Wow, I am impressed (and grateful) that you reacted to my initial post and coded the revision to the algorithm in response to my comments (especially while ill), and that goes to the latest modification too. I will test it and provide my feedback and any other information needed to help work out any "bugs".
As I mentioned I think there may some potential use for this indicator and will continue testing for results too. Concerning item 2, I'm a bit unclear what you mean by "practical tests". When I noticed the zero plot line disappearing I was using my normal version of Ninja, trading the instruments I usually do (in paper), and in my own style, so although the indicator may not practical for my trading style and use, at least to me the test was "practical". If desired I can provide whatever information you need to either replicate the problem or anything else that might help. Of course with the +/- 25 and 50 grid lines in the latest version that situation may not be as relevant.
I use PFE as part of my toolbox -- and helps me a lot with my trading (ES /GC/SI and many other instruments). What I try to say is as a stand alone indicator is probably ineffective but coupled or tripled with other indicators (do your own research) may give you the advantage you are looking for to be ahead of the herd.
I started using the revised PFE in my futures trading and it can point out retracements and reversals reasonably accurately, and especially as noted when used with other indicators for confirmation. One way that I've been using it is to put multiple instances of this indicator into the same panel, one set up as fast and the other a bit slower (fast + about 70%) . I've also added a third one that is setup to use a different data series (2 M vs. a 1M). The crossings can be used as a possible signal too, either as an entry or exit. I have noticed an occasional anomaly (use attached chart as a reference). The dark blue line is the fast PFE and notice that in the earlier example it does not change direction with the price and it does in the later example, even though the price moved about the same number of ticks. This condition does not happen very frequently and is most noticeable on 1 minute charts. For now, when I see this divergence I don't trade.
The following user says Thank You to Cheech for this post:
I used Snagit to do the screen capture as I wanted to add in some notations. When I saved the file I didn't know what format this site supported so I used the default which is a .pdf. I will use .jpg or .png in the future as I see it works much better. TU.
Any thoughts on the anomaly? Sometimes the indicator just fails to react to the price action. As another example, in the charts below the one on left is a 1 min chart and at 4:55-5:00 AM the blue indicator plot line is just about flat while the price is clearly down. The chart on the right is a 2 min chart and the indicator shows a down trend. Sometimes (but not always) it just doesn't react as expected when there's a quick change in price direction. The fast settings (11,6) are the same on both charts.
I took a quick look at your code and although I am a retired programmer/systems analyst, C (or C#/C++) is not a language I know so it's a bit over my head at this point.
I have some issues with PFE too. Sometimes the indicator's signal is working fine and sometimes not. And the user need to find the interpretation for its behaviour. For example if my stochastic hoovers at the top or shows a flat line this anomaly could be an indication of a very strong momentum building in certain direction.
BTW, all indicators are like that and if they help you with your trading most of the times and you also have an idea when the inconsistency will happen I think that this is good enough reason to use it.
The following user says Thank You to dario1 for this post: