Experiments - while waiting for an order to get closed
So went on a search spree for better ways to calculate it. There are some solutions and attaching an original paper published by Daniel Lemire in 2008 (it is a public document downloaded from here : https://r-libre.teluq.ca/207/1/webmaximinalgo.pdf) . In it you can find a pseudo code for how he solved. But, for time series data in trading terminal(s), that algorithm might cause errors in some cases (partial data reload, just one example). Hence, I made what could be called rolling minimum-maximum that does not suffer from that kind of issues (and is even faster due to different way how memory is allocated and used).
One might wonder what it (the minimum-maximum) is used for. And frankly, probably in majority of indicators used there is some sort of it used. Attaching just one simplest way of looking for trend via simple minimum and maximum
_________________________
All the parameters as usual except for one :
The rest is fairly simple to see what it does and how it can be used as any other trend calculating indicator
Stochastic with some small twists (all in line with the experimenting) : it is using streaming maximum-minimum for calculation it can use any price for calculation (including some completely "non-standard" prices that are specific to stochastic only) unlike the built in stochastic that can use only 2 price combinations PS: what is called a "close/close" price in the standard (built in) stochastic is actually a stochastic of close price. Results are going to be exactly the same PPS: average initial execution time for 10.000 bars is 8-900 microseconds...
At the example bellow :
Setting channel period to less than 2 or channel percentage to <= 0% or >= 100% turns the channel off
Of the rest of the "qualities" of such "wonder" stuff (like, God forbid, trying to use that from an EA) is out of the scope of this post - I leave it to the pundit(s)
______________________
PS:
period 1 : (p1 * 1) / 1
period 2 : (p1 * 1 + p2 * 1) / 2
period 3 : (p1 * 1 + p2 * 2 + p3 * 1) / 4
period 4 : (p1 * 1 + p2 * 2 + p3 * 2 + p4 * 1) / 6
period 5 : (p1 * 1 + p2 * 2 + p3 * 3 + p4 * 2 + p5 * 1) / 9
...
and so on and on ...
______________________
PPS: seems that these days knowing what the code actually does is not required to be a coder. Oh well : the times of "wonder" indicators seem to be finally here and the "wonder" indicators creator(s) are just flooding us with their "wonder" creations day in, day out. Hallelujah : we are all saved now /s
{ virus risk deleted by staff }
Frankly I can not decide if your stupidity is greater than your ignorance or it is your ignorance that is greater than your stupidity.
Erudite comments like this :
and???i know about this effect, that 17 = 18 and so on... or do you think, that you opened America? again?
when i was writing & testing these induks: i saw this effect.. but i don't care about it.
are always a pleasure to read and I frankly am thanking you for presenting to all of us :
Since the rest of your so "profound" answer is same as the above math evolutional and revolutional revelations, it is only fair to tell that you are as stupid as you are ignorant ... I would tell "Keep it up", but knowing you, you can not help your self but to do exactly that (and a tons of it) so I shall say the usual :
all the best durak
goodbye durak
