Disliked{quote} Steve, While these entries and exits look really promising I doubt they have any value when the bands are repainting?Ignored
Regarding the Entries -- subtle repainting may be okay. Remember -- repainting is an artifact of charting that occurs in hindsight. When you trade you only care about the current bar, and it doesn't make any difference what the useless indicator looks like to the left of the current bar. I mean, I could remove all markings to the left of the current bar and SpikeHunter would still work fine as a trading indicator. All we care about is the computation at the current bar -- everything else can be ignored.
I don't have any evidence yet because I've been busy coding and not testing, but I think there is a good chance that @Swingman's LazyTMA (or TMA centered) can still give good computation at the current bar. Furthermore, since we are looking for spikes then small repainting issues may not be important.
Another thing -- the latest version of SpikeHunter supports 8 different envelopes. Do you see how different they are? Yet they all sort of work. This means that when the goal is to detect spikes then minor variations in curvature probably can be accommodated.
I think that settling on a good Exit condition now becomes more important. Also, I'm trying to incorporate a primitive metric that can be used to compare results when testing the different types of envelopes.
1