I had never really looked at these values until now. What do you usually do with these values when you're designing a strategy? I know there's a book called MAE but I don't have it (yet) and don't know if it's worth buying. I've heard one tip. Finding the max adverse excursion of your profitable trades and use that as a stop. Someone said it could improve your profits by 25%. It actually made mine worse. Attached is a graph of Profit vs MFE of one of my systems. I haven't designed a stop/exit plan for this system (intraday). The results are from closing the position at 3:55 everyday. Looking at that graph, do any of you guys have an idea of possible exit techniques so I can capture the most profits as well as limiting the losers? Any stop/exit plan I try to design for it cause the profit factor/percent per trade/win ratio to go down. Ideally I'd like to get out of the market sooner but without sacrificing potential gains.
It looks like all but 4 of the trades that go your way more than 2.0 (MFE > 2.0) are profitable. 3 were losses and 1 was breakeven. The rest of the losses had MFE < 2.0. The trouble is, we don't know the MAE so we can't say if there is a stop that will cut off more losers without trimming your winners too. Why don't you look at all those trades where MFE >2.0 and see if they have any characteristics you can isolate. Tim
Yes you're right. I was going to include MAE but wanted to leave work early. Here it is attached along side with MFE vs Profits. Looks like I can maybe set a trailing stop at step intervals of 2. So if > 2, stop at 0, if > 4, stop at 2, etc. I'll see how that works out. Thanks for the input.
MFE and MAE, like any other performance measure can lead to curve-fitting the equity curve, just like curve-fitting the market.
I completely agree. But this system looks only at the current days data and makes a decision. I wouldn't think that a position taken at 10:30 would have an ideal exit at 15:55. Know what I mean? The system isn't fitted because the entry can't really be optimized. There aren't any variables to change. It just is what it is. So I figure I should give a shot at somewhat optimizing the exit. I know all too well about over-optimizing. I wasted too much time and money in my early attempts at creating systems.
You may try to look at volatility normalized MAE and MFE. If you exit at EOD, you're trying to catch fat-tail moves. Look to place initial stop narrow enough to cut off negative outliers, but wide enough to leave positive outliers intact. Again, your stops should be expressed in volatility terms, like 0,5*average_range(20), not like 2% or 15 points or any other fixed value.
That was a nice chart Tums. It looked like the MAE graph showed a value between 3.25 and 3.5 might cut off more losses than profits. That might be clear if you binned the values for net profit in vertical bands. Tim
Thanks. I'll check that out. By looking at my MAE and MFE charts, I decided to test entering into positions at +2% and -2% for the data set above (2007, 100 of the highest volume stocks). Profit factor for positions taken at -2% was 8.2, and for +2% was 3.3 (original at 0% is 4.9). About the same type of increase/decrease for other years tested as well. Looks like I'll try some position sizing strategies so I can scale in. I don't know what the overall profit factor is because my software isn't yet set up to average up/down like that. If it doesn't increase I suppose it's not worth averaging. Will let you know when I'm not too burnt out to do some programming.