in /afs/cern.ch/user/r/rauscher/public/rpc_cluster.root there is a TH1F called dblt_1_2, which shows an RMS of 2.882. If I do any zooming in on this histogram, the RMS changes to 3.361. As there are only a view bins around 0 filled, by the ZOOM (by mouse on the x-axis), only empty bins are excluded.
Could this be a bug in the RMS calculation?
PS: This refers to version 5.22/00b 14 May 2009 (athena 15.20)
When zooming, the mean value and RMS are recomputed using the existing bin range with the only possible assumption that all entries in a bin are at the centre of the bin. Of course one could be clever and in case one zoom a histogram by simply excluding the empty bins at the edges, do not recompute these values.
I have the same problem, but I didn’t understand Rene’s reply: why if I exclude only empty bins the RMS should change?
As soon as you zoom, the mean/rms is recomputed assuming that entries per bin are all at the centre of the bin, so you cannot get the original value computed at filling time.
As I said in a previous reply, one could retain the original values when excluding only empty bins at the edges.
OK, now it is clear . Thanks a lot