Most lab scoring asks one question at the end: does the finished thing work? It sounds fair, but it hides almost everything that matters.
Two people can end a lab with the same broken deployment. One got nine of ten steps right and ran out of time. The other skipped the hard parts. An end-state check gives them the same score.
Score the journey
On Gravity Labs, every lab is a journey of milestones, and every milestone is a set of tasks. Each task is checked in the live lab environment as the learner goes: the resource exists, the setting is right, the service answers.
That gives three things an end-state check cannot:
- Progress as you go. Learners see which tasks are done and which are not, so they fix mistakes while they still remember making them.
- A fair leaderboard. Partial work earns partial credit, so a strong attempt that ran out of time still ranks above a skipped one.
- Real insight for organisers. When most of the room stalls on the same task, that task needs a better explanation, not a harder test.
Checked where the work happens
Task checks run against the actual lab, not against what the learner says they did. There is no screenshot to upload and no form to fill in. If the work is there, it scores.
Built for events and courses
The same scoring powers a one-day challenge with a live leaderboard and a multi-week course with graded labs. Organisers choose the weighting of each milestone, and learners always see exactly what each task is worth.
Scoring should reward the work people actually do. Checking every task is how you get there.
