Meta Awards Every Game of the Year award, counted
Menu

Home › Statistics › Closest races

The closest Game of the Year races ever

Most years have a clear winner. A few were decided by a margin so thin that a handful of outlets voting differently would have rewritten the record. The tightest was 2007, where BioShock finished just 2 awards ahead of Super Mario Galaxy — a lead of 4.2%.

Sorted by relative margin, because an eight-award lead means something different in a 38-award year than in a 643-award one.

How close is close?

A margin needs a yardstick, and the spread across 23 complete seasons gives one. The median winner finishes 55.6% clear of second place — so the typical Game of the Year is not a photo finish, it is a comfortable win with a visible challenger.

  • 2Seasons decided by under 10%
  • 55.6%Median winning margin
  • 12Seasons won by more than half
  • 4.2%The tightest, in 2007

The two ends behave differently. A blowout says the press converged: one game was so far ahead that national taste, platform loyalty and house style all pointed the same way. A narrow margin almost never means the two games were judged equal — it means the season had two distinct constituencies, and which country you read usually predicted which one you were in.

What it would have taken to flip them

An award margin converts directly into outlets. Because one outlet's change of mind moves the gap by two — one away from the winner, one towards the challenger — a season flips when half the margin, rounded up, changes its verdict.

  1. 2007: BioShock over Super Mario Galaxy1
  2. 2010: Red Dead Redemption over Mass Effect 24
  3. 2006: The Elder Scrolls IV: Oblivion over Gears of War4
  4. 2003: Star Wars: Knights of the Old Republic over The Legend of Zelda: The Wind Waker5
  5. 2004: Half-Life 2 over Halo 27

This is arithmetic, not a counterfactual. Outlets do not vote against a field they can see, and nothing here says these seasons were nearly-run things in the minds of the people deciding them. What it does show is how little separates a result that gets written into history from one that does not — in 2007, 1 editorial decisions.

Updated 2 October 2026 · Source: the Meta Awards database, where every award row cites where it was published.