Ok guys thanks for the interest. I'm not sure what that bicycle comment was about... To answer your questions: My setups are for the open only and I only take a trade if my preconditions are set...again not very complicated there's only two conditions. However, this means I average about 15 trades per month. The backtesting used probably between 100-200 instances of trades taken. Cross validation used a sample of about 50-70 trades. I look at 30 minute candles. No indicators or anything else. I also don't look at other markets or anything. One screen with the 30 minute chart. Period. And BSAM trust me they are VERY simple and straightforward.
now if you go and backtest on these 4 months what do you see? is it also 30% of expected performance of prior backtests?
Let's get serious, the orginal question is can a 100% system (software) approach work ?. The answer is yes. The second question is can a discretionary trader , who really uses a system which is not 100% objective do better than a system trader ?. First a 100% system (software) approach can work if the system are well designed and you trade multiple uncorrelated systems so the failure of one does not cause large losses. The question about discretionary trader is that a great discretionary trader can beat a system trader, but great discretionary traders are rare. I would say you need to be in the top 5% or less to compete or beat a well designed mechanical program. On the other hand if you have a well designed mechanical program, all you need to do is follow the rules.
Murray, I hope you design systems better than you can read. Those were your questions, not his. His issue is that he thinks he has followed the classical system development rules, is following the system live, but is not getting the same results. So we are asking him those questions which will lead us to suggest what might be wrong.
Is that not the hardest part for a mechanical trader? Not second guessing the signals, be a robot no matter what happens? I find in the trading I do that computers and mechanical signals are a tool to help you monitor the trades that come up, but that using your human mind and insight set the results we can create as a trader...
The order type doesn't matter (limit/market). What OP has not accounted for is that market conditions are in a constant state of flux, and a 'system' that worked perfectly when backtested for even six years can stop working or need to be adjusted as market parameters change. Example: I use a system with a standard stop, and trade at different hours of the day. During different hours I cannot use the same stops and limits because market characteristics change (factors like volatility and volume) and the "robustness" of said system changes. Additionally, even trading at the <i>same time on dfferent days</i> requires adjutsments. I do not think it is possible to have a truly profitable system if it is a black box "plug and play". You have to be able to adjust for market conditions, even if it is only a little. A stop of six ticks vs. eight ticks can be the difference between a win and a loss. That said, one of the sponsors is constantly crowing about how wonderful his systems are and his "students" are making wild amounts of money because of him. I am skeptical (to put it mildly), but if he is to believed, he claims that what you are trying to accomplish is very simple.
"The order type doesn't matter (limit/market)." HAHAHAHAHAHA! And to be precise, if you have multiple stops and multiple profit targets depending on conditions, then you have MULTIPLE systems, not A system. Related? Yes.