Yo, posting it for the whole internet, took the whole day to find / to design the actual working solution for weighted percentile 'nearest rank' algorithm, almost no reliable info online and a lot of library-style/textbook-style solutions that don't provide on real world production level.
3) say we wanna find 50th percentile, get a threshold value: n = 50 thres = weights_sum / 100 * n 7.5 = 15 / 100 * 50
4) iterate through weights_cum until you find a value that >= the threshold: for i = 0 to size - 1 2 >= 7.5 ? nah 7 >= 7.5 ? nah 10 >= 7.5 ? aye
5) take the iteration index that resulted "aye", and find the data value with the same index, that's gonna be the resulting percentile. i = 2 data = 33
This one is not an approximation, not an estimator, it's the actual weighted percentile nearest rank as it is.
I tested the thing extensively and it works perfectly. For the skeptics, check lines 40, 41, 69 in the code, you can comment/uncomment dem to switch for unit (1) weights, resulting in the usual non-weighted percentile nearest rank that ideally matches the TV's built-in function.
Shoutout for @wallneradam for the sorting function mane ... Live Long and Prosper
Wassup every1, people been asking how to use the stuff, here's what I got at this point:
1) Resist the temptation to use percentiles to 'model' micro-structural support & resistances. First, organics that know what they're doing, ones who really know how to do it, are unbeatable there, ain't to no algo / auto learning method that will scalable & efficient enough to beat real traders in business env in terms for trading the levels. Second, if you really wanna try, algorithms 'can' be constructed to find support & resistances, but not to approximate via a formula (eg WPNR). Third, SR 'modelling' is hard to prove / disprove, who knows whether a bounce from SR was an accident or nah?;
2) Current use case for WPNRs that I've come up with is switching between different strategies. I don't wanna get too deep, just trust me shit has so delicate, beautiful & deep roots in mathematics & how markets, and the universe itself work, that all these textbook quants can sit 20 more years in a library and dem still won't get it. Alas, Take a window length 50-150, I suggest starting with 64, that will be your window for percentiles. Now, consider another window length 10-30, that will be the window for the actual trades, start with 16.
For volatility trades, get into momentum trades above 90th and below 10th percentiles, otherwise consider mean-reversion trades; For direction trades, notice the ''overall' slope of the 50th percentile: - If it's positive, directional trades above 50th percentile and counter directional trades below 50th percentile; - If it's negative, directional trades below 50th percentile and counter directional trades above 50th percentile.
Other use cases & including using 1st and 100th percentiles are possible as well, but so far that's it.