Hi Volker,
On 2013-06-28, Volker Braun <vbraun.n...@gmail.com> wrote:
> ------=_Part_728_19221634.1372430567336
> Content-Type: text/plain; charset=ISO-8859-1
>
> *Putting on my Physicist hat*
>
> You should be taking the actual standard deviation into account, too. Most
> likely, patchbot number one gave less consistent timings, possibly due to
> background jobs, mechanical harddisk, or memory pressure. This is why it
> failed to detect any effect, not because it is not there but because it was
> hidden in the statistical noise.
This is why I said that it did not give anything significant.
Fortunately, for statistics dummies like me, the plugin even tells with
what confidence it finds regressions or improvements.
Speaking about statistics: On my own machine, I made 20 runs of
"sage -startuptime" with (1) only the dependencies of the ticket
applied, and (2) with the patches applied. I found
(1) Mean: 1484.42 ms
Standard deviation: 6.35 ms
(2) Mean: 1494.93 ms
Standard deviation: 6.91 ms
(more details on #13589)
Can you tell me how to assess whether the change of the mean is
significant? Is it as simply as "the difference of the means is less
than twice the standard deviation, hence, it is not significant"?
Again, I am a statistics dummy.
Cheers,
Simon
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/groups/opt_out.
