Problem with histogram with strange error bars

[root0]

Hi,

Attached is a weighted histogram obtained by ROOT. I really can’t understand why some of the points have so large error bars? Actually I doubt that they are the error bars because when filling the histogram, I didn’t feed in the errors. So what are those very long tails along the y axis? Can anyone help me? Thanks!

[Axel]

Can you provide us with the input values you use to fill the histogram? Or you can e.g. call TH1::SetBufferSize(1000000) and store the histogram to a ROOT file, that will provide us with up to one million entries you fill into the histogram.

[root0]

Hi Alex,

Thank you very much for your response. The variable I’m filling is actually the pole mass of the Z boson (but this histogram corresponds to one background in my project which does not have a Z boson, so the reconstructed mass is not at the pole ), I do have a root file for the histogram and it’s attached ( The attached root file corresponds to another background, which also has this problem as you will see). Looking forward to your response again. Thanks!

mz.root (6.5 KB)

[root0]

Could someone help me out of this?

[Wile_E_Coyote]

You need to provide your macro which generates this histogram (together with any required input data files).

[root0]

Hi WE,

Thank you for your reply. I thought about this issue again and now I found the reason. I checked the number of events for those “problematic” cases and found that only a few of them “behave” that way. Then I suddenly understand that it’s my selection rules that make this happen: for different cases, I used different weights. I plot mz again and confirmed my thought.

Thanks again, your reply inspired me to think in that direction.