Resolve Research
Resolve Projections futbol · club Elo
snapshot loading…

Resolve Club Elo

Our own club Elo ranking — built from 28,967 matches across the top 5 European leagues × 16 seasons (K=32). The 5 era snapshots below — 2010 / 2014 / 2018 / 2022 / 2026 — read correctly era-by-era: Barca/Real on top in 2010, Bayern/Juve/Barca/PSG/City in 2014, City/Liverpool/Inter in 2022, PSG/Inter/Barca/Bayern/Real today.

This is the same model that produces our v3 calibrated WC picks — for each historical World Cup we counted how many squad members had a club in our top-20 at tournament start, then fit a one-feature linear model on outcome ordinal. Robustness retest: 4 of 5 N values BEAT v2 on champion hits.

matches28,967 · leagues5 · seasons16 · K32 · snapshots

Current Top Clubs

snapshot · click any column to re-sort

Live ranking as of the most recent snapshot. Switch eras via the dropdown to see who held the top spots in past tournament-start windows.

·
loading clubs…

Era Top 5 — How the Throne Shifted

2010 / 2014 / 2018 / 2022 / now

Five snapshot dates spaced four years apart — each pinned to the May/June window before a World Cup. Reads era-correct without us hand-tuning anything: Barca + Real owned 2010, Bayern + Juve + PSG + City emerged through 2014-2018, Liverpool + City peaked into 2022, PSG is the current #1 with Inter / Barca / Bayern / Real bunched behind.

loading eras…

Methodology

k=32 elo · 5 leagues · 16 seasons

Inputs:

  • 5 leagues: Premier League · La Liga · Bundesliga · Serie A · Ligue 1
  • 16 seasons: 2010-11 through 2025-26 (matches league transfers between snapshots)
  • 28,967 matches total — all available league matches plus relevant European-competition fixtures for cross-league calibration
  • K=32 standard Elo, initialized at 1500, snapshot taken at May-end / June-start each tournament year

Why this matters as a public surface: public Elo aggregators exist (Club Elo, FiveThirtyEight legacy) but we needed our own to drive the WC v3 calibrated picks honestly. By publishing the snapshot board we can be checked: if our 2018 top-5 looks insane, the v3 model is suspect. The 2010-26 progression tracks consensus club strength fairly closely, which is the bar.

Robustness retest: we re-fit the v3 lift formula using our own top-N (N ∈ {5, 10, 15, 20, 30}) and re-scored on the 4 historical WCs. 4 of 5 N values beat the v2 baseline on champion hits (3/4 vs 2/4). N=20 and N=30 are the cleanest configurations; N=20 ships as v3.

What this is NOT: not a player-level rating, not a transfer-market value, not a betting price. It's a per-club strength signal that aggregates match results into a single moving average. Use it as one input to club-comparison or national-team-squad-quality questions, not as a final say.