travis, Your modesty from the content of your post and what you have replied to other posters have led me to draw a conclusion that you were a Tradestation-only system trader aka. newbie trader. Apologies for making you feel like I didn't give you credit for all the hard work you have done.
DO automated system farewell together? Automated systems are good for certain stocks but not good for all stocks. Just my two cents.
MAESTRO, I was reading this wikipedia entry: http://en.wikipedia.org/wiki/Algo_trading It says: ...A third of all EU and US stock trades in 2006 were driven by automatic programs, or algorithms, according to Boston-based consulting firm Aite Group LLC. By 2010, that figure will reach 50 percent, according to Aite.[2] In 2006 at the London Stock Exchange, over 40% of all orders were entered by algo traders, with 60% predicted for 2007. American markets and equity markets generally have a higher proportion of algo trades than other markets, and estimates for 2008 range as high as an 80% proportion in some markets... Are we talking about the same things? A majority of trades (or % of volume traded) is being made by automated systems? Or are they talking about something else? If it is like they say and I understand right, it means we are in big trouble, because they say that number is growing and pretty soon everyone will be automated. A "drop in the bucket"? Can you explain this please?
"Algorithmic trading" and "automated trading" is often mixed up. "Algorithmic trading" often refers to the order being executed by some algorithm while the decision process might be human. "Automated trading" often refers to the situation where the decision process is also automated. But it is a mess, and it can be difficult to deduce what exactly the two terms covers when you read about them in some context. So when it is stated that X% of some exchange order flow is "automated" a large part of that is probably due to algorithmic order execution (and not necessarily the type of automation that you try to do).
I did not vote since I did not fit into the narrow age group represented here. About three systems are required; primarily. they relate to the capital limitations of trading markets. Trading can become related to taking the market's offer. Capital is shifted from one system to another and the towards the least sensitive capital applications. You have made the most focused and majorly primary observation as yet posted in the thread. The initial function of any automated stock trading system is the front end filter required to make the automated system "fit" what it is designed to do. About four coordinated filters are required. travis, Are you up for some critiquing that will remove you from the dilemma you have created?
Not as harsh as your predecessors please. Also, what is the problem with the "age group"? "> 10 years" includes everyone else. Or did it take you less than 0?
I put 5-10 years, as realistically it was hard to fully automate in the mid to late 90s. My algos ran after the close and generated the orders for the next day. Then on the phone with the night desk at Lind Waldock to place the orders. The last 3-4 years have really been a lot easier with good APIs from a variety of brokers. I could not run the auto-trading i do now, just 5 years ago without incurring huge expenses.
Being harsh isn't constructive. I think you have done a great job explaining your career and intellectual development. It is not easy for a person to provide a commentary about the effort required such as becoming a successful trader. There is a lot of humor in how things were in the past. My first attempt at automating trading was in 1957 and I had 700 series IBM machines available. There were few instructions and data was loaded using punched cards. At that time the 024 and 026 card punches were available. The convention was to go to a broker's office and sit around and look at chalked data on a raised wall. the chalkers simply carried strips of paper tape from "tickers" and posted the current prices and about nothing else. At some point charts became available on a weekly basis and they came through the mail service. Before that charts were made by hand daily from the newspapers. Saved papers let you back plot and you had to make the blank charts by blueprinting brownlines. When indicators were invented that became the front end of processed raw daily data once charting came into being. I took my manual trading and made sheets of blocks (binary block functions) and their data interconnections. this was the equivalent of designing logic for what is now known as hardware. I did assemble hundreds of relays for various ways to satisfy my curiousity. Periodically a person would blow air through the relay banks to clean any bugs that may have been making residences. Anyway I drew up 39 sheets of logic and a dashboard with display lights as the logic for trading indicator inputs that came from processed raw daily data. There was no place to put it though. I waited 18 years before there was a place to put the logic (it had become known as software by the end of that time passing.) The advent of real time data feeds changed things, so I adapted by buying what became known as PC's that could accept raw data. I called the companies that supplied data to see if they could make the displays work properly. It became known as" "We will have that in 6 months". This is a euphanism for: "we don't know what you are talking about". I was very lucky to be able to marinate all those years. today is about year 53 for me in terms of trading successfully; I did not have any period where I did not make money from day one. travis, Your dilemma is simple. You are trying to build something from an orientation that is not rational. About all posters have your irrational orientation. As has been posted alot of trading today is done using the same foundation and orientation you have. Refer to the bot and algo percentages. As we all see these constructs, do not make any money. They almost all stem from an orientation that has little to do with what is going on in markets. The basis you have and most all others have is that you use induction to think of ways to automate trading. The TSGannGalt lists provided to you, benevolently and sort of with constructive intent, provide a good example of using stages of induction to get to some place that is called a dilemma. The alternative is to use deduction and logic. When you do, you are dealing with defining the markets in a language rather than rules. Bots and algo do not do this as may be seen by looking at their coding and processing of data only, instead. I went through a trading period when regulators were doing the same thing as algos and bots to detect those people who were operating accounts that were in violation of the law and regulations. I had mulitple accounts as was found out by the regulators. the accounts, to regulators, were exhibiting illegal activity and it seemed to apply to all stocks I traded. as you think about by first comment to you and as you think about the regulators (SEC, in this case), you have to consider two things: Why would regulators screw up and why would what I was doing apply to all things I traded? the solution to straightening out the regulators was to prove to them I could not do what they cited me for in so many different applications of continuing to violate in in matter they detected. I was simulating, to them, "insider trading". That is it appeared I had advance information not publically available on what was going to happen to stocks. I was using deduction and the language of the markets (what is known as "operating before the fact". They were finding me by induction and "after the fact". They were looking at a level of "unbelievable" profit taking over and over in many parallel accounts (operated by limited POA's made known to them by my broker as required by law). There was also another issue which my broker WAS prosecuted for. My broker was forwarding my trading activity to favored clients throughout the entire US AND adding their specious accounting of the reasons for my trading certain stocks. Bottom line, you have to discontinue the use of induction. (As did the SEC). Secondly, you have to learn a new math as do the others using induction. Read John Boyd's OODA algorithm to find out how induction is misused to create a feedback system. Also read Carnap on Logic Theory to learn how language applies to logic and that logic is based on language.