We model football.
You decide.
A goals model that measures every team’s true strength, gives the probability of every match and projects every season, across 21 competitions. We deliver these numbers, dated and publicly calibrated, to media, clubs, agents, insurers and data teams.
Latest run published on (Champions League).
Each competition is recomputed once its matchday has been played, and carries its own date.
What the model says
Output of the latest published run, untouched
Research
Fixture of the weekSerie A · Sunday 11 October, 12:30
ComovsRoma
- Expected goals
- 1.2–1.3
- Most likely score
- 1-1 13%
- Both teams score
- 51%
Title race · Serie A50,000 simulated seasons
Inter2nd57%
Roma1st27%
Como8th6.0%
Juventus7th3.7%A decision to make,
a number to ground it.
A headline to write, a target to set, a club to choose, a premium to price, a pipeline to feed: one model, one answer per profession. Each card shows a number from the latest run.
The number that makes the story, before kick-off.
Pre-match probabilities, “who will be champions”, power rankings: graphics ready to publish, dated and credited to “Rondo Research”.
Media →Who will be champions · Ligue 1
Paris Saint Germain70%
Monaco14%
Olympique Lyonnais5.0%
LOSC Lille4.9%What your season is worth, luck aside.
True strength, opponent profile, quantified season objectives.
The table lies · Ligue 1
Olympique Marseille17thin the table4thon strengthChoosing a club means choosing a season.
Three destinations side by side: true level and projected season.
Finishing in Europe · Ligue 1
Lens34%
Paris30%
Auxerre1.8%Event probabilities you can price.
The full distribution of the final position, not a prediction.
Tottenham Hotspur · final position
Relegation: 45% · 8 times in 10 between 13th and 19th
The model, in JSON.
The same computations as our pages, ready to plug into your pipelines.
"marches": { "domicile": 0.349, "nul": 0.265, "exterieur": 0.385 }, "intensites": { "domicile": 1.23, "exterieur": 1.30 }
Como–Roma · computed on 27 September 2026
What the model sees this week.
Three graphics generated from the latest run, picked by a rule rather than by hand: among the major leagues, the most open title race (other than the one at the top of the page), the most finely balanced fixture, and the widest gap between table position and measured strength.
The most open race
Premier League 2026/2027 · title probability, 50,000 simulated seasons
Research
Manchester City1st · 15 pts58%
Arsenal2nd · 12 pts31%
Brighton & Hove Albion3rd · 10 pts6.7%
Liverpool6th · 9 pts1.6%
Brentford4th · 9 pts1.0%The other 15 clubs combined: 1.6%
Reading: Manchester City lead the race (58%), but Arsenal still have 31%: the title is far from settled.
On a knife-edge
Premier League · Matchday 6 · Saturday 10 October, 16:00
Research
Aston Villa16th in the table
Brentford4th in the tableReading: Aston Villa start as favourites, but at only 37%: no outcome dominates.
The table lies
Ligue 1 · table position against strength measured by the model
Research
Stronger than their position
Olympique Marseille17th in the table→4th on strength▲ 13
Lens15th in the table→3rd on strength▲ 12
Paris Saint Germain6th in the table→1st on strength▲ 5Higher than their level
Angers SCO7th in the table→17th on strength▼ 10
Le Mans9th in the table→16th on strength▼ 7
Paris3rd in the table→9th on strength▼ 6Reading: Olympique Marseille, 17th in the table, 4th on strength in the league.
The foundation is open.
What we build on it is our business.
We publish enough to judge us before you trust us. What we build on top — and how we deliver it to you — carries an honest status: what exists, what is done on request, what is still in development.
Open, free, verifiable
- The methodologyTeam strengths, score matrix, simulations: every step described.
- CalibrationForecast against observed, across 2,094 matches, published as is.
- The model’s outputsProjected tables, strengths, match probabilities for 16 competitions.
- The scopeWhat the model covers, and what it does not, stated plainly.
Ours: what we deliver
- Graphics and data packsThe infographics on our pages, adapted to your brand guidelines, ahead of every matchday.On request
- JSON APIProbabilities, score matrices, strengths, projections and calibration, fully documented.On request
- Reports and briefsSeason objectives, destinations compared, the risk of an event: one question, one quantified answer.On request
- Player impact modelWhat a player’s presence changes to his team’s goals. Open as a pilot, with its limits written down.In development
- Live modelProbabilities that move during the match, and projections conditional on the next result.In development
When we say 30%,
it happens 3 times in 10.
A model is judged by its calibration: in each probability band, the forecast against what actually happened. We publish it, gaps included — including where we perform worst.
- Home win
- 2.3%
- Draw
- 3.5%
- Away win
- 6.1%
Mean gap between forecast probability and observed frequency, 1X2 result, across 2,094 matches. The lower it is, the more the number keeps its word.
Each point is a probability band; its size is the number of cases. On the diagonal, the model keeps its word. Here, across 2,094 matches (1X2 result), the mean gap between forecast and observed is 2.3%.
From the latest matchday to your screen.
A Dixon-Coles-type goals model, re-estimated after every completed matchday. Nothing is adjusted by hand: what comes out of the computation is what you receive.
The matches are played
Each team’s attack and defence are re-estimated across the whole competition, adjusted for the strength of the opposition.
The model simulates
Each fixture receives its grid of scores; the rest of the season is replayed 50,000 times.
The model checks itself
Calibration is recomputed and published as is, good or bad.
You receive
Graphics, files or JSON feeds, each dated with the run it comes from.
24 competitions tracked.
What the model does with each.
The full model wherever the format allows, match probabilities for the European cups, strengths alone elsewhere — and domestic cups out of scope, labelled as such.
Premier LeagueFull model
La LigaFull model
Serie AFull model
BundesligaFull model
Ligue 1Full model
Primeira LigaFull model
EredivisieFull model
Pro LeagueStrengths only
Süper LigFull model
ChampionshipFull model
La Liga 2Full model
Serie BFull model
PremiershipStrengths only
Super LeagueStrengths only
Austrian BundesligaStrengths only
AllsvenskanFull model
EliteserienFull model
SuperligaStrengths only
EkstraklasaFull model
Champions LeagueMatch probabilities
Europa LeagueMatch probabilities
Carabao CupOut of scope
Europa Conference LeagueOut of scope
FA CupOut of scope
Tell us the decision you need to make.
Newsroom, sporting department, agency, insurer, data team: we come back to you with the numbers that bear on it, dated, with their limits written down.