I have to side with traderRX on this. I think most people envision a 'holy grail' as a fixed set of parameters for some TA indicators. Actually it can be anything you define as your own rules.. also it can be flexible and self-monitoring, self-learning. It can be programmed like a good trader, with iron clad discipline. In terms of how many people are successful system traders, nobody knows. The failed ones shout to the world 'There is no holy grail', while the successful ones hide as well as they can and collect dough. One hint.. IB has a number of institution and retail accounts that use there FIX interface for order entry. Those people must be doing automatic trading and some of them must be successful. I have been doing some research on emini futures with some rules I created from scratch, and in back test, it shows some consistency with acceptable drawdown, with over 3 years of tick data. The systems will probably put to live trading in the near future. It is probably a bit tougher to work on a system for stocks, as you have execution issues. But at least a 'trading-assistant' is possible, which can monitor multiple stocks and alert you for set-ups.
>"It seems that traders want to differentiate between "system" and "discretionary" trading. In reality, the differences are more a matter of degree. The discretionary trader uses a set of rules that takes into account his experiences in the market. Often these rules can be difficult to express in a way a computer can use this knowledge. Simply because it is difficult for systems to mimic the nuances of human thought is no reason to dismiss system development as ineffective. To do so is not only shortsighted but patently incorrect. In fact, in certain circumstances, removing the human element is advantageous." I'm becoming increasingly convinced that what is commonly relied upon as "experience", whether in trading or any other discipline, is really a subconscious simple algorithm that operates just beneath our level of awareness; and if we exposed to the light of day this "experience" or "market wisdom" that we cherish and rely upon to make the correct choices, we would be frightened at how tenuous the thread of reason really is that forms the basis for our decisions to buy this and sell that. And if I'm correct, a computer can do it better.
What Hitman is trying to tell you, I think, is that ultimately your supercomputer will be controlled by you...a person. A person that is exposed to all manner of extraneous data and theory. Do you have the discipline to let your "whatif" supercomputer do its' work and not let your emotions override the signals? Do you have the discipline to stick with the system after several consecutive losses? The computer is not trading with its' money. The computer, in effect, is paper trading.
traderx - couple of thoughts: 1. Yes, large continuous number crunching systems can be used to take advantage of short term arbitrage opportunities for very large profits. It's done all the time by large trading houses, some of which do little else and pocket millions each year doing it. The problem is that you basically have to be a large trading house to take advantage of the situations because in order to benefit you have to very quickly execute program trading transactions that you just wouldn't be able to do as an individual trader. For instance, when the premium between the S&P futures and the S&P cash index narrows or widens too much, programs will kick in that arb the two markets. 2. The ongoing difference of opinion on this thread seems to be one of whether any form of technical analysis can "predict" prices (apparently with the underlying presumption of doing so with 100% accuracy). For my two cents, that's the wrong issue and certainly the wrong perspective. TA isn't so much about predicting prices as it is about analyzing market psychology and predicting crowd actions. It's certainly not 100% accurate, but what is except hindsight? Nevertheless, traders at large trading firms use TA as the basis for their trading every day of the week. Many of these firms spend millions of dollars a year to provide their traders with as much of technical/computational edge as possible and do in fact put elaborate computing systems at their disposal. One large private (and very profitable) trading firm I did some work at years ago basically used human traders has executors (since the futures markets they traded were all open outcry at the time). The computer flashed buy and sell signals to the pits and the humans did what the computers told them to do. They had a backroom full of PhD quants making big bucks, they paid the floor traders based on their ability to execute as close to the computer's optimal entry/exit signal price as possible, and the partners in the firm were each raking in 8 figures each year. Since no TA is always accurate, all traders (even the completely electronic kind) naturally use position/money management techniques to limit their exposure and milk the positions for what they can. So is TA predictive of prices? That's not really the issue. When coupled with proper position management, it's sufficiently predictive of crowd actions to be profitable and is therefore valuable. That being said, there is a caveat. The big issue in the effectiveness of any TA approach is how closely the resulting analysis is to what the crowd is looking at. For example, if you just pull some abstract 141 day moving average out of the air and then start predicting what the crowd is likely to do when the price "breaks out" above that average - chances are you're going to be hopelessly disappointed. The flaw isn't in the technical analysis concept, anymore than a poorly framed house is the fault of the hammer. The obvious problem in this example is that the majority of the crowd isn't looking at your 141 day MA. On the other hand, if you're using the 50 day EMA you'd have a much better likelihood of predicting the crowd's actions (assuming that you've factored in any other resistance factors into your analysis - since the collection of support/resistance elements is needed to draw a reasonable picture of the price action, not just a single study or indicator).
A computer that simulated the universe with perfect accuracy would have to be at least as complex as the universe. Since the market is a reflection of the crazier people in one small backwater of the galaxy, you probably would need to simulate the whole universe to get something that truly took every variable into account.
