Light Crude Oil FuturesEducation
Flawed concepts: the way they sell volume & market profiles
Imagine daily profiles put on each day on the chart, and let the words in purple provoke some fruitful doubts in your mind.
The presented and advertised way of using volume & market profiles is essentially a way of approximating the real levels. This way is very bad.
1) The most fundamental and very obvious flaw is that it disregards the sequence of events. You can take a chart, reshuffle bars in order, or lol, just invert it horizontally. and you'll end up with different charts that have the same profile. All good bro? It's time series lol, sequence does matter. That's why you can't use profiles and non-weighted stats unless you have a very specific goal;
2) 70% rule, normal distributions & standard deviations have nothing to do with aggregated tick data. As a process, it can all can be modeled as a morphing distribution, a constant fight between normal and uniform distributions, the double auction dynamic distribution. But yeah, ofc course you can't read it in a book, gotta think for yourself a lil aye? In normal distribution 70% make sense, in the uniform one it makes sense to consider 100% (the whole distribution) as the area of interest. So overall it's somewhere in between 70 and 100. Also, confirmed with my R&D, bots give the best performance when a price channel includes 80-90% of data (mostly 80), best metrics converge around this number;
The real way to use profiles is when you consciously need to disregard an order of events within a certain period and have some kind of summary of it. It's applicable when a certain period has some distinguishable significance: when levels are formed, positioned and cleared. There, you are not really interested in order of things within these events, rather in summaries of these events, there you're doing the right thing. Otherwise, POCs of 'every' period doesn't have equal significance. Summarizing volumes within a week/day/months etc, making a profile & taking 70% so you gonna get VAHs and VALs of it won't magically calculate you the real levels, only approximate em, but 4 real there are better and less computationally intensive ways of doing it, just get a box plot with 10th and 90th percentiles.
This video might be really mind opening for you, I really respect it & its creator.
If you wanna know how to find levels 4 real, check the linked ideas & use it with pleasure.
The presented and advertised way of using volume & market profiles is essentially a way of approximating the real levels. This way is very bad.
1) The most fundamental and very obvious flaw is that it disregards the sequence of events. You can take a chart, reshuffle bars in order, or lol, just invert it horizontally. and you'll end up with different charts that have the same profile. All good bro? It's time series lol, sequence does matter. That's why you can't use profiles and non-weighted stats unless you have a very specific goal;
2) 70% rule, normal distributions & standard deviations have nothing to do with aggregated tick data. As a process, it can all can be modeled as a morphing distribution, a constant fight between normal and uniform distributions, the double auction dynamic distribution. But yeah, ofc course you can't read it in a book, gotta think for yourself a lil aye? In normal distribution 70% make sense, in the uniform one it makes sense to consider 100% (the whole distribution) as the area of interest. So overall it's somewhere in between 70 and 100. Also, confirmed with my R&D, bots give the best performance when a price channel includes 80-90% of data (mostly 80), best metrics converge around this number;
The real way to use profiles is when you consciously need to disregard an order of events within a certain period and have some kind of summary of it. It's applicable when a certain period has some distinguishable significance: when levels are formed, positioned and cleared. There, you are not really interested in order of things within these events, rather in summaries of these events, there you're doing the right thing. Otherwise, POCs of 'every' period doesn't have equal significance. Summarizing volumes within a week/day/months etc, making a profile & taking 70% so you gonna get VAHs and VALs of it won't magically calculate you the real levels, only approximate em, but 4 real there are better and less computationally intensive ways of doing it, just get a box plot with 10th and 90th percentiles.
This video might be really mind opening for you, I really respect it & its creator.
If you wanna know how to find levels 4 real, check the linked ideas & use it with pleasure.
Related publications
Disclaimer
The information and publications are not meant to be, and do not constitute, financial, investment, trading, or other types of advice or recommendations supplied or endorsed by TradingView. Read more in the Terms of Use.