How scoring works
When the CAT test ends, I freeze two things per domain: a score (0–100) and a level. Both come straight from the evaluation — no separate calculation, no margin for spin.
Per-domain confidence
Every domain on the certification starts at neutral confidence. As you answer questions in that domain:
- A correct answer increases my confidence in the domain.
- A wrong answer decreases it more sharply than a correct one increases it (I treat wrongs as stronger evidence than rights — overconfidence is the classic failure mode I am trying to surface).
- The score is clamped to 0–100.
If I never asked you a question in a particular domain — for instance because the certification has many domains and CAT stopped early — that domain stays untouched. Untouched domains contribute zero to your overall score, not a neutral 50. I refuse to pretend I have signal I do not have.
The four levels
Each domain's final score maps to one of four levels:
| Score range | Level |
|---|---|
| 0 – 30 | Novice |
| 31 – 55 | Developing |
| 56 – 75 | Competent |
| 76 – 100 | Proficient |
Untouched domains are reported as Novice by default — not because you are weak there, but because I have no evidence to claim otherwise.
Overall score
Your overall score is the average of all per-domain scores, including the zeroes from untouched domains. This is intentional. An overall score that ignored untouched domains would inflate against the actual breadth of the certification.
If your overall score looks lower than you expected, check whether several domains are at zero. That usually means CAT terminated early on a few high-confidence ones and never reached the rest — your real overall is somewhere between what is shown and a hypothetical "all domains tested" version.
Why the level matters more than the number
The level is what I use to shape your roadmap. Two users can have the same overall score and very different roadmaps, because the per-domain levels are not the same.
- Novice in a domain → I give you the most milestones for that domain. You need foundations before applied work.
- Developing → fewer milestones, focused on closing specific gaps.
- Competent → fewer still, mostly applied practice and one validation.
- Proficient → minimal — usually a single check-in to confirm.
This is why a 60% overall on one cert can produce a longer roadmap than 60% on another. It is also why retaking the evaluation produces a different roadmap if your domain mix shifted, even when the overall is similar.
What is locked in once scoring runs
Once CAT stops, I write your domain scores, domain levels, overall score, and the CAT confidence number to your evaluation session. That snapshot is the input to 
ARIA's roadmap generator on the next screen — and the baseline I compare your readiness against from day one.