Looks like the test was flaky:

0:39.89 INFO |                                   Test                                   | Subtest |          Results           | Messages |
 0:39.89 INFO |--------------------------------------------------------------------------|---------|----------------------------|----------|
 0:39.89 INFO | `/web-animations/timing-model/animations/reverse-running-animation.html` |         | **FAIL: 6/10, PASS: 4/10** |          |

Seems like pre-existing flake: https://wpt.fyi/results/web-animations/timing-model/animations?label=master&label=experimental&max-count=10&product=chrome&q=seq%28%28status%3APASS%7Cstatus%3AOK%29%20%28status%3A%21PASS%26status%3A%21OK%26status%3A%21unknown%29%29%20seq%28%28status%3A%21PASS%26status%3A%21OK%26status%3A%21unknown%29%20%28status%3APASS%7Cstatus%3AOK%29%29

Admin merging, and will ping CL author as an FYI.

The test actually pretty flaky on Firefox, https://bugzilla.mozilla.org/show_bug.cgi?id=1601322.

I suppose the test expects that the snapshot is taken after the animation in question finished? If so, can we wait for an animation finish event instead of the flaky 900ms delay?
It may be possible that anim.reverse() call in a callback of setTimout(500) is another source of the flakiness. Given that setTimeout callbacks aren't reliably invoked, it might be too late to call reverse().

@stephenmcgruer can you please this issue bounce to the test author Kevin? I couldn't find his account on github.

@stephenmcgruer can you please this issue bounce to the test author Kevin? I couldn't find his account on github.

I think Kevin might be @kevers-google

@hiikezoe - thanks for bringing this to my attention. This is being tracked in https://crbug.com/1029543