Quote:
Originally Posted by TheFrenchman  yeah I heard about that one, from what I hear it didn't do to well  but Steinitz has a new bot called "Steinitz HAS MTF" that he claims is REALLY 100% no loss "this time"  we shall see soon enough... hopefully he makes good on his claims this round. |
Same bot, new version.
This isnt new, this is the same robot he's always had. The basic principles are sound and always entice me, but in real action the huge drawdowns are hard to take.
The problem rises from the fact that this is entirely based on a smoothed moving average, on multiple time frames. As long as the upper trend holds all trades will eventualy become positive. However, when the upper trend changes direction you can quickly find yourself facing massive drawdowns.
I've tried manualy managing this robot, but its very difficult. The trade entries are so inacurate, due to the large lag in the smoothed moving averages, that it needs very large drawdowns just to run normaly. Stoping out prematurely results in unneccessary losses, but holding results in uncontrolled drawdowns.
I really like the system and indicators this EA is based on. But the system needs a bit more accuracy and discretion in choosing, entering, and managing its trades. It can be hugely profitable, to really incredible levels, when it has a good trend. On the other hand it can throw it all away, diving for pennies when the tide is changing.
Compare the actual price action to the HAS action. The HAS is a great trend spotting indicator. It could work very well as the basis for an EA, but it is an extremely laging moving average. By the time it gives a signal the move has almost always passed, and your really just catching trend continuation, generally from a very poor entry position. As long as the trend does continue its profitable, if the trend reverses the system fails (belly up) RIP.
Closing all open trades monthly is not good enough, by then it could already be thousands of pips in drawdown. Closing weekly is a much better idea, but any quick change is still going to cost you a lot.
This program does have many user adjustable variables, and there is a lot of potential for managing it better, but most of these alternatives have been tried. If they did work better, Don would have already incorporated them into the default settings.
This robot was a big disapointment. It is based on sound theory, but the implementation is a bit off. It needs a more accurate indicator for timing entries, and it needs to anticipate failure and have some loss cutting programing activated.