Every percentage on this site starts with a view of each team and ends with the season being played out 10,000 times. This is how it gets from one to the other.
Out of 100 games, how many does the home team win?
Start with a guess.
Drag to your guess
out of 100 games
Lock in a number to see how it compares.
2,775 games with a home and an away team, 2009 to 2026. Games at neutral venues aren’t counted.
Maybe that’s about what you expected. Maybe it isn’t. Either way, it’s there in every season, and the model has to allow for it. A win at home is worth a bit less than the same win away.
A view of every team
Where you play is one thing. Who you play matters more. The model gives every county a rating for attack and one for defence, and every result is read against both: where it was played, and who it was against.
The black line is what the model expected of the team in that game. Drawn to show the idea, not measured.
Select an option above to see what it does to the team’s rating
A drawing to show the idea. The dots, their sizes and how far the line moves are set by hand. They are not the Stats & Solos model’s weights or ratings.
Every game the team has played is in here, each one read against where it was and who it was against.
Even the biggest day moves it only a little. One big day doesn’t turn a team into contenders. A run of them will.
From ratings to a game
Put two teams side by side and the question is simple. When a team rated like Derry plays a team rated like Kerry, at Derry’s ground, how often does each side win? Years of results answer that, and the answer is the chances for the game.
86% means Derry should win about 86 of every 100 games like this one. Not all of them. On the day, anything can happen.
Play it once and the result could be anything. Play it 100 times and you start to see the balance of it.
No games played yet.
Kerry’s 10% doesn’t mean they won’t win. It means they should win about one game in ten. And Derry’s 86% doesn’t mean Derry always win. It means they should lose, or draw, about one in seven.
Then every game in a season
A league season is just a lot of games like that one. Here’s a division of eight: seven rounds, 28 games, and every one of them has its own chances. Play the season and each game gets rolled, round by round.
Mayo’s games are shaded
The table as it standsBefore round 1
Every result goes into the table. To keep it simple, teams level on points are split on score difference here, not head-to-head as in the real tables.
Where Mayo finishedOne dot a season
Before round 1 the table is listed strongest first.
One season tells you very little. Play it again and again, and count where Mayo finish. After 10,000, the count is the forecast. If Mayo win the division in 1,800 of them, their chance is 18%.
Because every fixture is played, at the right venue, the forecast already allows for who each team still has to play, and where.
The championship works the same way
The championship is played out the same way, game by game, 10,000 times, in its full format right through to the All-Ireland final. The difference is that the draws haven’t all been made. So every simulated season makes its own, following the real rules: who’s in which round, who’s at home, and no rematch of a provincial final or an earlier round where the rules forbid it.
Take Connacht, and say Galway, Mayo and Roscommon are equally good, well clear of the other four. The draw hasn’t been made, so every time it’s played it makes its own.
Connacht titlesNone played yet
Three teams just as good as each other, so they win it about as often as each other.
Why the draw matters
In a knockout, who you meet can matter as much as how good you are. Same seven teams, same ratings. Now make the draw: Roscommon get a bye to the semi-final, and Galway and Mayo meet in a quarter-final on the other side.
Connacht titlesNone played yet
Nobody got any better or worse. Only the route changed. That’s why a forecast moves when a draw is made, and why every simulated season plays out the draws that haven’t been made yet.
The 2027 forecasts go live in January, starting with the league.
See the leagueThe examples on this page use a simple teaching stand-in with made-up ratings, not the Stats & Solos model. Thanks to AI for helping me tidy up the writing.