A large case count is the single easiest number for a platform to advertise and the single least reliable number for a candidate to choose a platform by, because raw count says nothing about whether that breadth is clinical, meaningfully behavioural, or merely nominal, cosmetic variation dressed as distinct content. This article exists to redirect attention from the headline figure to the metrics that actually determine whether a large case bank represents genuine educational depth.
The metrics that matter more than raw count
Number of distinct clinical constructs: how many genuinely different underlying clinical problems the case bank covers, as opposed to many surface variations of the same small set of core presentations. Examination-blueprint coverage: whether the case bank maps proportionately to the actual examination's station-type distribution, rather than concentrating heavily in whichever station type is easiest to write or simulate. Percentage independently reviewed: what proportion of the case bank has been through a defined clinical and educational review process, as distinct from generated or lightly checked content. Number of genuinely different patient behaviours: whether patients across the bank display meaningfully varied personality, resistance, emotional presentation and communication style, or whether most cases share an underlying conversational template with different presenting complaints layered on top. Difficulty distribution: whether the bank spans a genuine range from straightforward to complex, or clusters at one difficulty level regardless of its advertised size. Proportion with current UK management: whether the clinical content reflects current UK guidance specifically, a currency check every case bank in this category needs applied individually rather than assumed from overall platform quality. Frequency of content updates: whether the bank is actively maintained as guidance changes, or was built once and left largely static. Number of repeated templates with cosmetic changes: the single most direct test of whether a large headline count represents genuine variety, since a case bank built from a smaller number of underlying templates with names, ages and presenting complaints swapped will produce a large count with considerably less genuine educational range than the number suggests. Feedback validity: whether the automated critique attached to each case is specific and accurate, independent of how many cases exist to generate that feedback. And unseen transfer performance: whether practice across the bank produces skill that generalises to a genuinely novel case, the ultimate test any case-bank size claim should eventually be measured against.
The worked comparison, read correctly
Simsbuddy advertises more than 100 expert-written cases distributed across five examination categories, meaning the per-examination depth is considerably smaller than the headline total suggests once divided across SCA, PLAB 2, PACES, CASC and UKMLA. MyMedi8 advertises more than 650 scenarios concentrated specifically within PLAB 2, a materially different comparison, since a single-exam platform's case count is not directly comparable to a multi-exam platform's total without adjusting for how many examinations that total is spread across. Quesmed advertises more than 450 OSCE stations with more than 340 AI-enabled among them, worth noting the distinction between total station count and AI-interactive station count specifically, since not every advertised station in a large total is necessarily the AI-simulation feature a candidate is actually evaluating the platform for. The much larger Geeky Medics ecosystem reflects its longer-established, broad-scope position in the market rather than a directly comparable single number to any of the above. And specialist SCA banks report roughly 100 to 350 cases depending on the specific provider, a range worth checking against the SCA's actual 12-station structure to judge whether even the smaller end of that range represents adequate variety for genuine unseen-case practice across a full mock circuit.
What this article deliberately does not do
It does not announce a numerical winner, because the honest answer depends entirely on the metrics above, which none of these headline totals directly reveal. The recommended approach for any candidate actually choosing between platforms: sample and classify roughly 30 cases from each product under serious consideration, checking for genuine construct diversity, difficulty range, and repeated-template patterns, a modest time investment that reveals far more about real depth than comparing any two headline numbers ever will. Every case-count figure cited above is provider-reported and should be treated as dated at the point of publication, since case banks in this category grow, and occasionally shrink or get restructured, over time.
Frequently asked questions
Is a smaller case bank ever the better choice?
Yes, when its smaller count reflects genuinely reviewed, well-distributed, clinically current content, which can provide more reliable calibration value than a larger bank containing more repeated templates and less independent review, exactly the depth-versus-breadth trade-off this article's sampling recommendation is designed to reveal.
How many cases does a candidate actually need to work through before an examination?
Fewer than most large headline case counts suggest, provided the cases worked through span genuinely distinct constructs and difficulty levels; working through 30 to 50 well-chosen, varied cases with proper review and correction typically builds more transferable skill than working through hundreds of loosely varied ones without that discipline.
Should case count factor into platform choice at all?
As a minor factor once construct diversity and review quality have been checked directly, yes, since more genuine variety within a well-reviewed bank is still meaningfully better than less; the caution is specifically against using the raw headline number as a primary or sole decision criterion.