Percent BB gives erroneous results!!??

[mbuta]

Hi,

Can someone explain why %BB (from Bollinger) is less than 1, when I read Close > BB.Upper, which makes me believe that all results are not correct?

Thank you.

[dloomis]

I am not sure, but what value of close() and BB.upper are use d in the calculation.

Perhaps there is a decimal point issue?

Hard to say with knowing how you are calculating the various parameters.

Do the values in the Cursor Window agree with the calculations?

[mbuta]

It is not decimal, and the values below are from the cursor window, althgough it is obvious that the %BB is wrong:
close = 35.38
BBUpper = 35.3759
BBLower = 35.2153
%BB = .95 !!!!!!!!!!!!!!!!!!!!!

QQQ, today, 9:30 PST, on 5 min chart.
I am using buitin BB and the %BB given by eSignal.

Anybody knows why?
Thanks.

[dloomis]

My qqq, 5 shows different numbers than yours, see pic

[ACM]

Mihai
At 12:30 EST and running 3 different Bollinger Bands, one from Basic Studies, one from the Bollinger formula folder and the third from the Built-in folder.
All of the BB values match between each other and the Upper is above the Close (see image)
The only way to get an Upper to match the value you are showing is to use OHLC/4.
Alex

[mbuta]

Alex, look at your chart around 12:30 and you will that YOUR %BB has incorrect values also!.


Thank you.

[dloomis]

I may be wrong, but here is what I see on Alex's chart...

close-bBBl / BBhi-BBlo = %b

35.83-35.2189/35.387-35.2189=0.95...

Where is the mistake there?

In any event, please stop shouting.

[mbuta]

Sorry, but %BB cannot be <1 when close>BB_UPPER.
Unless I am missing something!

[mbuta]

I though close(-1) will do it, but it does not.

[ACM]

Mihai

Alex, look at your chart around 12:30...

If you look at my chart you will see that the cursor is exactly on the 12:30 bar (look at the Time label at the bottom) and nowhere in that vicinity does the Close ever go above the Upper BB. The closest it gets is at the 12:30 mark.
As I said in my prior reply the only way I can plot the BB at the values you indicated earlier is by using OHLC/4 as the data series. But you cannot change the data series in the %b so that would justify the discrepancy.

Alex

[dloomis]

My bad, I had a typo, close is not 35.83 as I wrote, but 35.38, which is less than the bbHi.

At least you arn't shouting, blue highlight is much better, thx.

[mbuta]

Sorry guys, but if you just take the time to calculate couple of values, you will see what I mean.
The values are wrong. All of them, but it is obvious aroud 1 and zero, because that what I was looking for.
And I am not using HCL/3, just the close.

[dloomis]

You are right, my mistake, sorry for trying to help me understand what I did wrong.

later

[gspiker]

I have to agree with David and Alex.
After loading up an advaced chart QQQ - 5min 6:30-13:00 TT
and adding both the regular BB study and a version of PercentB.efs modifed to print out values, it looks 100% correct.

Like David I have very different values for today at 9:30 PST than you do. See attched chart.

In all cases where the close was above the upper BB I get values greater than 1, and in all cases where I have a close lower than the BB I get values less than 0.

I tried calculating a few dozen numbers by hand too...all seemed ok.

Garth

[gspiker]

Just for completeness, here is the output from the modifed version of the formula.

Garth