01 · The scoring arc
Rubric
the scoring sheet, done right
A short list of the specific skills you’ve decided to judge, each rated 1–5 on its own. Skills are judged deliberately, one at a time — so a dazzling overall impression can’t quietly stand in for the things nobody actually watched.
ExampleA "mite skater" rubric with separate criteria for crossovers, stops, edge control, and game sense — each 1–5, each scored on its own.
Criterion
one thing, scored on its own
A single line on the rubric: one specific skill, judged by itself. Because each stands alone, the quiet, complete performer never disappears into someone else’s spotlight — and presence only counts in the column where you decided it counts.
Example"Stops" is one criterion. A player weak on crossovers but flawless at stops keeps that 5 — it isn't averaged into invisibility.
Score
one number, one criterion
A single number, 1 to 5, from one person, about one specific skill — not a game score, and not the final verdict. Waza’s results are built from many of these small, separate judgments.
ExampleDan gives crossovers a 3. One evaluator, one criterion, one score — the smallest unit of judgment in Waza.
Blind (sealed) scoring
no peeking, no anchoring
Everyone judges on their own, and no one sees anyone else’s numbers until all of them are in. Copying the room never gets a chance to start — and it isn’t a setting you can switch off; it’s how the mechanism works.
ExampleThree coaches each submit before any score is revealed. Nobody thinks "Dan gave a 4, so I will too."
Weighted consensus
not a flat average
The group’s combined result when some voices are meant to count more than others — a head coach more than a first-year assistant. Not a flat average: the math respects who was trusted with more.
ExampleA head judge's score carries more weight than a trainee's, so the consensus reflects the panel's real structure.
Disagreement analysis
where the argument lives
When the group’s numbers split on the same skill, Waza points at the split instead of hiding it inside an average. The one person who saw what everyone else missed stays visible — the conversation a clipboard can’t give you.
ExampleTwo coaches score "game sense" a 4; one scores it a 2. Waza surfaces that split so the panel discusses the real disagreement.
Consensus
agreement you can defend
The end state: everyone’s independent judgments, combined and reconciled, with every point of contention left visible. Dan watched the crossovers, Sam watched the stops — same skill, different numbers, both right. That’s consensus.
ExampleThe event closes with a weighted score of 4.1 and one flagged criterion — a defensible result, not a number nobody can explain.
Checkpoint
the same look, later
The same evaluation, run again later — beginning, middle, and end of a season. Because the measuring stick doesn’t move, the change you see is the player’s, not the yardstick’s. This is what makes Waza longitudinal.
ExampleSeptember, January, March — three checkpoints on one rubric, one clear line of growth.
02 · People & structure
Subject
whatever you're evaluating
Anything evaluable — a person, a routine, a dish, a proposal, a place. Waza is type-agnostic: "people" is just the person-shaped case of a general subject.
ExampleThe same platform scores a hockey player, a thesis defense, and a plated dish — each is a subject.
Evaluator (Principal)
the one holding the pen
The person holding the pen — coach, judge, reviewer, committee member. What they’re allowed to see and do belongs to them and their organization, and never leaks onto the people being judged or the roster.
ExampleA coach can submit evaluations for their own org's events, but sees nothing outside that scope.
Collection
the roster / the group
The group being looked at together — the team, the cohort, the audition list. It’s how you follow the same set of people (or entries) across a whole season, beginning to end.
ExampleThe U10 roster is one collection, scored at tryouts, mid-season, and finals against the same rubric.
Domain
the world it lives in
The evaluation context that shapes the vocabulary and defaults — youth hockey, figure skating, clinical skills, hiring. The engine is the same; the domain gives it its language.
ExampleIn the clinical domain a "subject" is a candidate's skill demonstration; in hockey it's a skater.
Marketplace
rubrics you don't have to invent
A library of proven rubrics from people who’ve run this before. Clone one and shape it — heavier here, lighter there — instead of starting from a blank page.
ExamplePull a published "figure skating short program" rubric, then adjust the weighting to match your judging panel.
Agreement
the rules everyone signs
A versioned policy — consent, conduct, safety — that members accept before they participate. The platform sets a minimum enforcement level for each agreement; an organization can strengthen it, never weaken it.
ExampleA code-of-conduct agreement: the platform baseline says "must accept" — a youth league can raise the bar, but no one can lower it.
03 · Platform & trust
PWA
an app without the app store
A website that installs on your phone or laptop and behaves like a regular app — its own icon, opens full screen, keeps working when the connection drops. Nothing to download from a store, nothing to update by hand.
ExampleWi-Fi drops at the venue mid-tryout. Waza keeps taking scores offline, then syncs the moment the connection returns.
PII
the details that identify a person
Names, birthdays, addresses, faces, voices — any detail that points back to a real person. Waza screens for it the moment it enters the platform, so it never spreads further than it should.
ExampleA recorded comment mentions a player's full name and school. Waza flags it before the audio reaches the panel.
HIPAA
the US health-privacy law
The American law protecting medical information — the reason your doctor's office is careful about what it shares. Where health details may come up, Waza keeps everything inside that boundary: processed on your own hardware, never sent out.
ExampleA clinical-skills session is recorded. Transcription runs on-site, so nothing covered by health-privacy rules leaves the building.
SAST
software that checks the software
An automated review that reads the platform's own code looking for security mistakes — before that code ever ships. It runs on every change, alongside checks for risky dependencies and accidentally exposed secrets.
ExampleA change accidentally includes a password. The scan catches it before the change goes live.