Most people have taste. Few people can write it down. The gap between the two is where this project lives. By the end you'll have a half-page rubric that says, in writing, what makes work good in your chosen domain. It will surprise you.
This is the first Harness Studio project — the one before any system, before any judge. The finale (project 08) will only ever be as good as the rubric you write today. So write a real one.
The first rubric you write will be wrong. That's fine. Writing it down is what reveals where it's wrong. You will rewrite it twice before it's good. The rewrites are the project.
Step by step
-
Pick the domain. Be sharp about it.
Not "good writing." Pick "my weekly journal entry, the one I write every Sunday night." Or "the captions I write for my photo album." Or "my Slack message that I want my friend to actually reply to." The narrower the domain, the more honest your rubric will be.
-
Pick three dimensions.
Three — not two (too vague), not five (overlapping). Each dimension is one sharp thing the work either has or doesn't. Each must be checkable by reading the work, not by reading your mind.
Examples for a journal entry: specific (does it name real things?), honest (does it admit the awkward part?), arc (does the writer change between the first and last sentence?).
-
Write anchors at 1, 3, and 5 for each dimension.
Anchors are the calibration. A "5" on specific in your rubric should be unambiguous. Same for "1." This is where most rubrics die — people write the dimensions and skip the anchors, and then no one (including them) can score anything consistently.
-
Score three real examples by hand.
Find three real examples in your domain. Don't ask Claude to make them up. Use real ones — your own past work, a sample online, a friend's draft. Score each against your three dimensions. Add up the score (out of 15). Note your gut feeling separately.
-
Notice the surprise.
Almost always, one example you thought was good will score lower than your gut said. That's the project. Decide who's right — your rubric or your gut. If the rubric is right, you've found a place where your eye was lazy. If your gut is right, your rubric is missing a dimension. Add it. Re-score.
-
Write a one-paragraph reflection.
What surprised you about your own taste once you had to defend it on paper? Be honest. Most kids write something like: "I had been calling things good when they didn't actually score well on my own rubric." That sentence is the goal of this project.
A complete worked example
The rubric for "my weekly Sunday journal entry." Three dimensions, full anchors, three real journal entries scored.
domain: |
My weekly journal entry. Sunday night, ~200 words.
About things that actually happened, not things I wish.
dimensions:
## 1. specific
asks: "Does the entry name real things — a person, a place,
a smell, a sentence someone said?"
1: All abstractions. "I had a busy week. I learned a lot."
No moments, no people named, no specifics. Could
have been written by anyone, about any week.
3: One specific moment, but treated as decoration. The
specific thing is mentioned and then the entry returns
to abstractions.
5: Multiple specific things, and the entry is BUILT on
them. Names, places, exact sentences. Take them away
and the entry collapses.
## 2. honest
asks: "Does the entry admit the awkward part — what was
hard, what didn't go the way I wanted?"
1: Pure highlight reel. Everything went great. Nothing
was hard. (No journal entry is actually like this —
this score means the writer is editing.)
3: One mention of something hard, but quickly waved
away. "It was tough but I learned!"
5: Names a specific thing that didn't go right, sits
with it for a sentence or two, doesn't pretend to
have figured it out yet. The entry is more useful
because of the honesty.
## 3. arc
asks: "Does the writer change between the first sentence
and the last? Does the change feel earned?"
1: Static. The writer at the end is the writer at the
beginning. No change.
3: Small change. The writer notices something at the
end. The notice feels added, not earned.
5: Real change. By the last sentence, the writer is
thinking something they couldn't have at the start.
The arc was hidden in the specifics.
scoring:
scale: 1-5 per dimension
total: 3 - 15
ship_threshold: >= 11 (and no dim below 3)
## entry 1 — Sun Apr 13 (the one i thought was good)
excerpt: "Had a busy week. School was a lot. I felt
tired. Made some progress on my Skill project.
Looking forward to next week."
scored:
specific: 1 # zero specifics — "school", "felt"
honest: 2 # admits "tired" but waves it away
arc: 1 # no change, no notice
total: 4 / 15
my gut before: "this was a fine entry"
surprise: |
My gut was wrong. The entry is forgettable. The rubric
catches what my eye missed: I wasn't writing about
anything. I was just confirming I existed for the week.
