Community project measuring TF2 Casual matchmaking quality

Fix TF2 Matchmaking is a community project focused on measuring how TF2 Casual matches are actually constructed and how they change over time.

The goal is simple: collect enough real match data to move the discussion beyond anecdotes.

We are currently tracking things such as team composition, public TF2 playtime, roster churn, backfill timing, team swaps, kill-rate divergence, and how long matches remain competitive before one side begins to collapse.

The project is also testing proposed changes against recorded matches, including better experience distribution, smarter backfill assignment, targeted team balancing, and retention incentives such as randomized completion-based XP bonuses.

This is not an attempt to create perfectly even matches. Competitive games will still swing, players will still leave, and some matches will always be stronger than others. The objective is to determine whether Casual matchmaking can keep matches competitive for longer and reduce unnecessary rolls.

Data is being collected from normal Casual play using TF2 demo files and publicly available Steam information. Private or unavailable player data is treated as unknown rather than estimated.

As the dataset grows, this group will be used to publish periodic statistics, methodology, findings, and eventually the tools used to collect and analyze the data.

This is an independent community project and is not affiliated with Valve.