I see what you're doing now. Since you're backtesting gaps, you'll have an opportunity to see a large sample of how these set up.
You can make five minutes a rule for yourself, but in a setup such as that which occurred this morning, it's not a hard-and-fast requirement. Since the gap was pretty decent (1.5%), the probabilities were for closure, which made a fade the preferred tactic. As to waiting for five minutes, however, that really wasn't necessary since price made its initial effort to the upside, then began its decline rather than bob up and down within a range. Therefore, you could use two points below the opening low, which in this case was literally the opening low at 0930, and enter at 996.0 at 09:36.
Once you're in, the target becomes the gap, not the average range. In this case, Tuesday's close was 984, so when price got there, you had to be ready to exit. You don't want to be too tight since there's always the chance that price will continue its decline. And there's nearly always hesitation, and sometimes consolidation, when these targets are reached. But you don't want to be whipped out by a tick or two, either. So I used the same end-of-bar stop I use with the trend-following strategy I posted except that I used a 1m chart instead of the 3m. Other people just go ahead and take it without waiting to be stopped out, and that might be just the thing to do if a report is due out in only a few minutes. Having that money "in the bank" makes it easier to take subsequent trades if one has been having trouble pulling the trigger.
As for the upside, I entered at 1005.5 (six of one, half-dozen of the other). Target was 1011.5 (again, what's a tick?). When we reached that, I should have used the end-of bar stop according to the rules, which would have stopped me out at 1015.0 at 11:01. But due to a slight hangover, and being happy to begin the new year in the profit column, I used the 1m chart again for the end-of-bar stop instead and was stopped out at 1013.0. Which is why results often vary between the system and the actual play.
So results to the downside should have been 6.5 and to the upside 6.5 to 8.5, depending on the extent of the hangover.
Using the 1m chart for an end-of-bar stop is not a good habit. It's too easy to get whipsawed out of a nice position for no good reason. However, a gap-filling strategy has little to do with the trend-following strategy I posted since it has little to do with trend, that is unless price overshoots the target and establishes a trend. But there is, of course, no way of knowing that in advance. I need to look at all this again to see if stop placement on gaps can be made a little less fuzzy.
--Db
