Home › Methodology

Ratings built on evidence,
not reputation

Manager performance is often judged too much on results and not enough on the processes behind them. Results in football are noisy and affected by many factors outside a manager's control, therefore it takes a large number of games to make a reliable assessment. From a Bayesian perspective, we need enough evidence over time before drawing valuable conclusions. In reality, managers are often sacked before this happens, without enough time to develop their ideas, making proper evaluation difficult. Common industry metrics like win percentage and points per game are simple to use, but do not solve this issue. They miss important context such as squad quality, injuries, and the difficulty of matches, and do not fully reflect how a manager is truly performing.

Objective performance, measured continuously

Every manager and club has an ELO rating. Ratings start at 1500, which represents performance in line with expectations. After each match, the rating moves up or down depending on how surprising the result was, with older matches decaying in influence over time, so the ratings reflect sustained long-term performance.

A manager's ELO stays with them across jobs. It does not reset when they move between clubs or leagues. A rating built at a mid-table Championship club carries into the Premier League, and future results are judged relative to that starting point. This creates a continuous record of performance across every role they have held, where the rating adjusts over time as more evidence is observed.

1500 is the baseline. Above it, a manager is performing above expectations, and below it, they are performing below expectations. The scale is symmetric, so a move of 50 points above 1500 carries the same weight as a move of 50 points below it.
1600+Elite
1500 – 1600Above average
1400 – 1500Below average
< 1400Poor

Independent signals, one rating

Each manager's ELO is built from independent signals, each capturing a different dimension of performance relative to context. Each signal asks a distinct question: how did this manager perform relative to what was expected of them, given the resources and circumstances they were working with?

Using independent signals reduces variance and produces a more stable and consistent rating over time. It also reflects a deliberate choice in how the model is designed. Much of football analytics has shifted toward combining large numbers of metrics into composite scores, which can make outputs flexible enough to fit a narrative or justify a decision that has already been made. The signals we use capture the most meaningful dimensions of performance in context. The goal is not to maximise complexity, but to focus on measures that remain robust across leagues, squads, and time.


A tool for smarter manager recruitment

Clubs can use TouchlineAlpha to identify promising managers flying under the radar, or to evaluate a shortlist of candidates in a more structured way before making an appointment.

The ratings go beyond a single number. Because they are tracked continuously over time, they show not only where a manager ranks today, but how their performance has changed across recent seasons and across their full career. A manager coming off a strong season may have a long history of underperforming results. Equally, a well-regarded manager may have been quietly underperforming for years, struggling to adapt as the game has evolved around them. Reputation and current form do not always point in the same direction.

Performance can also be broken down by individual tenure. This allows clubs to see exactly where and when a manager performed at their best, and under what circumstances. Was the success tied to a particular type of club, a specific league context, or a level of resource? Does that profile match the hiring club? The ratings cannot guarantee future performance, but they give clubs a much stronger basis for judging whether a manager is likely to replicate their success in a new environment.


38 leagues. 25+ years. Updated weekly.

We cover professional leagues across Europe, North and South America, Asia, and Oceania, with data going back to 2000 for the major competitions. A manager in the J-League is rated using the same method as one in the Premier League, so comparisons across leagues are directly possible. The ratings already account for the different expectations that come with each role, depending on the club and the league, which means performance is evaluated relative to context rather than raw results alone.

5,585
Managers rated
1,197
Clubs tracked
38
Leagues covered
25+
Years of history