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)
Posted here because I didn't quite know where else to post it. (BM will move it if it's out of place).
Is there any way to determine what the "average" high/low for range/tick based bars happens to be?
that is to say, is there any way to determine what the average high/low for up and down advanced bars is?
What I'm trying to do is get a feel for the average price movement outside the bar. Obviously an advanced bar (like Renko) only shows the open close price and the time, it doesn't show the high low.
I'd like to be able to determine if the average high is at the midpoint of the bar range or is it centric to the bar type (i.e. do up bars have a lower average high outside the bar and down bars have a higher average low) or do they follow an even distribution.
If you took 10,000 renko bars and then manually calculated what the actual high and low for the bar was, would it be half the next bar? Or as I said, would it's center of gravity be closer to the bar center depending on up or down?
Maybe Fat Tails can help.....anyone?
Can you help answer these questions from other members on NexusFi?
Thanks FT, you're usually the smartest guy in the room.
What I'm asking, is over the course of a long run, what is the "average" high and the "average" low that a range bound bar experiences?
That is to say....if you took a sample of say 10,000 bars of 10 tick value. What would the average real high (not plotted or shown) of the bar have been and what would the average real low of the bar have been?
I'm guessing that up bars have an average high somewhat less than half the bar range value and that down bars have a low that's somewhat less than half the range. I'm not sure about the lows of upward bars and the highs of lower bars (maybe they end up being half, maybe more).
The problem is that I cannot craft an effective indicator because TS will not plot anything outside the bar for historical advanced bar types. So I either have to manually calculate going backward or forward sim. Obviously in order to get good confidence I would need a significant sampling size.
The next obvious question would be....what effect does bar size have? If you have a 2 tick bar range, obviously the high/low average would probably be pretty close to 1 tick.
If you have a 100 tick bar range, the high/low average is probably somewhat less than 50 (for peaks and valleys) and on the inside, I dunno.
I would expect that the average low of the following bar is half an average true range below the midpoint of the current bar and that the average high is half an average true range above the midpoint of the current bar.
If the market has a confirmed bias towards rising/falling prices or increasing/decreasing volatility this could be different. But then it can not be measured on smaller timeframe bars, but maybe on daily bars with a larger range.
At first glance I would agree. But some rough observations show that up bars complete the peaks of longer price action and down bars complete the bottom troughs.
Therefore I would expect the average low of an up bar to be half the bar increment and the average high of an up bar to be some factor less than that.
I would expect the opposite to be true for down bars...average high being half the bar increment and average lows being somewhat less.
The second challenge to this analysis would be to examine that "shift" and to what extent it occurs with respect to the size of the bar. Obviously, if you took 100 tick bars, that phenomenon would be greater than if you used 2 tick bars.
Over the course of time, on relatively smaller range bars, (n approaches infinity) I would expect an "average" high/low to reveal itself.
On smaller bars you probably cannot measure this effect.
On daily bars you might find
-> for stocks and index futures: average of |current high - prior midpoint| < average of |prior midpoint - current low|
-> for physical commodities: average of |current high - prior midpoint| > average of |prior midpoint - current low|
Thanks Mike, I was really wondering if anyone had done this type of analysis before. Not necessarily looking for an indie.
One of the reasons i'm hesitant to put this type of stuff in the TS/EL section is because it's more a general question and I'm afraid if i put it there, a lot of talented guys who aren't running TS will never see it. Seems like the majority of guys on futures.io (formerly BMT) are ninja guys.
