Global
rankings.

The only board in the Office World Cup that put whole companies against each other. Inside a workspace colleagues raced colleagues; here an office raced the rest of the world.

A qualifying squad was five World Cup predictors

Five World Cup predictors was the minimum to be ranked. Below that, one lucky forecaster carries the whole office and the table stops measuring anything. Of the 269 companies that played in 2026, 143 fielded a squad that size, and 65 of those chose to appear here. Nothing was automatic: the company board was opt-in, and every player chose separately whether their own name could show on the players tab.

Average points per World Cup predictor, not total

A total points board is a headcount board: a 500 person company beats a 10 person company without a single good prediction. Averaging takes headcount out of the contest, which is how a small office won the whole thing. Custo took the 2026 title with seven players on 364.6 average points, close to double the tournament average, and the frozen top ten is in the 2026 Hall of Fame. The board above currently has Custo on top.

Why the order kept moving

Every settled match rewrote the averages. A correct result was worth 3 points, an exact 90 minute scoreline 9, and calling an upset added 2 more. The stage multiplier did the rest: group games counted once, the round of 32 and round of 16 at 1.5x, quarter finals at 2x, semi finals and the third place play off at 2.5x, and the final at 3x. One knockout evening outweighed several group match days, so an office that coasted through the groups could still be caught in July. The two tournament picks, 30 points for the winner and 30 for the Golden Boot, settled at the end and moved the table one last time.

What the page cannot show is the history. We kept no public archive of every position change, so there is no movers table here. The ranking rule itself never changed, so the order is directly comparable from the first group game to the final.

What it actually meant

A cross-company board gives an office pool an outside opponent, which no internal spreadsheet can. That idea now runs for the NFL, where consenting companies are ranked on their weekly pick accuracy at the NFL company leaderboard. Season long league pools have no cross-company table on purpose: companies join at any point in a 38 gameweek season, and a cumulative table would punish everyone who joined late.