Maya shows up to board game night every single week. Twelve games in, she's won 8 of them - a solid 67% win rate. Ben, on the other hand, has only made it to five games. He's won 4 of those, giving him an 80% win rate.
When you sort your leaderboard by win percentage, Ben is ranked #1. But something feels off about that. Maya has proven herself against way more opponents, in more situations, over more weeks. Is it really fair to rank Ben higher based on just five games?
Or take another scenario: Zoe has a 70% win rate, but she's only played against beginners. Chris has a 60% win rate, but he's been facing the toughest players in your group. Who should rank higher?
These are the kinds of questions that come up when you're trying to keep a leaderboard fair. Simple win-loss records don't capture the full picture. So what actually makes a ranking system fair?
Let's look at different approaches to tracking rankings, with honest trade-offs for each.
If you're tracking scores for any kind of ongoing competition - board game nights, office ping pong, video game tournaments - you've probably noticed that simple win-loss percentages can feel... off.
The problem shows up when people play different amounts, face different opponents, or when someone wins 3 games and suddenly they're #1. New players are hard to place fairly. People who take a break come back to find their ranking meaningless. And if the rankings feel broken, people stop caring.
You need a system that stays fair across different situations. Here are three common approaches and when each makes sense.
The idea: Players must complete a minimum number of games before they're ranked. Simple and straightforward.
Set a threshold like "Must play at least 10 games to be ranked." Until players hit that number, they show as "Unranked." Once they cross the threshold, use win-loss percentage or points.
Example:
This works well for short-term competitions where everyone plays roughly the same amount anyway. If your season is only 8-10 weeks and the group is consistent, requiring a minimum filters out lucky streaks from small samples.
But it breaks down for casual ongoing competitions. New players take weeks before appearing on the leaderboard. Take a month off and you drop from #3 to unranked. The cutoffs feel arbitrary - is 10 games really that different from 9? And when you're unranked anyway, why show up?
Your office ping pong ladder has a 15-game minimum. Tyler is ranked #1 with a 70% win rate over 20 games. Then he goes on a 3-week vacation. When he returns, a rule kicks in: "Rankings expire after 30 days of inactivity." Tyler drops from #1 to unranked, even though he's still the best player. That feels broken.
The idea: Give more credibility to players who've proven themselves over more games.
Calculate win percentage or total points, then apply a confidence multiplier based on games played. Players with more games get a boost to account for statistical reliability.
Example formula:
Adjusted Score = (Win % × Games Played) / (Games Played + Bonus)Maya (67% × 12 games) would score higher than Ben (80% × 5 games) despite Ben's better record.
This rewards consistency and lets everyone be ranked from game 1 without arbitrary minimums. More data means more trustworthy rankings.
But it still doesn't account for opponent strength. Your board game group uses this system. Zoe has played 15 games with a 60% win rate, mostly against beginners. Chris has played 8 games with a 75% win rate, only against the three best players in the group. Zoe ranks higher because the system doesn't know Chris faced tougher competition.
It's better than raw percentages, but it still misses the context of who you beat. Works when competition quality is relatively equal across all games.
The idea: Instead of tracking raw wins and losses, track a number (a rating) that goes up when you win and down when you lose. The amount it changes depends on your opponent's rating.
Everyone starts at the same rating (e.g., 1200). When you beat someone rated higher than you, your rating jumps up significantly. When you beat someone rated lower, it goes up a little. Losing does the opposite. Over time, ratings converge on your true skill level.
This is how Elo ratings work (originally designed for chess, now used everywhere from video games to sports).
Example:
After these games:
Now Maya and Ben can be fairly compared even though they've played different numbers of games against different opponents.
This approach shines when participation is irregular. Ratings stay accurate whether you play 3 games or 30. The system accounts for who you beat. New players can join mid-season and be fairly ranked within a few games. Take a month off, come back, your rating is still valid. No minimums needed.
The trade-off is complexity. You can't just "count wins" - it requires math behind the scenes. And you need to track who played whom, not just final scores. The first few games swing ratings around, though this settles quickly.
Your friend group tracks Smash Bros rankings with ratings. Players come and go as schedules allow. Some weeks you have 8 people, other weeks 4. The leaderboard stays meaningful because beating the #1 player moves you up significantly, new players quickly find their level, missing a few weeks doesn't erase your rank, and you can't game the system by only playing weak opponents.
It requires slightly more setup but stays fair automatically for ongoing competitions with irregular participation.
Here's a simple decision framework:
Minimum game requirements work for short-term leagues (8-12 weeks) where everyone plays roughly the same amount. You want dead simple "just count wins" and have a small group with consistent attendance.
Weighted by games played works for ongoing competitions when you want to value consistency and opponents are roughly equal strength. You're comfortable with moderate complexity.
Rating-based systems work when irregular participation is normal, people play different numbers of games, opponent strength varies significantly, and you want rankings that stay relevant over time. You're okay with math happening behind the scenes.
If you decide a rating-based approach makes sense for your situation, you have a few options:
Manual calculation: Use the Elo formula yourself after each game. This works but gets tedious fast.
Spreadsheet: Set up formulas to calculate ratings automatically. You'll need to input results and it'll update rankings.
Automated tool: Use software that handles all the math for you - you just enter who won each game. Tools like shmelo are built specifically for this: track any kind of competition, Elo ratings update automatically, and your leaderboard stays fair whether people play every week or once a month.
Regardless of which approach you choose, the key insight is this: fair rankings need to account for both how much you've played and who you've played against.
Simple win-loss percentages only capture the first part. Rating systems capture both. The trade-off is simplicity vs. sophistication - pick what matches your group's needs.
For most casual, ongoing competitions with irregular participation - board game nights, office game leagues, friend group tournaments - rating-based systems solve the problem elegantly. You don't need to understand the math. You just need to know it keeps things fair automatically.
If you're currently using a simple spreadsheet and running into the issues described here - players skipping games, rankings feeling unfair, newcomers hard to place - it might be worth trying a different approach.
The good news is you don't need to become a statistician. Modern tools handle this for you. Whether you go the DIY route with a spreadsheet or use an automated tool, the important thing is picking an approach that keeps your competition fair and fun.
After all, the goal of a leaderboard isn't just to rank people - it's to keep everyone motivated and engaged. And that only works if the rankings feel fair.
Want to try Elo rankings for your group? Shmelo handles all the calculations automatically - just track who wins and the leaderboard stays fair.
Create a leaderboard →