Methodology

Sources

• Games, scores and schedules: the NCAA scoreboard (via a self-hosted mirror of the open ncaa-api). Every 2026 game is stored once; NCAA sometimes lists a postseason game twice under two ids, and the second listing is dropped.
• Team records: the NCAA's official team W-L. Conference-only records are not shown anywhere on the site, because no feed we read publishes them.
• Player stats: each D1 school's official 2026 statistics page (6,541 players, all 304 teams), postseason included. Rates are recomputed from totals. Each player card links its source page.
• Rankings: the ESPN.com/USA Softball Top 25 as published.
• Transfer portal: public reporting (articles, school announcements). An entry or commitment is listed only with a source link; statuses never move without a new source. Each player page shows its sources.
• Leader boards: built from the player stats above. Counting stats rank by season total; rate stats rank qualified players only. The NCAA's own published lists (on /leaders) can differ slightly where a school's final sheet and the NCAA disagree.

Update cadence

During the season (February to June) scores and records refresh several times a day and the boards follow. The 2026 season is final: those pages no longer change, and each says so with a “Final 2026 season” line and the date of the last game it covers. The transfer-portal tracker refreshes twice a day all year. The next NCAA D1 softball transfer window is Dec. 1–15, 2026.

The DiamondIQ Rating

DiamondIQ asks one question: how many runs would this team score, and allow, against an average Division I opponent on a neutral field? It answers it with a single ridge regression over all 7,634 decided D1-vs-D1 games of 2026: every half-game is a row, runs = league average + offense(batting team) + defense(fielding team) + home edge, all terms solved together. A run scored against Texas counts for more than a run scored against a twelve-win team, which is the point.

Since October 2026 the regression has two more pieces. Each conference gets its own offense and defense term, so a league that mostly plays itself is still measured against the rest of Division I. And each team starts from its rating of the previous season instead of from zero: the 2026 ratings are pulled toward each team's 2025 rating (from 2025 game results), and the season's games move them away from it as the evidence builds. Both changes were tested walk-forward on eight NCAA seasons (2017–2026, without 2020 and 2024), scoring every game only with ratings built from games played before it: together they raised game-by-game accuracy from 69.6% to 71.9% and improved log loss on NCAA tournament games, with confidence intervals that exclude zero. A defending-champion bonus, recency weighting and run-margin caps were tested too and rejected — they predicted games no better, or worse.

The rating is predictive, not a résumé. It doesn't know who won a trophy, so a champion can rank below a team it beat in June. The matchup predictor turns two teams' expected runs into a win probability with a logistic curve fitted walk-forward (each game scored only with ratings from games before it) on the 2025 season and checked on 2026: on 2026 games the favourite won 72.9% of the time (Brier 0.181, log loss 0.536). Model: 304 teams, league average 5.09 runs a game.

Terms

DiamondIQ Rating

An opponent-adjusted team rating: one ridge regression over every D1-vs-D1 half-game of the season, solving each team’s offense and run prevention together with conference terms, a home-field term and a pull toward the team’s previous-season rating. 50 is an average D1 team; the top team is scaled to 97.

Final season

The 2026 season is over and its numbers are frozen: regular season, conference tournaments and the NCAA tournament through the last game of the Women’s College World Series.

Qualified (hitters)

At least 2.0 plate appearances per team game, the NCAA minimum for rate-stat leader lists.

Qualified (pitchers)

At least 1.0 inning pitched per team game, the NCAA minimum.

Official record

The NCAA’s own team W-L, all opponents. Never a tally of scoreboard games, which misses some games.

Limits

• The NCAA scoreboard doesn't flag neutral sites, so “home” means the listed home team.
• Some schools publish a game the NCAA scoreboard leaves out (or the reverse), so a team's schedule here can be a game or two short of its official record. Records always use the official figure.
• Players listed only by jersey number on a school's stat sheet keep that label and aren't indexed.

Independent coverage. Not affiliated with the NCAA, USA Softball or any school.