bright has it exactly right. Pro traders "partially automate" existing profitable strategies... And then constantly evolve and adapt everything to changing markets... And ongoing maintenance of Automation is expensive. ATS is manned by experienced Pros for an insurmountable reason: Leading edge AI cannot even come close to simulating the complex decision making of a Pro Trader... Therefore Pro + ATS >>>> ATS Unlike 1000s of Pros... Mark Brown has obviously never been successful trading stock... And his less than measured claims are highly suspect.
If MarkBrown is THE MARK BROWN that would become an understatement. It's not my style to reference other people's words but for ETers who don't know who he is... THE MARK BROWN's posts in other forum has been studied and respected by professional system developers. Acrary(being the ET celeb) has mentioned that Mark Brown and a few others are his goal and his indirect mentor as a system developer. One of the system that was offered to the public (not sold... but offered in Active Trader Mag.) was called Oddball which literally gave new insights to how trading systems are developed. The convention impacted the developers, in public, to the degree of Turtle trading and Toby Crabel. I'm not much of a system trader yet but I've valued his past studies and insights on developing.