Weekly feedback on every section of every practice test
Your teams grade each other's practice tests against the official checklist, and the AI grades every test too. Scores and comments reach every student before the next practice.
Already on your club's roster? Sign in
Grading a stack of tests takes hours
A test has thirteen sections, every team takes one each Saturday, and every line is graded against the same checklist.
Students need feedback more often than one coach can give it
Students fix a weak section by hearing about it while the test is still fresh, week after week.
The screens below are the app's own, filled with sample data. The teams and students in them are made up.
Students grade the sections they own
Each student owns a few of the thirteen sections and grades just those on other teams' tests, so nobody grades a whole test.
Maya R.
Owen T.
Nina C.
Each student grades four or five sections, on other teams' tests.
Every section comes back scored and explained
Each grading team scores every checklist line and writes what went well and what to fix. Your students see them as Grading team 1 and Grading team 2.
HGraph
Grading team 17/12
Clear title and labels. The x axis jumps from 10 to 30, so the steps are uneven.
Grading team 28/12
Good graph type for this data. Add units to the y axis label.
The AI grades every test as it lands
Upload the scans after practice and the AI grades the whole test in a couple of minutes. Its scorecard sits beside the students' and is never merged into one score.
- BayGraded by AI71/107
- SierraGraded by AI64/107
The AI's scorecard lands about two minutes after the upload.
The AI reads every section too
Its feedback names what went well and what to fix, section by section, often with a sentence the student can model.
KAnalysis (CER)
- AI · what went well
- Your evidence quotes the averages from your own data table.
- AI · what to fix
- The claim repeats the question. Say what the data shows, for example: “As the drop height increased, the bounce height increased.”
You read one page of results
A practice plan ranks the sections where your teams are losing the most points, so Saturday starts with the one worth the most.
- 1HGraph14 points on the table3 of 3 teams need help
- 2KAnalysis (CER)9 points on the table2 of 3 teams need help
Weak sections come back as corrections
Once scores are out, each student rewrites any section under 90 percent, typed or as a photo of the page, and gets a practice score back.
HGraph
- On the test
- 7/12
- Practice score on the rewriteHanded in
- 11/12
Then it all runs again the next Saturday.
One section, eight Saturdays
A sample season, not a real student. The Graph section gets graded and commented on every week, and each rise follows the comment before it.
Week 8 of 8
11/12
Sample data, not a real student · Section H, Graph, 12 points · each week averages the peer, coach and AI scorecards
What the graders wrote
Wk 14/12Grading team 2
There is a graph, but it has no title and neither axis has a label or units.
Wk 25/12AI
Title and labels are in. The y axis still needs units, and the scale only fills the bottom third of the grid.
Wk 35/12Coach
Units on both axes now. Pick a scale that uses most of the grid so the trend is easy to see.
Wk 47/12Grading team 1
Much easier to read. The steps on the x axis jump unevenly between 10 and 30.
Wk 58/12AI
Even scale. Plot the averages you calculated, not every trial, so the trend stands out.
Wk 69/12Grading team 2
Averages plotted. You drew two conditions, so add a key that says which line is which.
Wk 710/12Coach
Key added. Make the title name both variables, not only the topic.
Wk 811/12Grading team 1
Clear title, labels, units, scale and key. One point left: a line graph suits this data better than bars.
Join the waitlist
I built this for the Division B club I help coach in Albany, CA, and I'm opening it to more clubs a few at a time. Leave your details and I'll write to you when there's a spot for yours.
It grades against the Division B checklist today. If you coach Division C, say so in the form; that's how I'll know to build it.
David Martin · Experimental Design coach