You are correct. The entire premise of this badge is flawed and there is no way to fix this without totally rewriting the definition and rules.
@zogstrip we should talk to @dsj and remove this from the forum.
The only concern is the message that not all bugs are fixed and some are just ignored.
Gary
Was a user complaining about this? I’m still unclear as to the nature of the actual issue, is this the best example:
Your objection is that this is solely based on average score per post, then? Remember too that there is a floor, a minumum number of posts one must have to even be considered for the badge.
@zogstrip what we could do is change the math a bit to equalize it, add a small constant so someone with a tiny number of posts doesn’t accidentally dominate:
9 / (3 + 5) = 1.13
60 / (30 + 5) = 1.71
Your thoughts @eviltrout? This seems simpler than what you proposed, perhaps?
Here is real world data:
Original code:
[["michael", "2.1666666666666667", 3],
["vreihen", "1.4210526315789474", 19],
["weather-display", "1.1500000000000000", 10],
["pierre", "0.62500000000000000000", 4],
["GaryFunk", "0.48055555555555555556", 180]]
Adjusted code:
Last number is the number of unique posts with at least 1
[["GaryFunk", "2.5540540540540541", 37],
["vreihen", "2.2500000000000000", 12],
["weather-display", "2.2142857142857143", 7],
["michael", "2.1666666666666667", 3]]
Raw score:
{
"GaryFunk"=>96.5,
"vreihen"=>27.0,
"weather-display"=>15.75,
"michael"=>6.5
}
Yes, I very much like the minimum posts threshold. I think that would fix these issues.
OK so @zogstrip just add a literal hard-coded arbitrary +5 to the SQL as indicated above. That will reduce the impact of people who just barely met the minimum post threshold to get the badge, which is already 3 posts I think.
Has this update been deployed? No rush or pressure.
This is now done 
https://github.com/discourse/discourse/commit/77ac14d47552b1ee703da0f916aa7bc0d0b6dde8