RE: Error bars of 100% of the estimated value

From: Joachim Vollaire <joachim.vollaire_at_cern.ch>
Date: Fri, 9 Sep 2011 14:21:23 +0000

Dear Tomek

I think you did not run your calculations long enough (or did you try to bias
your problem or modify the binning structure) as the results are not
statistically significant. The 99 % error is just indicating that the error
could not be calculated, for example if you ran 5 independent calculations
and a contribution to a given bin was added in only one of those 5 runs then
the error cannot be computed.... Even for bins with a better statistic you
still get several 23 / 40 % which is quite high....

Cheers
Joachim
Received on Fri Sep 09 2011 - 18:34:17 CEST

This archive was generated by hypermail 2.2.0 : Fri Sep 09 2011 - 18:34:22 CEST