TestBuddy is an adaptive Italian preparation platform for the Concorso SSM: a large question base with an AI assistant, an "intelligent calendar" that reshapes your schedule, and a scoring signal that estimates your level per subject. Adaptivity is powerful, but it has one hazard — a feed tuned to keep your overall score rising can quietly stop showing you the syllabus areas you avoid. This workflow is for Italian medical graduates using TestBuddy for the written national ranking exam who want the adaptivity working for coverage, not against it. Its principal limitation: no bank reproduces the pressure of the national graduatoria.
What TestBuddy offers for Concorso SSM right now
Figures below are vendor-reported from testbuddy.it and were last checked on 19 July 2026. Confirm current numbers and pricing on the product page.
| Item | Detail (vendor-reported, last checked 19 July 2026) |
|---|---|
| Exam covered | Concorso SSM (national specialty entrance), structured around the 140-question exam and 25+ specialties |
| Volume | 30,000+ questions; 700+ "microargomenti" for targeted practice |
| Adaptive / AI | "Buddy" AI assistant; "Calendario intelligente" adaptive schedule; "TB Score" per-subject level estimate; "Trova il mio livello" diagnostic |
| Extras | Digital manual with 300+ images; simulations; mnemonics; explanations for all questions |
| Access / price | Free plan available; paid from around €20/month (about €239.99 for 12 months); 3- and 6-month options (verify current pricing) |
| Not reproduced | The live national exam-day conditions and the ranking pressure of the graduatoria |
The exam you are actually preparing for
The Concorso SSM (Scuole di Specializzazione in Medicina) is a national, computer-based examination of 140 multiple-choice questions, each with five options and one correct answer, sat in 210 minutes. Scoring is +1 for a correct answer, −0.25 for a wrong answer and 0 for a blank, so the negative marking makes disciplined omission a real skill, not a weakness. The maximum test score is 140, plus up to 7 curriculum points, for a maximum of 147; the resulting national ranking, the graduatoria, decides who is placed in which specialty. The syllabus is broad — medicina interna, chirurgia, farmacologia and the rest. Verify the current structure and scoring against the ministerial decree, the official syllabus and any released questions for your year; the decree is the authority, not any third-party platform's description.
Two features of this exam should shape how you use an adaptive tool. First, negative marking means your workflow must train a calibrated sense of when to answer and when to leave blank. Second, it is a ranking exam: a rising personal score means little if your competitors are covering the syllabus more evenly, so coverage discipline matters more here than on a simple pass/fail test.
Baseline week: measure before you let it personalise
Before you accept a single adaptive recommendation, take a small, blueprint-stratified unseen sample across the major subjects and record it. TestBuddy's own diagnostic can seed this, but keep your own record too. This baseline does two things: it gives you a reference point the adaptive feed cannot later flatter, and it tells you which subjects are genuinely weak versus merely unpractised. Do not let the algorithm personalise your feed until you know your own starting distribution — otherwise you cannot tell whether an improving TB Score reflects real gains or a feed that has drifted toward your comfortable subjects.
First pass: set domain floors the algorithm cannot hide behind
The single most important control with any adaptive tool is a domain floor. Decide, up front, a minimum number of attempted questions and a minimum accuracy for every major syllabus area, and hold the feed to it. An adaptive engine optimises for a rising aggregate; left alone it can let a whole subject — say, a lower-yield surgical area — go under-sampled while the overall TB Score climbs on internal medicine. By fixing floors, you force even coverage. Check weekly that no subject is sitting below its attempt floor, regardless of what the aggregate score is doing.
Reserve a protected pool as well: hold back 15–20 per cent of items, untouched, for end-stage unseen mocks. With 30,000+ questions this is easy and costs you nothing, and it preserves an honest measurement that recognition cannot inflate.
An error taxonomy that decides the review action
Not every miss deserves the same response, and the negative marking makes the distinction sharper. Code each one:
- Knowledge gap: you did not know it. Action: short source read, then a spaced re-test.
- Misread stem: you knew it but misparsed. Action: a stem-parsing rule.
- Premature closure: you anchored early. Action: a "check one alternative" habit.
- Guideline or fact error: you held an outdated or wrong rule. Action: correct against the source and log it.
- Calculation error: arithmetic or unit slip. Action: a calculation checklist.
- Time-pressure error: rushed at the end of the 210 minutes. Action: pacing drill.
- Miscalibrated guess: you answered when you should have left it blank, and lost 0.25. Action: tighten your answer/omit threshold.
That last code is specific to this exam. Track how your guesses fare: if items you were "somewhat sure" on are net-negative once you subtract the penalty, your threshold for answering is too low.
Review interval: repeat, space, read or transfer?
Match the interval to the error. A knowledge gap wants a short source read now and a spaced re-test in a week — not an immediate repeat, which only tests short-term memory. A misread or calculation slip wants a rule and a fresh item to prove it. A genuine conceptual gap wants a transfer question on the same principle, not the identical item. Reserve immediate repeats for almost nothing; they are the least informative review you can do.
Mixed-block switch: objective criteria
Reduce topic-filtered practice and increase timed, random, mixed blocks once a subject clears its floors — its attempt minimum and its accuracy target on unseen items. As with any ranking exam, the aim is to spend your last weeks under conditions that resemble the real 140-question, 210-minute session: mixed subjects, full negative-marking rules, and a clock. Do not wait to exhaust the 30,000 questions; clearing floors, not finishing the base, is the trigger.
Exit criteria: five signals, not bank completion
- Coverage floor: every major subject attempted to its minimum, with no syllabus area hidden behind the aggregate.
