Median scoring leagues explained
A median league gives every team a second game each week, against the middle of the league. Score in the top half and you win it; score in the bottom half and you lose it.
Below, the rule is explained step by step, then tried out on a real 12-team Sleeper league from 2025. That league did not use median scoring, which makes it a useful test: we can replay its 14-week regular season with the extra game added and see what changes. It is the same league as the sample report, with teams shown as Team 1 to Team 12.
How the median game works
Each week, every team still plays its normal head-to-head matchup. Then all the week's scores are lined up from highest to lowest and the middle score is found. With 12 teams, that is the point halfway between the 6th and 7th best scores. Every team above it gets an extra win, and every team below it gets an extra loss.
So each week counts as two games, and a season of 14 weeks becomes 28 games. A team that wins its matchup and also scores in the top half goes 2-0 for the week. A team that loses its matchup and scores in the bottom half goes 0-2. The interesting weeks are the split ones, 1-1, where the matchup and the median disagree.
On Sleeper this is a league setting the commissioner turns on, and it applies to the regular season. Playoff games are decided head-to-head as usual. Records in the standings include the median games, so a median league's records add up to twice the number of weeks.
Why leagues use it
The median game exists to take some of the schedule out of fantasy football. In a normal league, your record depends partly on who you happened to play. Post the second-highest score of the week and still lose, because your opponent was the one team that scored more, and you take a loss you did not really earn. Our schedule luck guide measures exactly that.
The median game does not care who you play. It only asks whether you were better than half the league that week. That makes half of every week's result a pure test of how well your team scored.
How often the two games disagree
In this league, the head-to-head result and the median result were the same in 83% of team-weeks. That is expected: a team that scores in the top half usually beats its opponent. The other weeks are the ones the median game is designed for. 14 times a team scored in the top half and still lost its matchup, usually by running into an even higher score. The median game would have given each of those teams a win back. And 14 times a team scored in the bottom half and won anyway, because its opponent had an even worse week. The median game would have handed those teams a loss.
The season, replayed with a median game
Here are the 12 teams with their real head-to-head records, their record against the median, and the two combined. Teams are ranked by combined record, with ties broken on points scored.
| Team | Head-to-head | Vs median | Combined | Rank before | Rank after |
|---|---|---|---|---|---|
| Team 4 | 12-2 | 12-2 | 24-4 | 1 | 1 |
| Team 12 | 11-3 | 11-3 | 22-6 | 2 | 2 |
| Team 9 | 9-5 | 10-4 | 19-9 | 3 | 3 |
| Team 2 | 9-5 | 9-5 | 18-10 | 4 | 4 |
| Team 3 | 7-7 | 8-6 | 15-13 | 6 | 5 |
| Team 10 | 8-6 | 7-7 | 15-13 | 5 | 6 |
| Team 8 | 6-8 | 7-7 | 13-15 | 7 | 7 |
| Team 11 | 6-8 | 6-8 | 12-16 | 8 | 8 |
| Team 7 | 5-9 | 6-8 | 11-17 | 10 | 9 |
| Team 1 | 6-8 | 5-9 | 11-17 | 9 | 10 |
| Team 6 | 3-11 | 1-13 | 4-24 | 11 | 11 |
| Team 5 | 2-12 | 2-12 | 4-24 | 12 | 12 |
Only 4 of 12 teams changed places, and none moved by more than 1. The 7 playoff teams would have been exactly the same. Among the risers, Team 3 went from 7-7 head-to-head to 15-13 overall, because it scored in the top half more often than its matchups showed. Among the fallers, Team 1 went from 6-8 to 11-17.
That is a quieter result than many people expect. In this league, records already followed weekly scores closely, so there was not much for the median game to correct. In a season where one team keeps meeting the week's top scorer, the median game can move a team several places and change the playoff field.
Does it really reduce luck?
Yes, though not by as much as it sounds. We measure luck as the gap between a team's actual wins and the wins its weekly scores deserved (its all-play win rate times its games). Head-to-head only, the average team's luck was 7.1% of its games. With the median game added, it fell to 6.0%. The median game itself has no schedule luck, but the head-to-head game still does, and it is still half the record.
What it cannot tell you
- This is a replay. Managers in a real median league might have made different moves: chasing points late in the season matters more when every point counts toward a second game.
- The median is the middle of one week's scores, so a week where the whole league scores well still hands half the league a loss. It measures you against the league, not against a fixed bar.
- One season of one league is a small sample. The effect on standings depends on how uneven the schedule was, which varies a lot from year to year.
Shoulda Started counts median games automatically: if your league uses one, the season report's records and expected wins include it. To see how much the schedule decided this same season without a median game, read how much the schedule decides a season, and for the all-play record that the luck numbers are built on, see all-play record.