This "discretionary" trader VS "system" trader business makes zero sense to me, and never has. Hitman is one of the champions of the "discretionary always wins" credo, but Hitman, in my experience, rarely analyzes what he means when he utters credos. Hitman's definition of "system" trader: regardless of market conditions, I repeatedly use the same exact rules over and over again to buy and sell. I never deviate from this. I've said this to him a million times (no effect), but: Sorry, but you'd have to be a moron to trade this way. I have a specific 'system' that I like to use. It works great in certain market conditions. When those conditions are present, I use that system religiously, and never deviate from the rules. I make lots of money. When the general market conditions favoring this system are NOT present, I do not trade this system. I either use another one better suited to the present conditions, or I don't trade. Because I follow a strict set of rules when I trade -- each time falling under some system or other, I consider this a "system" based trading routine. Maybe this should actually make me a discretionary trader, because I don't use a particular system ALL THE TIME. I'm not sure. What IS a discretionary trader? Is this a trader that follows no system-like rules? A trader that willy-nilly picks XYZ long for no apparent reason? LOL. Or is a discretionary trader someone that moves from system to system based on the trader's perception of general market conditions? If that's what it means, then fine, I'm a discretionary trader. But if that's what it means, then "system trader" has no meaning (I don't know anyone that would use the same system all the time regardless of market conditions), and this whole "Discretionary vs system" distinction is meaningless. Wet
Wet: If you have a fixed criterias for your system (the market index moving average crossovers or tea leaves reading doesn't matter), then of course it is a system. You do not trade a system until the conditions are met, as long as the conditions are pre-defined then yeah it is a mechanical trading system. A discretionary trader makes many many more decisions, entry, stops, exit, pare in, pare out, while a system trader has a pre-defined set of criterias and once he is in a trade, it is basically on auto-pilot. Your system can be as complex as it has to be, but you are NOT involved in making those decisions, your system generates all the signals, does all the work. That is my definition of a system trader. I stand by, and insist, that at equivalent tiers (skill/capital/ experience), a discretionary trader beats a system trader hands down, simply because the market is a dynamic animal. Being discretionary also allows you to trade VERY aggressively and you typically hit more shots, to be top tier in this business, you have to be able to trade VERY aggressively (this is not to say trading foolishly), and a one of the most important trait of a trader is the ability to tune up his aggression as needed, so when the market is slow, he trades conservatively but as soon as he sees an opportunity, he can go after it with heavy size. By the time a computer can trade as well as a senior trader at my firm, most other occupations in this world will probably be long obsolete, and I am dead serious about that. I think a little clarification is needed. You swing trade, I said one can swing trade with a system profitably (because the fills doesn't matter as much, and backtesting actually mean something), but intraday trading is a completely different story. Not to mention ultimately to define success one has to be trading for a living, I wouldn't associate "make a lot of money" with either one of us yet, so I am not ready to brag about anything just yet . . .
Hitman, When I say "made a lot of money" I mean percentage wise. I've made a great percentage using my system. That's a fact. I'm not "bragging", simply pointing out that system trading is not as "doomed" as you always seem to suggest... You said: ================== A discretionary trader makes many many more decisions, entry, stops, exit, pare in, pare out, while a system trader has a pre-defined set of criterias and once he is in a trade, it is basically on auto-pilot. Your system can be as complex as it has to be, but you are NOT involved in making those decisions, your system generates all the signals, does all the work. =================== So a discretionary trader has no rationale behind what they do? They willy-nilly pare in and out of positions? They flip a coin, I guess? That sounds absurd and probably totally untrue. My point is that if you pare in/out whatever, you do it only in certain situations. You don't flip a coin or roll a pair of dice to decide whether or not to do it. If you don't think X is one of those situations, you don't do it. If so, then you apply rules to determine if X is or isn't such a situation. If you use rules, then why isn't a 'discretionary' trade a system trader after all? It makes no difference if you use many systems a day. You use rules. You said: ====== That is my definition of a system trader. I stand by, and insist, that at equivalent tiers (skill/capital/ experience), a discretionary trader beats a system trader hands down, simply because the market is a dynamic animal. ====== Again, you make the silly assumption that a system trader does the same thing, regardless of market conditions. How many times do I have to say it? A system trader need not do the same thing in every market condition. In fact, you'd be stupid to do so. So, if "discretionary beats system" means "a person that uses many systems to match current market conditions beats a person that use one system all the time", well, then NO kidding. But I don't think there are many people that use one system all the time. So you've constructed a straw man argument. You said ================ Being discretionary also allows you to trade VERY aggressively and you typically hit more shots, to be top tier in this business, you have to be able to trade VERY aggressively (this is not to say trading foolishly), and a one of the most important trait of a trader is the ability to tune up his aggression as needed, so when the market is slow, he trades conservatively but as soon as he sees an opportunity, he can go after it with heavy size. ================ So, in other words, use different systems depending on the current market conditions. A system trader, for the 100th time, does that if he/she is smart. You said: =========== By the time a computer can trade as well as a senior trader at my firm, most other occupations in this world will probably be long obsolete, and I am dead serious about that. =========== Probably, but system trading isn't computer generated trading. Different subject. You said: ================ I think a little clarification is needed. You swing trade, I said one can swing trade with a system profitably (because the fills doesn't matter as much, and backtesting actually mean something), but intraday trading is a completely different story. ================= Maybe. But in your first post you said "There are entire teams of developers very well funded doing this as a full time job, but the reality is the market is dynamic and that's why you will see very few succesful system traders and even fewer (not sure if there are any) successful system day traders." Sounds like you are implying swing and intraday. You said: ============= Not to mention ultimately to define success one has to be trading for a living, I wouldn't associate "make a lot of money" with either one of us yet, so I am not ready to brag about anything just yet . . . ============= That's a lot of self serving crap. Success trading does not entail "trading for a living". I know you'd like for it to mean that (since it's your job). And for someone that does it for a living, sure, if you can't succeed at it then you're out of a job. But sorry, "success trading" doesn't mean "trading for a living" for everyone. For the vast majority of traders, it means "making money". Period. Nice try, though. I guess successful investing means working for an investment house too. All those people with multi million dollar retirement accounts -- that never invested as a full time job -- must be pretty unsuccessful in your eyes. ROFLMAO Wet