- Stable first-attempt performance: unseen accuracy holding across subjects, not just the ones the feed favours.
- Pacing: you complete 140-item simulations within 210 minutes with margin.
- Retention: spaced re-tests of earlier misses hold up.
- Official-material calibration: your performance agrees with released official questions for the exam — the closest thing to a criterion standard you have.
Your TestBuddy TB Score is a useful internal signal, but it is not your ranking. Read it the way the percentage explainer reads a Q-bank percentage: one input, not the verdict.
Worked example: a seven-day week for a graduate
Give TestBuddy one job — adaptive first-pass coverage with floors — and iatroX one job — transfer practice on fresh, unseen items. No proprietary-algorithm claims are needed; a second non-overlapping bank simply supplies clean stems.
- Monday: a timed mixed TestBuddy block under full scoring rules. Code every miss; watch the miscalibrated-guess code.
- Tuesday: for the principles you missed, answer fresh iatroX items on the same principles; right cold means retire, wrong means step back.
- Wednesday: targeted work on the subject sitting lowest against its floor — not the one the feed offers, the one your matrix flags.
- Thursday: spaced retrieval of last week's coded misses; a short source read on the weakest concept.
- Friday: another timed mixed block; log guess calibration and pacing.
- Saturday: a longer simulation drawn partly from your protected pool; read unseen accuracy and time management.
- Sunday: rest or a 30-minute review; update the coverage matrix and set next week's floor targets.
Decision checklist: continue, supplement, switch or stop
- Continue while TestBuddy misses teach new content and unseen accuracy is rising evenly across subjects.
- Supplement with iatroX for unseen transfer once misses are recognition rather than knowledge gaps.
- Switch only where a coverage audit shows a subject TestBuddy under-serves and another resource covers better — a measurable gap.
- Stop new questions when all five exit signals are met, and move to full official-condition simulations.
Bottom line
TestBuddy's adaptivity and large question base make it a strong first-pass engine for the Concorso SSM, and its scheduling and diagnostic tools are genuinely helpful. The discipline it cannot supply for you is coverage against a ranking field and calibration under negative marking. Set domain floors so the feed cannot hide a subject, protect an unseen pool for honest measurement, code your guesses as carefully as your errors, and exit on signals and official-material agreement rather than on finishing the base. Used that way, the algorithm works for your ranking rather than for a comfortable score.
Frequently asked questions
Is TestBuddy enough for Concorso SSM on its own? Its question volume and adaptivity can carry most of the first-pass load, and many candidates prepare largely within one platform. But a single adaptive feed cannot both teach you and honestly measure you, and its optimisation for a rising aggregate can mask uneven coverage on a ranking exam. Pair it with the official released questions for calibration and a second, non-overlapping bank for unseen transfer, and it becomes a strong core rather than a single point of failure.
Which Concorso SSM component does TestBuddy not reproduce well? It reproduces the item format well; what it cannot reproduce is the exam-day condition and the ranking pressure of the national graduatoria — the full 140-question, 210-minute session under negative marking with a place on the line. You approximate that only with disciplined full-length simulations under exact scoring rules and, ideally, official released questions. Practising items well is necessary but not the same as rehearsing the whole session.
How many TestBuddy questions should I complete per day for Concorso SSM? Coverage and review quality matter more than raw volume. A sustainable target is one or two timed mixed blocks a day with full error-coding, rather than hundreds of skimmed items. If your daily count is high but you are not coding misses or checking guess calibration under the −0.25 penalty, you are training speed without judgement. On a negatively-marked ranking exam, calibrated accuracy beats volume.
When should I stop using TestBuddy and move to mixed mocks? Move a subject into timed mixed blocks as soon as it clears its attempt floor and accuracy target on unseen items — not when you have exhausted the base. Stop adding new topic-filtered practice altogether when coverage is complete, unseen accuracy is stable across subjects, pacing fits 210 minutes and your results agree with official released questions. From there, full official-condition simulations are the highest-value activity left.
How should I combine TestBuddy with iatroX without duplicating practice? Assign non-overlapping jobs. TestBuddy is your adaptive first-pass and coverage engine; iatroX is your unseen-measurement and transfer layer. When TestBuddy exposes a weak principle, do not immediately repeat the item — test the principle on a fresh iatroX item to confirm transfer. That keeps your measurement clean and turns two resources into a learn-then-verify loop rather than the same questions twice. The two-Q-bank rule shows how to keep them from overlapping.
Editorial notes and references
Written by Dr Kolawole Tytler, NHS GP and founder of iatroX. Last checked 19 July 2026; platform figures are vendor-reported and may change — verify the current count, features and pricing on the product page. Exam structure and scoring should be confirmed against the current ministerial decree and official syllabus for your year. Disclosure: iatroX operates a Concorso SSM question bank that competes with TestBuddy; iatroX's role here is confined to unseen transfer practice and measurement, and it makes no claim to a proprietary adaptive algorithm. Corrections via the feedback route on iatrox.com.
References: Ministero dell'Università e della Ricerca — bando and decree for the Concorso SSM, official syllabus and released questions; TestBuddy SSM product pages (testbuddy.it); iatroX Concorso SSM bank (https://www.iatrox.com/concorso-ssm); "Your Q-Bank Percentage Is Not Your Exam Score" (https://www.iatrox.com/blog/qbank-percentage-not-your-exam-score); "Question-bank completion is not coverage" (https://www.iatrox.com/blog/question-bank-completion-is-not-coverage-how-to-build-a-blueprint-coverage-matrix-for-any-medical-exam).