The problem with 'study everything'
Re-reading notes feels productive and tests nothing. Real exams punish specific gaps (the three objectives you keep missing) while rewarding nothing for the tenth review of what you know. Weak-spot training flips the default: every session aims at the gaps first.
What the engine watches
- What you miss, by objective. Misses cluster on objectives, not questions. Three misses on firewall rules is a firewall-rules problem even if the questions differed.
- How sure you felt. After each answer you can mark sure / think so / guessing. Confident-and-wrong burns hotter than unsure-and-wrong (misconception vs gap), and correct-but-guessing schedules a review anyway (luck is not knowledge).
- When you will forget. Spaced-repetition scheduling (FSRS) predicts each memory's decay and serves reviews just before they fade, the cheapest durable learning known.
- Speed. Slow correct answers on timed skills (subnetting, ports) count as shaky even when right.
How a session gets built
Roughly: due reviews first (never skipped), then weakness-targeted questions, then fresh coverage of unseen objectives, then a small stretch. Same-day misses return the same day; confirmed misconceptions get contrast cards that teach the difference, not just the fact.
How to use it instead of fighting it
- Answer honestly, mark confidence honestly. Gaming the inputs poisons your own schedule.
- Do the reviews. Skipping due cards to chase new questions feels faster and learns slower.
- Trust the queue order. 'Why this card' is shown on every flashcard. If the reason makes sense, the algorithm is working.
- Check Progress weekly, not daily. Daily noise hides trends; weekly domain heatmaps show whether the weak spots are actually shrinking.
Calibration in action (a true story in miniature)
You answer a firewall question correctly but marked 'guessing', the engine schedules a firewall review in two days instead of two weeks. Next you miss a crypto question marked 'sure', that misconception jumps the queue ahead of everything except due reviews, and a contrast card (symmetric vs asymmetric, side by side) appears the same day. Two honest taps, two correct scheduling decisions. That is the whole system.
Settings worth changing
- Confidence prompts: on. One extra tap per question buys the calibration above.
- Explanation detail: full while learning a domain, concise during review weeks.
- Daily goal: smaller than your ambition. A 20-minute goal you hit daily beats a 60-minute goal you abandon. Streaks compound, guilt doesn't.
What it cannot do
It cannot study for you, and it cannot fix zero foundations. Brand-new topics need a first explanation (a video, a guide, an objective page) before drilling helps. The engine aims; you fire.
A week of weak-spot training
Monday: due reviews first, then the two weakest domains. Stop while you still feel sharp. Tuesday: one lab from the weakest domain, then twenty questions on the same objective. Wednesday: reviews plus a short drill on a skill that has gone quiet, usually ports or subnetting. Thursday: a full domain mix, no new material, and a note on every miss. Friday: the domain that dropped this week, plus a lab that proves it. Saturday: one simulation or a longer mixed set, depending on how close the exam is. Sunday: off, or flashcards only.
The week has one rule: the weak domain appears at least three times. Frequency beats duration. A domain you touch for ten minutes four times will stick better than one long session on Sunday.
When the queue feels wrong
Sometimes the engine serves a question you know cold. Do it anyway, quickly, and mark confidence honestly. The schedule is not only about that question. It is also confirming that the memory survived the gap. Skipping reviews because they feel easy is how people arrive at exam week having forgotten the first three domains.
If the queue serves a topic you have never seen, that is a coverage gap, not a mistake. Read the explanation, follow the link to the objective page, and let the next review confirm it. Do not judge the algorithm by one surprising card. Judge it by whether your domain heatmap improves over two weeks.
How to read the domain heatmap
Look at two things: the trend and the spread. A domain that is red but rising is working. A domain that is green but falling means you have not touched it in a while. The spread matters because exams weight domains unevenly. A weak 28% domain costs more than a weak 12% one, and the heatmap does not know the weights. Multiply the gap by the weight before you decide where the next hour goes.
Calibration without guilt
Confidence prompts are not a personality test. Sure, think so, and guessing describe the evidence you had at that moment, not how smart you are. People who mark everything sure to look confident poison the schedule and then wonder why it keeps serving easy cards. People who mark everything guessing flood themselves with reviews. Aim for honest, not impressive.
Where the method stops
Weak-spot training is a targeting system, not a teacher. It cannot explain a concept you have never met, and it cannot make you practice a hands-on skill you have not tried. When a domain stays red after two weeks of reviews, the problem is usually upstream: the explanation never landed. Stop drilling, read the objective page and the guide again, then do one lab before returning to questions.
It also cannot fix exam technique. If your accuracy is fine in practice but simulations go badly, the problem is time and reading, not knowledge. Practice under the clock, and read the PBQ strategy before the next sitting.
And where the gap is a missing explanation rather than a fading memory, no scheduler will fix it: the Security+ resources comparison covers where to get the explanation that actually lands.