All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=811468

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5991bfdefd722dc07ffae0b810e2ea943f718863896f1163ac897574cb353cb1


Bot(s) for this bug's original alert(s):

android-one

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16a4b32d840000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/16a4b32d840000

Don't composite trivial 3D on low-end devices. by chrishtr@chromium.org
https://chromium.googlesource.com/chromium/src/+/812a74e7bc08bf89f8d8dd6b2260492e4bc318b7

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

http___mobile-news.sandbox.google.com_news_pt0?scroll:

 Regressed because this version of the site uses translate3d(x, y, 0) as a
way to scroll the content, and does not use will-change: transform. The live
version of news.google.com on mobile devices uses regular (no-script) composited scrolling of the body for the content, plus fixed-position for the header.

I conclude that this old site is not representative any more of best practices. It
should be re-recorded to the new site.

Still trying to find pathological_mobile_sites_000.wprgo.

Filed bug 813158 for the dated recording of news.google.com

Found pathological_mobile_sites_000.wprgo. Now https://recode.net fails cert
lookup, asking Ned...