I think the discussion here is on two different (related) topics. The type of trading that you, STOCKBROKER, Don Bright, etc. requires tape reading, looking at order flow, and a very short-term time span. When most people refer to "systems trading" I believe the are usually talking about a longer time frame where reading tape and viewing order flow is not required. (Although it is conceivable that a a complex system could be programmed - that is a theoretical topic for a different thread.) I've seen plenty of mechanical systems on individual stocks that have performed very well, both intraday and EOD. I prefer to focus on futures where you can get more leverage per dollar (and risk), even when compared to someone at a pro firm. Stocks are also a lot choppier intraday than futures, which lends itself well to the type of trading you do (lots of small opportunities). The risk of a systems trader is that the system may not work in the future, whereas if you're an expert tape reader and can make profits on lots of small trades then you would (in theory) be able to do this day-in-and-day-out for the rest of your life. However, you're not going to get $1B under management tape reading.
Hitman: I thought I began the conversation with the admission that I was focussing on day trading e-mini futures (Nasdaq specifically). Your comments on individual stock day trading is probably correct, although I would imagine that many high volatility stocks (>10% swing per day with > 1 million shares and > 1 bill cap) probably have similar intra-day patterns which common rules could be applied to. In reading so many messages on this site, it just blows me away how many people seem to trade on their "gut" feel. For some, perhaps it works, but for me I need a more objective approach with less subjectivity or witchcraft. It's just the type of person I am. Very mechanical, yet adaptive, rules based system that attempts to capture what people are interpreting in their "gut" feels. I would surmise that many gut feel trades can be put into objective measurements. There is something going on that is prompting you to enter those trades and exit those trades. You have generated your own internal representations and rules that you use, whether you know it or not. It's capturing those rules that is difficult. As for technical analysis, I would think that day trading futures contracts would have to lean significantly on the technicals, which day trading a bunch of different stocks which you each day would probably not require as much Technical Analysis. I'm rambling, but hopefully you all understand my thoughts. Have a nice holiday everyone - gotta go cook the bird now... TAHOE
I found a Neural Network trading system from Ward systems. It's 3K but comes with a full money back guarantee. Sounds legit. Anybody use these things in systems trading? Any of you programmers try them?
Intelligent machine trading is the future. Those who cannot see this didn't see the explosive rise of the internet or the significance of the drive-up window at McDs :~D "I can guarantee you that tape reading is the most important element on NYSE trading and I don't believe anyone on this board can program that." ~Hitman Hitman, you aren't interpreting the tape through divine inspiration. I assume by your comments that you cannot define your tape reading methodology, unfortunate for you, if you could you would be much further along. Even though you are not consciously aware of what it is that you do, your mind is subconsciously working/creating if/then algorithms. You are continuously making subconscious choices and acting on them. Would you really argue that these meaningful and fiscally significant processes are better left to flounder vague & illdefined in your 'reptilian mind'? Or, as I posit, wouldn't fetching these vague concepts to our frontal lobes and illuminating them the light of reason bear far more fruit? Can you name one thing that is not bettered through improved definition? I doubt you can. Yes, with better understanding comes improved utility. If you cannot make explicit what you do implicitly, then you do not have a true and sufficient conception of what it is that you are doing or attempting to accomplish. If, right now, you cannot write down how you make your tape reading choices, the variables and how you weight them etc, then you do not really know what you are doing, do you? And how can one improve efficiently what he does not understand to begin with? The answer of course is, you cannot. Finally, when one does possess genuine knowledge, when these rules become explicit, it then becomes an easy task to transfer your trading acumen to another person....or MACHINE. In time, these rules will become explicit, and, if there is any merit to their predictive value, smart machines will be mining the financial markets and your slow plodding methods will be an anachronism. There are too many here that place too much stock in the 'holistic' approach to the markets (and I suspect in their lives too), and do not truly comprehend the power of the mechanistic approach. Intuition is HIGHLY overrated, in all things, especially trading. Inutition is equivalent to not really knowing what it is that you 'know' but placeing faith in some deep inner 'voice' of unknown origin and of unknown veracity. I have to laugh at every post that invokes intuition in trading as something to be revered and achieved. Such a goal is second rate at best. The rational trader eschews such nonsense for what it is, and strives for increased clarity, not more uncertainess. FIREHAWK "Intuition is not a prescription for thinking, but the very opposite, a declaration that certain things should not be thought about further."
