I've done several hundred, possibly thousands of WSG and AB as alliance on the whirlwind battlegroup. I've kept logs of wins/losses in MS Access. The win % for alliance on these 2 bgs (as pug vs pug) is about 15%. AV is about 50%.
Other people I talk to on my server and in the battlegroup agree with my numbers. It would be interesting to see these statistics on your site for the mod because one would expect roughly 50-50 or even 60-40 win percentages, but that is clearly not happening in WSG and AB if they are played through to completion. I'm curious what people think the reason is for the huge disparity in win/loss that your mod could illustrate.
I believe the mod does not currently show this disparity not because of a logic problem with the code but rather a subtle usage pattern.
What happens in WSG is about 50% of the alliance team AFKs out after the Horde's first flag cap. I guess they had AV queued as a backup plan and shoot for that 50% chance to win at AV.
After the 2nd Horde flag cap a few more afk out.
My suggestion is:
1) treat AFK (left bg early) as a loss since not many will leave a wining WSG
or,
2) display the number of AFKs for each side (the mod might only count BGs not completed for the user, not try to detect "xyz has left the raid").
or,
3) sum losses + afk into a separate row so that the page displays wins, losses and losses + afk
The score (# of flag caps or AB score) and game length would also be interesting. What I notice is a long string of Horde 3-0 wins in 10 minute games, followed by one agonizing Alliance 3-1 or 3-2 win that takes 45 minutes. These numbers are pug vs pug, not pug vs premade.
Hopefully my curiosity can be handled with a simple html display issue!
Being able to detect premades (say 75% of the team from the same server) would also be interesting.
Thanks.
