Five times between 2017 and 2021 we asked our rating model to name the twelve best under-23 players in the Swiss Super League. Then we gave every one of those 60 players five full seasons and counted how many became regular starters at Swiss Super League level or above, meaning at least two seasons of 1,800 minutes, which is twenty full matches.
Thirty-five of those sixty became regular starters, a rate of 58%. Among the players the model ranked lower, compared with ours at matching prices, 34% did.
That is the question worth asking about a rating. Not whether a player is interesting, but whether he is strong enough to hold down a place at this level, at the money a club can actually spend.
The model saw only matches played before each date. We took its top twelve, grouped every under-23 in the league by what they were worth, and compared each pick only against the players who cost the same.
Where it helps most
| Price on the day | Picks | Became regulars | Others | Became regulars | Times as often |
|---|---|---|---|---|---|
| No public value | 3 | 0% | 138 | 9% | none |
| Up to €150k | 2 | 100% | 85 | 12% | 8.5× |
| €150k to €400k | 5 | 40% | 112 | 11% | 3.7× |
| €400k to €800k | 11 | 55% | 121 | 27% | 2.0× |
| €800k to €1.5m | 12 | 67% | 45 | 33% | 2.0× |
| Above €1.5m | 27 | 63% | 47 | 47% | 1.3× |
| All 60 picks | 60 | 58% | 548 | 34% | 1.7× |
The edge is widest in the lower price ranges, where the market is least sure of itself, and narrows as prices rise. Above €1.5m it falls to 1.3 times, because at that price the market already knows what it is looking at. Between €400k and €1.5m, the range most clubs in the league are actually shopping in, our picks became regulars twice as often as the players standing next to them.
And not because they were already playing
Half our picks were already regulars on the day we named them, which raises the fair question of whether the rating is describing the present rather than predicting anything.
So we asked the model a different question. Same five dates, same league, same age group, except this time it could pick only from players who had not yet held down a place: under 1,800 minutes at that level in the season before the list.
| The model choosing from | Picks | Became regulars | Others | Became regulars | Times as often |
|---|---|---|---|---|---|
| The whole league | 60 | 35 (58%) | 548 | 34% | 1.7× |
| Only players not yet holding a place | 60 | 24 (40%) | 484 | 24% | 1.7× |
The same 1.7 times, whether the model picks from the whole league or only from the players who had not established themselves yet. Strip out everyone who was already a regular and the edge does not move.
The names
The twelve widest gaps across the five lists, ranked by how far a player's own total exceeds the average for his price. Club and price are as they stood on the day the list was made. His minutes are his own five-season total. The last column is what the other players at that price averaged over the same five seasons, with how many of them there were in brackets.
| Player | List | Club then | Price then | His minutes | Others at his price |
|---|---|---|---|---|---|
| Timothé Cognat | 2020 | Servette | €900k | 14,176 | 5,958 (13) |
| Miro Muheim | 2020 | FC St. Gallen | €900k | 11,008 | 5,958 (13) |
| Stefan Knezevic | 2017 | FC Luzern | €100k | 10,937 | 2,824 (19) |
| Arthur Cabral | 2020 | FC Basel | €2.7m | 9,562 | 3,953 (7) |
| Leonidas Stergiou | 2020 | FC St. Gallen | €1.8m | 9,398 | 3,953 (7) |
| Baltazar | 2021 | FC Sion | €800k | 9,141 | 2,708 (11) |
| Eray Cömert | 2020 | FC Basel | €2.3m | 9,085 | 3,953 (7) |
| Jérémy Guillemenot | 2019 | FC St. Gallen | €500k | 9,076 | 3,223 (35) |
| Michel Aebischer | 2017 | BSC Young Boys | €400k | 8,765 | 3,765 (18) |
| Toni Domgjoni | 2018 | FC Zürich | €250k | 8,233 | 1,881 (23) |
| Silvan Sidler | 2018 | FC Luzern | €200k | 8,060 | 1,881 (23) |
| Betim Fazliji | 2020 | FC St. Gallen | €100k | 6,463 | 1,266 (15) |
Miro Muheim plays for Hamburg now, Eray Cömert for Torino, Michel Aebischer for Pisa. Others simply stayed and started, season after season, at clubs that could afford them. At these prices that is the outcome that pays.
Five lists, all five ahead
Average minutes per player at that level across the five seasons following each list.
| List | Our picks | Others at the same price | Difference |
|---|---|---|---|
| 2017 | 6,142 | 5,143 | +1,000 |
| 2018 | 5,626 | 4,302 | +1,324 |
| 2019 | 6,262 | 5,076 | +1,187 |
| 2020 | 7,727 | 4,123 | +3,604 |
| 2021 | 6,128 | 4,303 | +1,825 |
What this shows
A rating is easy to validate badly. Rank the players, wait a few years, then point at the ones who did well. That test flatters every model ever built, because the expensive players were always going to play.
This one is built the other way round. Every pick is judged only against players who cost the same, on the same day, in the same league, at the same age. The ratings are archive values written before any of the football happened, with no projection or ceiling anywhere in the calculation. Every pick from every eligible list is in.
It holds through all of that. Fifty-eight per cent against thirty-four on two regular seasons, seventy-five against fifty-two on one, and 34% more league football per player across the five seasons. All five lists ahead of the players who cost the same. And the identical edge again when the model is allowed to pick only from players who had not yet held down a place.
In plain terms: at the same budget, this is the difference between signing a starter and signing a rotation player. And it matters most between €400k and €1.5m, which is where most clubs in the league are actually shopping.
We rate about 164,000 players in more than 500 leagues across 125 countries. We can run this same replay on your league and your squad, live, on a call.
How we ran it
The ratings are strictly point in time. Each comes from an archive written on the day, and the model had seen only matches played before that date. No projection or ceiling is involved anywhere in this test, so nothing here is fitted with hindsight.
Every under-23 in the league on each date is used, and every pick is compared only against players standing in its own price band. The pooled figure is weighted by where our picks actually fall, so it answers what the rating is worth for the players it selects rather than for the league as a whole. Unweighted, those 548 players became regulars 19% of the time; weighting them to our own price mix lifts the comparison to 34%, which is the harder benchmark and the one used throughout.
Lists run to 2021 because every player is given five complete seasons. The 2022 and 2023 lists are not eligible yet.
Playing “at that level or above” means the 24 leagues our classification rates at Swiss Super League level or higher: the big five, the second tiers of England, Germany and Italy, and the top flights of the Netherlands, Portugal, Austria, Belgium, Scotland, Denmark, Norway, Poland, Czechia, Greece, Turkey and Russia, plus MLS, Brazil and Mexico. Minutes anywhere else do not count, including in the Swiss Challenge League.
Minutes are league minutes, totalled per player and then averaged across the group. A figure of 6,377 is one player's five-year total, on average: roughly 1,275 minutes a season. Measured August 2026.