Cert Studio

Problem

The certification guides list domains and weights, but no way to practise that respects those weights or explains why a wrong option looks right. I wanted to measure myself with my own questions, spread like the real exam, with no accounts and no server.

Approach

One HTML page and three data files: the browser does all the work. A mock spreads the questions across the domains by their official weight using the largest-remainder method, runs against a timer and returns a scaled 100–1000 score with a pass at 720, the per-domain result and the traps you fell into. Every wrong option in the bank carries a trap tag, and the validator rejects one without it.

Review mode serves first what you have not seen and what you missed, and comes back to what you got right at growing intervals (simplified SM-2). Progress is kept in local storage and exports to JSON; with ?sample=1 everything lives in memory under a fixed seed, which is how the screenshots are recorded.

Limits: an unofficial study aid, not affiliated with the certification’s issuer; the real scale is not public and a weighted linear one is used here; the bank was reviewed by a model, not a human panel; progress does not travel between browsers. Code under Apache-2.0; content under CC BY 4.0.