Bump BookmarkPrompt experiment up to 10% in experiment group.

Bump BookmarkPrompt experiment up to 10% in experiment group.

This also bumps up the Control group to 10%, which is the preferred practice for experiment group sizes (makes it easier to look at the stats later when we have comparable groups that are the same size).

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=186186

[SteveT, 2013-03-04 19:12:02]

yosin: review
sky: OWNERS review

Hi Yoshifumi-san,

We've received an OK to bump up the Bookmarks Prompt experiment to 10%, since
we're not seeing a lot of data from the new installs.

I'm also changing the size of the Control group to 10%, which should make it
easier for us to compare it to the Experiment group. Note that this does not
change the behaviour for the remaining 80% in the default group.


Steve

[sky, 2013-03-04 22:08:40]

LGTM

[yosin_UTC9, 2013-03-05 00:35:37]

LGTM

Is this for M25 stable only?

Just curiosity, what's happened rest of 80% users, 100% - 10% control + 10%
experiment?

[SteveT, 2013-03-05 14:10:13]

On 2013/03/05 00:35:37, Yoshifumi Inoue wrote:
> LGTM
> 
> Is this for M25 stable only?

We are discussing merging this into the M26 beta branch so we get the data
sooner, but we probably won't mess with the Stable version.

> 
> Just curiosity, what's happened rest of 80% users, 100% - 10% control + 10%
> experiment?

The 80% leftover goes to the DEFAULT group. The way it works is that we set out
a total weight of 100 units. We set aside 10 for EXPERIMENT and 10 for CONTROL,
which leaves the rest to DEFAULT.

In the old version, we had a total of 100. You set 1 to EXPERIMENT and 99 to
CONTROL which left 0 for DEFAULT. This is fine, but it's easier to compare 10
and 10 rather than 99 to 1.

[commit-bot: I haz the power, 2013-03-05 14:10:34]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/stevet@chromium.org/12398016/1

[commit-bot: I haz the power, 2013-03-05 16:19:45]

Change committed as 186186