Quoting volanDislikedIso, I think you may be implementing the 50 pip rule in an opposite manner to which post #13 prescribes. The following figure shows my understanding of post #13. The lines that say 50pips are just approximate because I eyeballed them.
http://i28.imagethrust.com/t/113534/eurgbp.jpg
The two horizontal red lines were generated using Iso's "tun_lines" indicator. The turquoise horizontal lines show (about) 50 pips above the turquoise 8 SMA. The grey horizontal lines show (about) 50 pips below the grey 55 SMA. Even though the two regions overlap a little, it can plainly be seen that the 8 SMA is more than 50pips away from the 55 SMA. But the "tun_lines" indicator shows that all indicators (including price) are within its bounds.
Dial, is my understanding in line with your post#13?
Iso, is my understanding the same as yours?Ignored
It does not matter which ma is higher or which is lower. The 8 can be above the 55 or below it. It only matters that when the two ma's are within 50 pips of each other, the upper boundary line will be 50 pips above the LOWER ma, and the lower boundary line will be 50 pips below the UPPER ma. Think of it this way. If the 8 and 55 are exactly 50 pips apart, and the 55 is uppermost, then the upper boundary will be on the 55 and the lower boundary will be on the 8. As the 8 and 55 move closer together, the boundary lines will move further apart until both ma's are at the same level (crossing). At that point the two boundary lines will be at their maximum distance of 100 pips apart.
Iso's reasons may be different from mine, but I'm considering ditching, or at least modifying the 50 pip rule. In my backtesting, I have seen slope changes that would have been ignored because the 50 pip rule was in effect that ended up garnering well over 200 pips. I have also seen instances when the 50 pip rule was not in effect, but the 8 sma and price were in bed with each other for an extended period. Now nothing works all the time, so the best we can do is come up with rules that work most of the time in order to stack the probabilities in our favor. Even though I am considering ditching the 50 pip rule, I would not do so just because of the occasional occurence of the two scenarios just outlined. Ignoring the current 50 pip rule just to capture the above two scenarios might also let it a lot of losing trades. It may be that the current rule, warts and all, may be the best current solution. I just suspect that it can be improved. And that will require additional testing. As always, rely on your intra-aural computer for case by case discretion.
Dial
sxaxlxvxaxtxixoxnxbxyxgxrxaxcxexdxoxtxoxrxgx