Player Ratings — Current State & Improvement Resource

1. The Big Picture: Three Systems, One Rating

A player rating starts with a coach's observation and ends up balancing a draft. Between those two points it passes through three systems we run, plus — now — a fourth capability PlayMetrics has built natively. Here is the whole landscape at a glance:

2. Where Ratings Live Today

The automation pipeline stores every rating as a durable record in a player_rating table — the system of record. A rating is modeled as "an observation by a rater," so the same player can carry many ratings across seasons, divisions, and raters; nothing is overwritten.

What a rating record holds

• Overall — the single number the draft uses. Sub-scores are also kept: Technical, Tactical, Athletic, Personal, plus Team Rank (where the player ranked within their team).
• Provenance — who rated (the coach/source), the season it describes, the division, the player's prior team, a free-text comment, and the source system.
• The coach's raw scores — the original columns the coach filled (Speed, Kick, Ball Control, Passing, Aggressiveness, Defending, Attitude, Game Knowledge, Attendance, Rank in Team) are preserved verbatim. The four sub-scores are a roll-up of these; a few raw fields (e.g. Defending) survive only inside the weighted Overall.

The three ways ratings get in

Linking ratings to players

Ratings are matched to the registered player by AYSO ID first, then name + date of birth, then a guarded fuzzy name match (with a nickname table). A rating whose player can't be resolved is kept anyway — stored with the name/DOB so it can be re-linked later once the player appears. This is why a mistyped name on a coach sheet doesn't lose the rating, but it is also a source of manual cleanup.

3. How Ratings Drive the Draft

The pipeline runs a rating-balanced auto-draft, then hands Division Managers a kit to finalize by hand. The automatic pass does three things in order:

1. Pins each coach's own child to that coach's team (the kid rides the coach, and consumes that team's pick).
2. Keeps siblings together on one team — unless a parent asked to separate them.
3. Places everyone else strongest-first onto the lightest team, where "lightest" means the lowest total team rating. This drives every team's average rating toward the middle (about 5 in a 1–9 system).

A player with no rating is treated as a 5 (neutral) for the balancing math and flagged as unrated — so missing ratings don't break the draft, but they do dilute how well it balances.

What the Division Manager actually drafts on

The pipeline generates an Excel draft kit — a "Draft Pool" sheet with every player sorted strongest-first, split into older and younger birth-year cohorts, showing Rating, Rated By, Coach's-Kid-Of, Sibling-Of, and Comments, with an editable TEAM column (a dropdown of the DM's team names, each head coach's kid pre-filled onto their team). The DM drafts live in that column; it imports straight back as confirmed placements. The sheet header literally reads "Keep ratings CONFIDENTIAL." A second "Draft Order" tab shows the serpentine pick order.

4. What the Portal Does with Ratings

The Region 58 portal is a read-only consumer — it displays ratings but never originates or edits a score. It's where directors actually look at ratings between drafts. Everything here is gated to admins and directors, and a director only sees their own divisions; ratings are labeled confidential throughout.

• Ratings Worklist (/ratings) — every enrolled player in your division(s) with their latest rating and status (Rated / Unrated / Under review), plus bulk "email parents." This is the tool for seeing who still needs rating.
• Draft Board (/draft) — teams with each player's rating and the team average, plus a full rating-detail table (the sub-scores, rater, season, prior team, comment).
• "Under review" flag — the one thing the portal writes: a director can flag a player's rating for follow-up with a note. It records who flagged it and when. It never changes the score.
• Draft workbook & retention list — a downloadable per-division workbook joining roster to ratings, and a non-returning-player list sorted highest-rated-first (so the strongest departures surface for outreach).

5. PlayMetrics' Own Native Ratings — the Built-In Alternative

Since our pipeline was built, PlayMetrics has added native ratings and auto-draft, and — per their help center — made them available to AYSO regions for recreational core leagues. This matters because it overlaps a lot of what our custom pipeline does, in-app. We are not using it yet; this section is here so the comparison is on the table.

Player Tryout Ratings in a Team Assignment

• Turned on per team assignment (Enable Tryout Ratings), with a configurable 1–10 whole-number range.
• Admins and directors rate players right in the Players tab — a score plus a note — or in bulk by exporting the players to CSV, filling a rating column, and re-importing.
• PlayMetrics explicitly suggests importing prior ratings as a note (e.g. "Legacy Rating: 4") to carry rating history forward — the same idea our pipeline does with multi-season history.
• When assigning, you can filter by score (a specific rating or a range) and place accordingly.

Auto Assign / Auto Draft (Beta)

• Automatically forms balanced teams from the rated pool — prioritize by rating or age, choose standard or snake draft order, and it seeds lower-average-rating teams higher so rosters even out. This is essentially the same balancing goal as our run_draft.
• In-house leagues only; it is a beta that PlayMetrics must switch on (email success@playmetrics.com).
• Without ratings it falls back to balancing by age.

6. The Opposing-Roster Question (Steve Post)

At the board meeting Steve Post noted that in Sports Connect coaches could see the rosters of other teams, but in PlayMetrics they cannot — and that this bears on the ratings process. Here is what is actually true in PlayMetrics, verified against the help center and the product:

Why it matters for ratings — and why it's worth probing

Our rating model rests on coaches rating their own players, which the PlayMetrics limitation does not block. But cross-team visibility in Sports Connect plausibly supported the ratings process in ways worth naming before deciding whether its loss is a real problem:

• Calibration. Seeing other teams let a coach sanity-check their own 1–9 scale against players they'd watched elsewhere, so a "7" meant roughly the same thing across coaches.
• Validation / scouting. A coordinator or coach could eyeball whether the draft actually balanced, by seeing who landed where.
• Cross-team observation. Any rating of a player a coach only saw as an opponent now has no roster to attach to.

The honest framing for the board: PlayMetrics' model is "coach sees own team only," full stop — there is no hidden setting. If cross-team visibility is genuinely important to how Region 58 wants ratings to work, that is a feature request to PlayMetrics, not a configuration change. The question to settle first is what the visibility was actually used for, so the request (or the workaround) is precise.

7. Gaps & Friction in Today's Process

8. Recommendations & Questions to Probe

These are starting points, not decisions — framed as questions so the board can work them.