You are right, my system is not that great, although it doesn't seem to be terrible. To be honest, it's not even very well defined. I'm trying to define some rules, but what I always see is that there is something I would lose, sometimes because of pure randomness (I guess), sometimes because of something I didn't get, usually something in the news. Being a software developer I would love a system that can be automated, but looking at the complexity of the market, and mainly the importance of the news, a completely mechanic system doesn't look very doable; if it does, I would guess it's a system with a poor performance, compared to a more discretionary system. Or maybe I just want too much. That's why I was curious about the way people here do their trading
it looks more like you are tallking about the course of a variable, rather than its value. I guess you can still do that automatically, to some extent. You can program for example rules based on average, max, min, variance, slope, frequency.
My systems are tightly defined. I always follow the rules with the exception that I take no trades on FOMC days. Learned not to trade those days the hard way. I'd like to completely automate my systems but the real world doesn't seem to be too accommodating. For example, if I want to exit at end of day, I put in a MOC order in the system. In backtesting it always gets the closing price for the day. In the real world it's different. Tradestation 8.1 doesn't support MOC and OpenECry places the order 1 min. prior to the close. In the ES market this means you're not going to be close to the closing print. To make it work I manually enter a market order with about 5 sec. till the close. This means my system and the brokers system must both have their time synched correctly. One broker I used let their clock drift and about 1/2 the time my order would be rejected because they thought the market was closed. (dumped that broker). Anyway, I still only get the closing print about 1/2 the time. The rest of the time I get the spread (which is ok since I use the spread for slippage in backtesting). There are lots of other examples like this that make autotrading tough for me to see being a reality anytime soon. Manual trading is good enough for me.
In trading Grob still does not know the parts he does not know. At his age he is well past the option of getting 'a wake up call'. Parts of his procedures are Stone Age. I won't be enlightening him although undoubtedly his teachings have benefit for independent traders.
*** Just try using very public information to determine how you can get a system which has more wins than losses, with the wins being larger, for starters. Your system should include but not be limited to the use of: Pivot Points Moving Averages using $TRIN If you google any of the above categories you will find a plethora of public information for free (and some at small cost) and/or go to Amazon and plug-in the categories and see what you get. Pull up a chart using the 3min or 5 min criteria, plot the Pivots, incorporate the Moving Average(s) (length and type) of your choice, incorporate volume measurements if you like, for the past three months, and look at what the information is telling you. Code it and give it a shot (you're actually a lot closer to being profitable using a mechanical system than you think, as you already have all of the programming knowledge, you just need the trading experience/skill). With effective money management you should do pretty well. Best, Jimmy P.S. There are also plenty of books on coding trading programs at Amazon also, you just have to burn the midnight oil while you're putting together your system. P.P.S. Thanks to you and Grob109, (from a previous thread), I'm rewriting my manual to include coding the setups as if they were for a computer program (I already do VBA and Excel formulas) and I've incorporated graphics (color pictures of the examples) as well.
... and one more thing based on your comments about programming andread, and Grob109's reflections, I put together a Trading Grid which instantly made the criteria for my setups "crystal clear" (when you have 4 or more setups which require different parts of 9 different parameters (internals, different time frames, multiple indicators to confirm/deny) it gets pretty complicated ... until you know how to put it together! I'd have to say "thank you" for that one. Later, Jimmy Jam
trading is complicated, programming isn't It can be because of what my job is, but I think that every system with strict rules (and I mean really strict, so that everything is a number) can be automated, although sometimes it can be a bit harder because of technical limitations, like in 40yotrader's case. The "grid" in Grob's posting was not completely clear to me, but I believe it's no exception, if it's specified a bit more