Practice scores
Practice mode is designed for learning a method. A module may report accuracy, completed items, response speed, control error or a composite measure depending on the task. Settings can change between runs, so two practice percentages are not automatically equivalent.
Use a practice result to answer a local question: did accuracy improve at the same difficulty, did tracking become steadier, or did a specific error happen less often?
Assessment scores
Assessment mode uses a more fixed workload so results are more suitable for comparison over time. The displayed percentage is calculated from the Academy module’s own scoring rules. Depending on the task, those rules can combine correct decisions with missed targets, false responses, timing or control precision.
An assessment score describes performance in CBAT Academy. It is not an official RAF or CAF test score and it is not converted using an official proprietary algorithm.
Estimated stanines
Where the interface shows a stanine-style value, CBAT Academy normalises its own practice result onto a nine-band training scale. This makes different modules easier to compare at a glance, but it does not mean the value was produced by an official assessment’s norm group.
The useful interpretation is relative: lower bands suggest a module should return to focused practice, while higher bands suggest the current Academy workload is becoming more reliable.
For Cognitive Updating, Target Recognition, Angles, Bearings & Degrees, Vigilance, Sensory Motor, Rapid Tracking, System Logic, Digit Recognition, Trace Test 1, Trace Test 2, Spatial Integration and Airborne Numerical, the bands are set from how CBAT Academy users score on their first standardized attempt, so a stanine there describes where a result sits among other CBAT Academy users. That group is not an official applicant norm group, and the bands are reviewed as more results come in.
System Logic, Digit Recognition and the two Trace Tests have only 8 to 16 questions in a standardized run, so each possible score takes the stanine that matches where it ranks among those first attempts, and a perfect score is always a 9. With so few possible scores, some stanines can’t be reached in those modules.
Standardized runs of Table Reading, Numerical Operations, Mathematics Reasoning and Instrument Comprehension score correct answers out of every item in the run, so an item left unanswered when time runs out counts as wrong. Practice runs still report accuracy over the items answered.
The Auditory Capacity Test counts what you leave undone as well as what you get wrong. A number order that isn’t answered within 10 seconds, a wrong number key, a number key pressed when none was ordered (including for an order meant for another call sign) and a press of the trigger when nothing beeped are all mistakes, and every avoid order, yours or another call sign’s, is followed by shapes it is about. Standardized runs use six digits. Results from before this change can’t be compared directly with the new ones, so they are marked as an earlier scoring version, and the stanine bands for this module stay the general ones until enough CBAT Academy users have taken the new version.
How modules map to cognitive domains
Each training module is assigned to one or more cognitive domains such as spatial reasoning, symbolic reasoning, task management, perception, attention or psychomotor performance. These mappings are independently developed and group module results into broader skill areas.
Role requirement comparisons
Where public material describes a role, domain structure or historical requirement, the interface can place Academy estimates beside that reference. Public information is used only as context. Gaps are not silently filled with official-looking claims: module mappings, trainer weights and readiness calculations remain Academy methodology.
A comparison can help identify which domain deserves more practice. It cannot determine eligibility, predict a selection outcome or replace current advice from a recruiter.
How score history is used
History keeps individual module attempts so you can compare like with like, see whether a change persists and avoid overreacting to one unusually strong or weak run. Broader averages use eligible assessment attempts rather than treating every adjustable practice run as equal evidence.
Trend direction, workload and recurring error type are usually more useful than a single best percentage.
Why every estimate is unofficial
- Official scoring algorithms, current norm groups and complete role thresholds are not public.
- Academy exercises differ in wording, timing, controls, display design and workload.
- Browser performance, screen size, input hardware, fatigue and chosen settings affect results.
- Historical candidate material can become outdated and is labelled separately from official sources.
- Selection includes factors beyond aptitude testing.
Review and correction process
Scoring documentation is reviewed when a module’s rules, result format or domain mapping changes. Public-source claims are checked separately from Academy calculations. If a source does not establish a claim, the wording is narrowed or the material is labelled as historical, candidate-reported or independently inferred.