An adaptive AMC question bank may help organise revision, but it does not automatically reproduce the real computer-adaptive test. Adaptive learning selects useful study activities; adaptive testing supports measurement under an examination design. A platform can do the first without demonstrating the second, and neither label alone establishes a valid prediction of passing.
The AMC's MCQ examination information, checked on 24 September 2026, describes the examination as computer-adaptive. This article is published by iatroX and includes iatroX alongside Amedex and eMedici, applying the same limits to all three rather than treating adaptive marketing as a unique advantage.
Different systems can adapt for different reasons
A revision system might choose your next question because you recently missed the topic, because the concept is due for review or because the selected material is suitable for your current learning stage. Its aim may be to support learning, even if the resulting session is deliberately unrepresentative of an examination paper.
An adaptive examination has a different purpose: measuring performance within its assessment design. The examining body's item-selection and scoring methods cannot be reconstructed merely from the fact that a commercial platform changes the next question after an answer.
A tutor conversation is another form of responsiveness. Asking a follow-up about your misunderstanding may be educationally useful without being either a computer-adaptive test or an item-calibrated assessment.
A terminology table that prevents overclaiming
| Term | What it should describe | What evidence is still needed |
|---|---|---|
| Adaptive revision | Study selection changes using information about the learner | What inputs are used, what changes and how breadth is maintained |
| Spaced review | Earlier material returns after an interval or under a review schedule | How the schedule works and how the learner can inspect the queue |
| Difficulty adjustment | The selected difficulty changes according to a rule | How difficulty is defined and whether the rule has been evaluated |
| CAT simulation | An attempt to reproduce aspects of adaptive examination delivery | Which aspects match the official test and which are approximations |
| Mock assessment | A practice assessment under specified conditions | Its content selection, exposure history, timing and intended interpretation |
| Score prediction | An estimate of a future examination outcome | Appropriate outcome-linked validation and uncertainty reporting |
These are definitions for evaluating claims, not statements that every named provider offers every function. A custom topic filter, for example, can be useful without being described as an adaptive algorithm.
What can be established about the named resources?
Amedex and eMedici offer commercial preparation resources, as reviewed on 24 September 2026. Their availability does not establish that either reproduces AMC's item-selection or scoring model. This review did not verify such equivalence, and does not assign an adaptive function to a provider merely because the article discusses adaptivity.
The official AMC-eMedici preparation resource has a separately documented AMC relationship. That provenance should not be extended to eMedici's entire commercial catalogue or converted into a claim that every practice result is an official score estimate.
Per iatroX product information, September 2026, its question banks use adaptive sequencing and spaced repetition. Those are revision design features. They are not evidence that iatroX reproduces the AMC CAT or can determine a candidate's probability of passing.
Why a lower adaptive-session percentage may not mean deterioration
Consider an original fictional candidate, Imran. He completes a broad introductory session, then works through a queue deliberately concentrated on concepts he found difficult. His next percentage is lower.
One possible explanation is that the second session sampled harder material for him. Another is that the original gaps remain unresolved. A third is that the sessions were conducted under different conditions. The percentage alone cannot distinguish these possibilities.
Ask what changed in the selection process before drawing a conclusion about ability. If the platform explains why an item was recommended, use that information. If the selection rules are not visible, keep the interpretation appropriately narrow: this was performance on a particular selected set, not necessarily a comparable examination of the same construct.
A rising percentage requires the same scrutiny. Repeated items, narrower topics or easier selection can improve a result without establishing broader transfer.
What a mock percentage cannot establish
A raw commercial percentage does not reveal the official examination's scaled result. Nor does a colourful readiness indicator establish predictive validity. The provider needs evidence that connects the measure to the outcome it claims to predict.
For a defensible prediction claim, useful information would include the candidate population studied, the dates of data collection, how practice exposure was recorded, which examination outcomes were obtained and how the model performed on learners not used to develop it. The report should explain uncertainty, missing outcomes and whether the result remains applicable after product or examination changes.
These are methodological requirements proposed for evaluating a claim, not a statement that any named provider has or has not completed an unpublished study. A provider's inability to display every technical detail does not prove its learning resource is ineffective; it limits the conclusion a buyer can draw about prediction.
Inspect the adaptation rather than the adjective
During a trial, create a simple record of what you attempted, what you found difficult and what appeared next. Ask whether the next activity addresses the same concept, an adjacent prerequisite or simply another item in the topic.
Then look for control. Can you still access mixed practice? Can you inspect areas that the system has not recently selected? Can you deliberately revisit a correct answer that you guessed? An automated recommendation is more useful when it does not prevent an informed learner from checking its blind spots.
Do not intentionally fail large numbers of questions and then claim to have validated the algorithm. An informal probe may show that the interface responds, but it cannot establish the educational quality of the underlying selection policy.
Rehearse decisions that remain useful without algorithmic equivalence
You can practise reading the lead-in accurately, selecting the best-supported answer, managing time and continuing after a difficult item without possessing the official CAT algorithm. These are legitimate preparation activities in their own right.
Use unfamiliar mixed material periodically so that your study plan does not consist only of repeatedly targeted weaknesses. Follow the current official instructions on examination navigation rather than assuming that a commercial review or backtracking function is available on the day.
After a difficult question, explain the uncertainty before reading the answer. Was the obstacle missing knowledge, conflicting interpretations or discomfort with not knowing immediately? Different problems require different follow-up. A dashboard label such as "weak" is a starting signal, not a full educational diagnosis.
Keep the study record honest
A useful record separates supported learning, repeated retrieval and unfamiliar assessment. Note whether the session was topic-targeted, adaptive or manually selected; whether explanations were visible; and whether the questions were previously encountered.
For iatroX, a Tutor-assisted answer should not be recorded as equivalent to an unaided decision. The same applies to hints, reference access or explanations used on any other platform. Assistance is not a failure; concealing it makes progress harder to interpret.
Ask-iatroX and free question access are genuinely free without trial expiry or a verification gate, per iatroX information for September 2026. That permits exploration of a learning workflow without requiring a claim that a free or paid session reproduces the examination.
The decision by candidate need
Choose a resource for adaptive revision when its recommendations help you study relevant weaknesses while preserving breadth. Choose a mock for a defined assessment purpose and interpret it under its stated conditions. Use official AMC material for official familiarisation.
Amedex or eMedici may suit a candidate seeking their particular content and learning presentation. iatroX may suit someone who benefits from question-linked tutoring and subsequent review. None should be selected on an assumed equivalence between the word "adaptive" and the real CAT. Useful learning is a sufficient proposition without an unsupported prediction claim.
Frequently asked questions
Does adaptive revision use the same algorithm as the AMC examination?
That equivalence is not established by the adaptive label. A provider would need to document and substantiate which examination processes it reproduces.
Can my score fall while I am learning effectively?
It can fall when the selected questions, conditions or exposure history change. Inspect those factors and your ability to apply the reasoning before interpreting the result.
Is an ordinary mixed question bank useless for a CAT examination?
No. It can support knowledge, discrimination and time-management practice without claiming to reproduce the official adaptive assessment model.