Background
The current hall of fame / leaderboard appears to rank users mostly by the generated score. As a result, once very famous or legacy GitHub users are checked, they can immediately jump to the top and stay there for a long time.
Problem
This makes the leaderboard easy to dominate by a few historically high-score accounts:
- Community-active users and ordinary users become hard to discover.
- The top of the leaderboard becomes relatively static after well-known users are added.
- The ranking does not reflect current community activity or recent interest.
Proposal
Add a composite ranking model that combines hotness + score instead of relying only on the roast score / total score.
Possible signals:
- Roast score or GitHub profile score, normalized.
- Recent query/check count.
- Recent unique visitors or unique users who checked the profile.
- Time decay, so old entries gradually lose ranking weight unless they continue to receive attention.
- Optional separate tabs such as
Trending and All-time if preserving the existing all-time ranking is still useful.
The exact formula can be adjusted, but the key point is that old high-score accounts should not permanently occupy the top of the default leaderboard.
Expected outcome
- The default leaderboard surfaces users with both strong scores and recent activity.
- Famous legacy accounts can still rank highly when they are currently popular, but they should not permanently dominate only because they were checked once.
- Community-active users and normal users have a realistic chance to appear near the top.
Background
The current hall of fame / leaderboard appears to rank users mostly by the generated score. As a result, once very famous or legacy GitHub users are checked, they can immediately jump to the top and stay there for a long time.
Problem
This makes the leaderboard easy to dominate by a few historically high-score accounts:
Proposal
Add a composite ranking model that combines
hotness + scoreinstead of relying only on the roast score / total score.Possible signals:
TrendingandAll-timeif preserving the existing all-time ranking is still useful.The exact formula can be adjusted, but the key point is that old high-score accounts should not permanently occupy the top of the default leaderboard.
Expected outcome