Granted #2 is something you can't do anything about, but #1 and #3 you could have been prepared for. Catching bugs in your platform should be done while on simulation, otherwise use a more mature commercial platform for automation. #3, well, all I can say about that is "UPS". RoughTrader
Well, there is obviously some disconnect here. From the charts that Frost posted in the past, it looks like his "trend-fighting" system made a net gain of about 125 ER2 contract points over the 3-month paper trading test period, from which it is fair to conclude that trend fighting does work (or did work for that test period). The other possibility, of course, is that it works for trends of certain durations, and doesn't work for trends of different durations (either shorter or longer). Yet another possibility is that Frost might have missed something in his backtest, so it showed spurious results. What I think is the most likely scenario is that the system was in fact profitable over that 3-month period of time, but the test period itself was too short and too abnormal in terms of ER2 action to be meaningful. It was just like 3 heads flipped in a raw: fairly easy to observe in a long series of throws, but worthless to bet on as a pattern. I'd also say that the results of the live run are totally inconclusive, as well. Over the last one month of live trading, the system lost about 20 ER2 points, which is less than 3% of the contract value, while the ER2 contract itself ended about where it started one month ago. So, the only conclusion that can be made about Frost's retreat from the system is that he is undercapitalized. The system itself may be just fine.
I don't assume anything...and the facts remain...that was a nice "trend"....albeit short-lived. Even if the bot got 50% of that move, it would be noteworthy.
Crazy question - if you see things that a computer program cannot, why not trade manually then? Why does there need to be a computer program involved here? While human knowledge and emotions can get in the way of trading, once you hone this skill your trading may be far superior to a computer program.
Well, after all, this is the AUTOMATED trading forum. But you are right. Despite IBM's Big Blue beating a human at chess, they've not been able to do the same with trading. It's been said it'll be another 10 years before the technology to put human intuition and cognitive knowledge into a piece of code is mature. Also, Ricky Cheung, the famous trader who lost a fortune, then built a few strategies, and came back a winner said the same. He never relied totally on his "systems".
Hmmm, let's see, if you really could get a computer program to trade exactly as you would then you could go off and do something else enjoyable while your computer is busy making money for you. Seems like a worthwhile goal to strive for. Also you could add additional strategies that trade concurrently, something that a human may not be able to do easily if at all. Then again, some people still think paper and pencil is better then word processor and spreadsheet apps. To each his own I guess.
there is a reason 90% of volume on the london sx comes from automated systems.. i think they have already been quite successful replacing the human trader.
Speaking of the data outage frost experienced the other day, has anyone else noticed that the IB quote for ER2 is stale? I'm seeing the last price stuck at 792.5.... RoughTrader