All-VCT Awards

A fully data-driven, unofficial selection — no votes. For each season we score every player whose series count (distinct matches played) is above that season's median — capped so no more than 200 players qualify — then build the teams from those scores. The floor self-adjusts per season, so a season still in progress needs no special handling: its median reflects the matches played so far.

Coverage. China matches record fewer of the stat blocks the score draws on (65.7% carry opening duels, against 99.7% elsewhere), so a season played mostly there is scored on a thinner sample. Each component is averaged over the maps that carry it; one missing entirely is dropped and the remaining weights renormalised. Coverage by league →

• Performance score. A composite that blends per-map combat rating, KAST, K/D and opening-duel win rate (first kills as a share of the opening duels you take, so it rewards winning entries without taxing duelists for their first deaths). Rating and KAST carry the most weight. ACS is shown for reference but left out of the score — it’s largely redundant with rating and inflates in lost rounds. This is combined with team success (below) to rank the teams.
• OVR. The OVR badge is a FIFA-style 0–99 rating of that performance score alone — pure play, adjusted for role, with team strength left out (that’s the Elo column). It measures how well a player played, not how valuable their season was, so it won’t always count perfectly down a team — a player on a weaker team can out-OVR someone ranked above them. The scale is stable across seasons: a typical selection sits in the high 80s, and only a genuine once-a-season outlier reaches the mid-to-high 90s.
• Winning matters. Production alone misses the players who anchor deep-running teams. Each season, every player carries their own Elo, updated map by map (one global pool, so international results rank the regions against each other); each 5v5 result moves all five players against the opponent’s average Elo. Ratings start fresh every season, so each year is judged on its own. A great stat line on a team that wins beats the same line on a team that doesn’t.
• Big stages count more. Maps at international events (Masters, Champions, and the season-opening international LAN) are weighted 2× regional play — for both production and Elo. You don’t have to play internationally to make a team, but shining against the world’s best counts for more than padding a regional schedule.
• Roles compared fairly. Scores are normalized within each role, so a dominant Controller or Sentinel can outrank a merely-good Duelist on the overall teams instead of losing on raw fragging.
• Role & Flex. A player’s role is the one they played most — unless their maps are split across roles (no single role ≥60%), in which case they’re a Flex. The five All-Role teams use the same scores, grouped by role.
• Sustained > spiky. Scores are shrunk toward the role average by sample size, so a hot 30-map stretch can’t outrank a great full season.

Seasons marked “in progress” (★) are provisional and will keep moving until the season ends.

How percentiles and peer groups work