Are you talking about just an NN simulator (e.g. Neuroshell Trader) or are they selling you the actual model with all the data inputs, data preprocessing rules, NN structure, etc.? If you're just talking about the software then you still have to create the model yourself. It's not as easy as it sounds. Most users build a model, test it with out-of-sample data, looks promising, then trade it and the model stops working (due to over-optimized nature of most NN models). Even when you adhere to proper validation methods in testing out-of-sample data, the problem is you don't know if the nature of the markets as captured in your model has changed or not. One approach is to use adaptive parameters in your network so that it at least has a chance of surviving in different market conditions. If you train your model on 1990 data, test it on 1970 and 1980 markets to see if it works or not. NN's in general definitely do work and work very well for numerous real-world sytems. Whether they work for trading has been debated for many years.
Firehawk, I agree with some of what you have said. But the main reason I am writing this is to express the extent to which I am enamored by your literary artistry. Twas indeed sheer bliss to cast my eyes upon your artistic prose and eloquent expression. Happy Thanksgiving, Candle
candle, thanks, but reading many of your posts, i must say your prose is far more eloquent than mine. My sentences are wooden & my syntax clumsy next to your great literature. But thank you all the same....... Happy Holidays Your friend, FIREHAWK
Dearest Firehawk, I express sincere gratitude for your compliments. It would not be an understatement to affirm that I am inordinately flattered by your good words. But I must disagree with some of what you have said. I cannot concur with your dual assertion that your sentences are woody and that your syntax is clumsy. On the contrary, my friend, your sentences are a marvel of positive anfractuousity and your syntax is the antithesis of clumsiness. It makes for consummate delight to cast my organs of vision upon your majestic discourse. In deference to a litterateur of the highest order, Candle
Dottom: The vast majority of the people with 1B hedge funds are probably still discretionary traders. People who can trade successfully with fully mechanical systems are exceptions in this business, not the rule. I don't care if you trade 100 shares or 100000 shares, if you intraday trade NYSE, you have to be a good tape reader. "Seen" a few systems in action is non-sense, you either put yourself at the front line every day or you don't. Firehawk: There are many things in life that can not be defined using scientifical expressions. Love is one of them, no one can tell you that you are in love, you just know it, a lot of the times you can't even put it to words that why you love someone as much as you do, sure you can list a few reasons, but it is far more than the sum of its parts, and it boils down to that "feeling". There is no "if-then" statements, even if another person comes loaded with similar (or even superior) "features", doesn't mean it will produce the same impact. There are indeed things that are impossible to quantify or define, and that's what makes this world a little bit more dramatic than it would have been. I have said it before, and I will say it again, by the time the machines replace most of the traders, 99% of the existing occupations on this planet would be long obsolete. Let's not forget, if your system is simple, then you probably can't trade it mechanically as a discretion is needed to filter out the noises. If your system is complex, then the changing market environment will kill it at some point for sure, and to be able to maintain the edge, you have to be a great trader to begin with, and why would you bother with a system when trading is at its best when it is free flowing? No offense, more than half of the traders (DT-waw for example) on this board touting mechanical trading systems are not even trading real money. What they don't understand is that it takes a great trader to build a great system, if you don't have the deepest understanding of the market, then it can't be done. By the time you acquire that understanding, you probably will not even bother with mechanical trading because you will find that trading is more than the sum of its parts. Last but not least, if you are talking about mechanical trading intraday stocks, then backtesting is total garbage with the slippage, and let's not even get started with the short side, when your beloved system generates the signal chances are you can't short it because the thing is downticking. Until the number of mechanical day traders (stock, I don't trade futures hence I would rather not comment on them) grow large enough to be even 10% of the profitable traders population, I say it is impossible.