WHY THIS PREP
Built to be checkable
Most free ATSA practice gets the look roughly right. The harder problem is being correct: that the answer marked right is actually right, that scoring reflects the real objective, and that nothing is invented where the public record is silent. That is the bar this project holds itself to.
Verified answer keys
The spatial left and right key is unit-tested against fixed orientations, and the logic and reading answers are checked by a solver on every build. The keys are provably correct, not asserted.
Minimum-removal radar scoring
Radar computes the true smallest set of removals that clears each wave and rates your efficiency against it, rather than only counting saves and mistakes.
Deterministic replay
A seed reproduces an exact session. You can repeat the same wave order or question set to measure improvement against a fixed target.
Harder, honest math
The radar arithmetic distractor runs in larger digits and higher difficulty than lightweight practice math, closer to the pressure people report.
Full-length personality format
The personality module runs the reported length in the correct most and least forced-choice format, not a short sample, and it does not invent an FAA profile.
Answer review with explanations
Reading and logic show a post-test review that explains why each answer is correct, so a wrong answer becomes something you can learn from.
Eye-viewpoint accuracy, broken out
Spatial reports your accuracy on the eye-viewpoint items separately, so you can tell whether the perspective switch specifically is the weak point.
A real full-mock protocol
A full sitting respects the reported active ceiling and a separate break bank, which per-section drills alone do not rehearse.
What it will not do
It will not hand you a numeric ATSA score or a definite band, because the FAA does not publish the cut points or section weights. It will not present disputed timings as settled. Where candidate reports disagree, the matching setting is left configurable so you can practice both rather than trusting one guess.