Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
The chart on the right is a composite profile. The composite has four settings and the one on the chart is set at dynamic which covers the bars seen on the chart The first setting the composite profile has is a longer time frame (Dated composites) which starts on a specific date and runs through to the present time. The other time frames are where only a particular set of bars from a specific point are profiled (Fixed composites), a set of visible bars are profiled (Dynamic composites) or a specific number of sessions are profiled (Rolling composites).
In any mode the composite profile can automatically mark prices of high and low volume on the chart with horizontal lines which extend all the way across the chart.
The profile on the left is the Range Profile drawing tool which can profile any section of the chart over any time frame for which you have data. It was not for the day before but where the market was balancing .
first of all let me say big thanks for all the folks who contributed to that thread, and big thanks to @JonnyBoy, not trying to discard other folks because most of the comments was very helpful! Was nice few days of reading/trying.
And I hope there is still some steam is going on and people who is interested in discussing the subject is still around and would not mind to wake up on the subject again
-----------
there is important point which was appearing through whole thread, that context is very important, before we even try to go into any intraday action, we should know what to expect, etc...
So I start thinking how to collect such context. The task for me looks like we have X -> Y, where X is some prior context, Y is how the session was closed(type of the session which was after specific X). So I will have separate questions about X and Y.
#1 question about X:
it appeared that approach which was shown/discussed was mostly around prior day profile/range.
in your opinion does it make sense to collect the context where PriorDay WVAP/STDs will be used? Or PriorWeek, Or developing week (on monday close is same as session vwap). Examlpes I was thinking is something like:
question is from perspective do you see any reasonable logical aspect in that approach and if it make any sense at all, and not from perspective if is possible to do or not, ofc, we can collect all possible numbers, but is good to know if there is potential logic around it which make sense. Because some people mentioned that they using VWAP+Bands as dynamic value areas.
#2 question about Y:
this is where I have a dilemma, because I start thinking how even approach the Y, because we're not doing buy-and-hold, where we enter the position with first tick on open, and close last second before session closed, which means we do not care so much about open/close, but more about granular intraday action.
a) first approach I considered is to split the Y session range into 3 parts (0%-33%, 33%-66%, 66%-100%)
Y-session closed between 0%-33% I considered SHORT
Y-session closed between 33%-66% I considered RANGE
Y-session closed between 66%-100% I considered LONG
it works OKeish when there is strong long/short days, but there is cases where it does not looks good (imho), where for example day was a range day more or less but last few bars(or lets say last 30 minute of the session, not to confuse with bars here) strong reversed into the opposite direction. Attaching charts for examples what I mean.
b) other approach could be to consider to use averaged POC across all bars during the session (something like avg(session_bar0_poc,session_bar1_poc,...))
c) other approach could be to consider to use LinRegression across all prices during session, and then look where last point of Regression landed in 0-33%,33-66%,66-100%. but not sure about that because seems also I'm trying to overcomplicate things?! Which is fine if this would give me acceptable results
d) consider only where POC of session closed in what part of the session range, but is also giving a lot of debatable situations
e) RANGE if session closed in profile value area (70%), and if above = LONG, below = SHORT. To me sounds like logical, but visually also look sometimes/somehow strange because price can reverse quite fast and close below Value area or something like that.
f) cluster the profile distribution shapes into LONG/SHORT/RANGE, this can get more complicated because there is variation of shapes, especially when there is spreaded shapes across whole range, and not clear P/b/D shapes.
Could anybody who did that and happy with results maybe share how you're doing this? How you labeling the day/session by LONG, SHORT, RANGE, if you want to respect the granular intraday action.
Or this whole idea is somehow is the wrong direction of labeling the session like that(LONG/SHORT/RANGE)? And is better to look to something more specific what price did in the Y-target-session (which can also give many intersecting Y labels)?
#3 general question
do you guys consider collecting such things on all data you have or some kind of rolling periods like (10years, 5years, 2year, 1year, ...)? The reason I'm asking that it can be that for example you found that on 10years data there is positive expectation that if session opened between yesterday vwap_std1+ and vwap_std2+ it will be a LONG session, but then if you look deeper you notice that all this positive action happens lets say 5 years ago, and all the data after was just destroying the stat which is going down, but is still positive but whole trend of that stat is moving towards 50%. How do you go about that?