Scaling-in/Adding near your average MAE; why not?

The answer in this case is that for most strategies the best trades do not take much as much heat. It is very likely that if you look were to look at your best winning trades that they would take the LEAST MAE. The trades that take more MAE are more likely to end up as a loss. Adding trades that are losing with a small stop is more likely to produce a type of skew where most trades you add too will lose. Adding to trades with a large stop will produce a skew your most trades you add to will result in profit but you will take bigger hits.

More precisely, if you can determine where an optimal re-entry would be, you would most likely need to reduce your original size in half to keep your risk the same. In general, adding to losing trades with futures compounds your risk. This is not the case with binary options where you merely multiply your risk.

Let me add that most mean reversion systems will show an ever decreasing average loss as the MAE increases. This is because such systems that are net profitable will never basically "max out or die" in the backtest simulation and because many markets do tend to be mean reverting. But, the tendency for the MAE to be much higher then the closed loss for mean reversion systems is also why it is difficult to threshold them with a stop loss around the average loss because most of the trades that have a small MAE will go on to produce a larger profit.

The stop loss point that proves where you are wrong is just one way to view stop losses. There are many other rational ways to view stop losses. For example, your stop loss could simply be a risk limit where it says you don't want to risk any more then a given amount on a particular trade. The stop loss could also be a global optimal which was simply the optimal found during the backtest. The global optimal may not be the local optimal though, i.e. the best value to use in your trading for your account size and psychology. As long as the stop loss doesn't stop out the majority of good trades, it is basically doing its job of limiting your individual trade risk.

Gary 

Hey all, Riddle me this.... Let's say my statistics tell me that on my winning trades I have an average MAE of 8 ticks. For this example, let's assume that I am using a 12 tick initial stop. Why would I not want to add to my position if/when price is trading against my initial entry around the 8-12 tick area with a smaller stop? Example: I enter using a buy stop market order on CL @ 75.11, my initial stop is $74.99. Shortly after entry, price is trading around $75.03. Why not add more here with a 4 tick stop? Statistics tell me that this is as far against me as it will go, on average. I add to my position with relatively little more risk, but the upside is potentially much larger. If you think your initial stop is no good any longer, then why are you still in the trade? Why wait for it to hit your stop? Your stop should be placed in such a way that you know you are wrong if your stop is hit. On the other hand, if you know your stop is good based on your experience and statistics, why not add more with a small stop (still honoring your initial stop) thus allowing a slight bit more risk, but maximizing your upside much more? 2ND part to my riddle: Best way to handle this logistically with Ninja? I currently do the following: Enter trade with buy stop market order with "ATM Strategy 1", change to "ATM Strategy 2" (add with 4 tick stop) and place a limit order 8 ticks from initial entry, and then manage the exits as I normally would; possibly even look to pull off some of the trade near my original entry. Exits will and should be based on statistics and price action at the time the order is placed. Ideally, price trades against me enough to fill my additional order, and then moves on like I expected it to move in the first place. Something to think about...... Thoughts / comments / suggestions? Thanks, Gary