ATR Calculation....

A

amigasearch

Hi. I am calculating ATR myself, and it works fine.

Except, for the S&P EMINI contracts, atr shows, as an example, 3.44, or similar. Looking at other charting programs, at this time, the ATR is usually 1, or around 1.
Do I need to adjust for the cost of each tick (.125)? Or, is my value ok?
As a solid example, at 10:59 today (NY time), (open is 1033, close 1032.25), I have an ATR of 3.8611, on a 14 period lookback.
Can someone point out what I am overlooking to get successful reading?

Thanks.
 
I show the 1 min-14 bar lookback of the ATR on the ESZ3 to be .775 on the bar you asked about.

You can check this manually by noticing that NO bar over the last 14 as a range of greater than 1.0 so it's very unlikely that the average of that would be greater than 3.0

Wells Wilder published the original formula. I don't have time to post it, but I'm sure you can find on net somewhere.

Good Luck
 
Thanks for checking. I can assure you my ATR is just as Wilder intended - I am correct with my calc on all stocks.
But, futures, I get this high number. Am I supposed to factor in tick price or something like this?
 
Dr. Just something I am not getting. You say that by observation, no bar over the last 14 has a range > 1. I am looking at my chart, and a 14 minute period shows a point range of nearly 3... (high of 1033 low of 1030 lets say).
This accounts why my app. is getting ATR of around 3. Am I misunderstanding range here?
 
It's not the totality of the 14 bar range that may happen to have a 3 point range that you re concerned with.

You're basically averaging the TR of the last 14 bars individually. The total range over the full 14 bars is irrelevant. So the 14ATR will be closer to the TR of each individual day of any in the 14, than it will be to the total range of the entire 14. Make sense?



I wouldn't know why your ATR function proves correct with shares but not with futures. There's nothing inherently different in the calculation between shares and futures. There's no "minimum tick" used in the calculation.

Are you trying to do this calc by hand? I would recommend you do so if you aren't. Charting packages can distort your results if you aren't totally familiar with all the moving parts of a calculation such as this.


Sorry, pressed for time today.
 
Again, I am using the formula myself (I have created it in my C++ app., trying to see if it can be used in my strategy). I got the calc. off a website - I will check that. Thanks again for the time you took to answer. It will help lead me to a solution.
 
As I said, will check calculation again on website (charts school or something), and I got it right now. Seems I was adding an extra value somewhere. Thanks for the help. I really do thank you Dr.
 
You're welcome.

So what are you researching? I'm always looking for developer talent. PM me if you wish to speak in private.


Regards,
Dr. Zhivodka
 
You must log in or register to reply here.
Back
Top