Hello everyone,

Perftools has successfully completed the first part of the PerfCompare
Statistics Project to significantly improve the accuracy and
interpretability of performance data analysis!

You can test it now on PerfCompare <https://perf.compare/> by selecting
Mann-Whitney-U from the test version dropdown (default is set to Student-T)
in the table controls of the results page.

[image: image.png]

The new test uses Mann-Whitney U Test, Cliff's Delta, and the Common
Language Effect Size (CLES) to determine if the changes you made are
statistically significant. Additionally, more statistical information is
provided to help with interpreting the results when the row is expanded.

The goal is to provide developers with clearer, more reliable, and easily
interpretable results on performance changes.

The Mann-Whitney comparison technique is still in the experimental stage.
If you encounter any issues, have suggestions for improvements, or wish to
provide feedback, please report it on the Testing :: Perfcompare Bugzilla
component, or reach out to us in the #perfcompare
<https://matrix.to/#/#perfcompare:mozilla.org> Matrix channel.

-- 
*Kala Severe*
Senior Software Engineer
Firefox Performance

-- 
You received this message because you are subscribed to the Google Groups 
"[email protected]" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/a/mozilla.org/d/msgid/dev-platform/CAFCgvPqQVyAQYyByaw--Vb%3D5BbywDMbD8YgzdkRtoP3Ar5EFKQ%40mail.gmail.com.

Reply via email to