## entry 2 — Sun Apr 20
excerpt: "Mr Lee asked me on Tuesday why I always sit
in the back of math class. I said I don't
know. Walking home I realized I do know — I
don't want him to call on me when I'm not
sure. Sat in the front today. He didn't call
on me anyway."
scored:
specific: 5 # Mr Lee, Tuesday, math, the back, the front
honest: 5 # admits the discomfort, doesn't resolve it
arc: 5 # writer changed seats. real action from the realization.
total: 15 / 15
my gut before: "good entry"
surprise: |
No surprise. The rubric agreed. Specific + honest + arc.
## entry 3 — Sun Apr 27
excerpt: "Played basketball Friday. We lost. Coach
said we need to communicate more. I think
she's right. Will try harder next week."
scored:
specific: 2 # Friday, basketball — but "communicate more" is generic
honest: 2 # admits the loss, but no specific awkward moment
arc: 2 # claims an arc ("will try harder") but doesn't show it
total: 6 / 15
my gut before: "this was thoughtful"
surprise: |
Halfway between my gut and the rubric. The entry sounds
thoughtful but doesn't DO any of the work: no specific
teammate named, no specific moment of bad communication,
no actual change. The rubric caught what felt-thoughtful
but wasn't.
## what surprised me about my own taste I thought my journal entries were mostly good. The rubric disagreed about two of three. The pattern: my entries sounded thoughtful when I read them — they had words like "learned" and "important" — but they weren't doing any of the actual work my rubric says good entries do (specific, honest, arc). The biggest realization: my eye was forgiving entries that SOUNDED reflective without actually being reflective. The words "I learned a lot" appeared in two of three entries. That's a tell. Real reflection is concrete: something specific changed because of something specific that happened. Going forward: every Sunday, before I close the entry, I re-read it once and ask — does this score above 11? If not, either I had a forgettable week (fine, write that honestly) or I'm hiding behind generic words. This rubric is now taped inside the front cover of my journal.
Live demo 1: score a piece of work right here
Paste a piece of work. Score it on the three dimensions. The widget computes your total and flags any dimension below 3.
Hand scorer (1–5 per dimension)
Live demo 2: write an anchor right now
Pick a dimension. Write what a 1, a 3, and a 5 look like. The widget checks: are your anchors specific? Are they testable by reading the work, not by reading the writer's mind?
Anchor writer
What makes this hard
The hardest move is scoring something you like badly. You'll have a piece of work — maybe one you wrote — that you thought was good. You'll score it 5/15. You will be tempted to fudge it up to 10. Don't. The rubric is a tool to find out where your eye disagrees with your written taste. The disagreement is the gold.
The second hard thing is anchors. Writing a "5" anchor feels easy. Writing a "1" anchor feels mean. Skipping the "1" is the most common mistake — and it's exactly the anchor that lets you tell mediocrity from real bad.
Self-check before you ship
- Domain sentence is narrow enough that an adult couldn't have generated it from scratch.
- Three dimensions, no overlaps; each one asks one sharp question.
- Anchors at 1, 3, AND 5 for every dimension — actually written, not implied.
- Three real (not generated) examples scored by hand.
- I noted at least one example where my gut and the rubric disagreed, and decided who was right.
- Reflection paragraph names at least one specific thing my eye had been calling good that the rubric flagged as not.
Try it · once your rubric is on paper
A 3-dimension rubric is the floor. Want to push?
- Get a friend to score the same 3 examples. Hand them your rubric. Have them score independently. Compare. If you both gave the same example a 12, your rubric is doing real work. If you gave it 12 and they gave it 6, the anchors aren't sharp enough.
- Add a 4th dimension after a week. Live with the 3-dimension rubric for 7 days. Notice if anything keeps slipping through that you wish the rubric caught. Add it as a 4th dimension. Real rubrics grow this way.
- Use the rubric in a different domain. Pick a 2nd domain — same 3 dimensions or different ones, your call. Score 3 things. Notice what transfers and what doesn't. Some dimensions (like "specific") are universal. Some are domain-bound.