Speed: are tests with implicit waits faster or slower than tests with only explicit waits?

gistfile1.txt

**QUESTION**: Are tests any faster if we use implicit waits (onboard the 
              remote browser) or explicit waits (polling remotely & setting
              browser waits to 0)?

**ANSWER**: There's no significant difference. The difference is less than
            10% on average. This isn't a significant factor in choosing the
            wait implementation to use in writing tests.

**DATA**:

All times are eyeballed from the 10 recent runs on the build time trends
page (click on an individual build, then click on the time at upper right).
Example URL: https://ci.mozilla.org/view/Persona/job/identity.browserid.dev.windows.7.firefox.16/buildTimeTrend

win/chrome
implicit time: 3:00
explicit time: 2:50 (excluding one outlier at 6:45)

win/ff
implicit: 2:50
explicit: 2:50

IE9
implicit: 4:10ish (relatively wide spread)
explicit: 4:10

IE8
implicit time: 5:50
explicit time: 5:45