[Mac Triage] Looks like this was added in https://codereview.chromium.org/2261713003. Assigning to rnephew@.

Other than the one bump, it looks like the ref build follows it for the most part. The BattOr tests are still very noisy and prone to environmental issues.

sullivan, charliea: I thought that BattOr was no longer noisy? Or is that still a work in progress?

Charlie, do you know what's going on here? This one is a lot noisier than the one you showed me in your doc.

Something about these stories makes them significantly noisier than the battor.power_cases benchmarks: https://chromeperf.appspot.com/report?sid=516b507aa5b40f7e9b527d8b6f8b4653461883ae309ca0713247f3e34a482178&rev=435383

I definitely think that this is worth exploring. It should be noted, though, that even with the noise, the shape of this graph never should have resulted in an alert: the reference build moved with the ToT build, indicating that no regression took place. 

Also, if you extend the end of the time range from 12/2 to now, the graph has gotten significantly smoother since the start of December: https://chromeperf.appspot.com/report?sid=30ca24c9235c23fcf50b9812c874b396fe67ddea8fe19f15f917870478911abe&start_rev=431798&end_rev=440849&rev=435383, although I'd be lying if I said I could tell you why. Is it possible that the cooling systems in the lab have become more reliable since then?

Looking at the traces, it looks like the biggest difference in November between the high and the low runs is that *all* power samples in the high runs were much higher than those in the low runs, indicating that perhaps the fans were constantly running for the high power runs and not running nearly as much for the low power ones.

Available, but no owner or component? Please find a component, as no one will ever find this without one.

I think the Battor tests are gone, so there's nothing actionable here.