Quote:
Originally Posted by Ken Long  Interesting. If you look at trades # 10 and 11, order #'s 19 and 20 from year 2006. They had just passed a major loss and doubled up on the position size as per the RecoveryMode. Trade #10 was for 8.39 lots, followed by trade #11 for 4.40 lots. Both of these trades were open simultaneously for over an hour, while the account balance was listed as $12083.40... Obviouly they are running at more than 100:1 leverage or this would not even be possible, even so, this is an extremely high risk situation. And I'm sure there are more like it to find.
I have to wonder if these posted results arent doctored a bit to prevent blowing the account up. As was previously mentioned by BB, when they had a longer string of losses they didnt run the recovery mode. Easy to do with hindsight. Not so easy going forward. These user parameters, RecoveryMode and ReduceLots false, will take some thought before implementing. |
Good points. When I first started backtesting and building trading systems (in stocks initially), I learned a few things the hard way. ALWAYS expect real-time drawdowns to be worse than backtested ones, and returns to be lower. Sometimes there are exceptions, but prepare for the worst. Even if you don't intentionally "cheat" on your backtests (and many EAs do), there are many subtle ways to artificially enhance your results in hindsight. If you're trading at risk levels that barely survived backtesting, you're in great danger going forward.