The most difficult question in the world

New comment
 
After asking a 20  of programmers without receiving an answer
I understand that the question is very difficult
Premise
a long long long time ago
1 tick = 1 price movement
therefore there were reliable tick indicators
in 2020, however, ticks are the average
so they are tick = zero movements
tick = 2 or 5 or 10 price movement
so I ask
is it then possible to see the real price movements of the last closed bar ??
it is possible to see that a bar has been made as an example
12045 up movements and 4367 down movements ??

is it possible to do this analysis based on real price movements ??

Thanks to all of you 

 
Yes it's possible.
 
ElisaP: a long long long time ago 1 tick = 1 price movement

That has not changed. PIP, Point, or Tick are all different in general.
          What is a TICK? - MQL4 programming forum

 
William Roeder:

That has not changed. PIP, Point, or Tick are all different in general.
          What is a TICK? - MQL4 programming forum

you have to believe me for many reasons of data transmission from the server and for other reasons 1 tick in 2020 is not 1 price movement

one of the many examples is to average the price to keep the spread equal

so count ticks up and down today gives very erroneous results, especially in the hours of higher data volume

 
Marco vd Heijden:
Yes it's possible.

Thank you !!

you are the first person to say this

I just started a job to find a programmer capable of counting the REAL up and down movements of the last close bar , 

make speed and minute movement thresholds

and understand if after xx minutes of the new bar open near a level we have continued price trend or rebound reversal of the delta of the up and down values .

 

The problem with that is that most of the logic, or robots that trade algorithmic mostly still use the simplest pips (do something @ x ticks up or down) concept and this creates trigger clusters somewhere above and below the bar open.

When price moves into such cluster zone there will be a slight market reaction because a large portion of the connected bot's will start to act together around the same levels, because their triggers are all around and in close vicinity of these levels.

You can kinda think of this as testing a S&R level, price moves into such zone but is immediately pushed back by all the robots that start to fire at the same support or resistance levels, of course in S&R these are visible levels because they are visually connected to an earlier high or low on the chart.

But there are many of these zones or levels that are hidden, that are set as a triggers in software, but which do not correlate to any earlier point on the chart other then anchor points in time, and these events have a smoothing, or flattening effect on the market feed, they work like shock absorbers, or have comparable action to a dynamic range compressor.

Brokers also know this and they often (at least some of them) blow up the spread to kick all these robots out of orbit, the most extreme event of this i have seen was with a shady broker that spiked the spread from a few 3 or 4 points all the way up to 200 and then back to 3 points in a matter of millisecond so you have to protect your positions against these undesirable and... unfair practices and the only way to do this is to use a stealth stop and first monitor the incoming spread, and if it widens, also adjust your stop in the opposite direction so that it becomes dynamic in both directions with a direct relationship to the incoming feed, this is an essential part in robot design, if you want your account to survive..

Moreover, such protection mechanism is impossible when you modify your stop on the sever side because you will never be fast enough to move your stop in case such event occurs.

Ok back to these invisible cluster zones.

The better you become at identifying and especially calculating these hidden levels, the better you will be able to use these zones in your entry and exit decisions.

But if you convert the data into another format, and make the equation a bit more complex you risk missing out or either undershooting or overshooting these levels i mentioned above, and the thing simply stops working.

The resulting plot does look like it follows the same curvature, i mean it's form resembles the same figure, but the levels involved are totally different, and incompatible with what is needed to make it work.

I am sure they can be made to work but that would involve a complete re-calibration of the weights and thus of all involved neurons.

It's an exact science.

The bottom line is that simple things tend to work best and you learn or discover key elements by analyzing massive amounts of data yourself because all that is hidden shall be revealed to you, given that you invest the time and effort to look for it.

New comment