Log inSign up
The Drill Method

Your material. 
Drill's protocol.

Create Cards from what you need to know. Train against Truths and Traps. Count proof only when a Run clears cleanly.

Create CardsStart verifying
The protocol

Five moves the student actually makes.

No ornament. No optional path. Each step is what Drill makes the student do — and what makes the verdict mean something.

01.

Cards hold concepts

A Card is the unit of Mastery. It should be small enough to test cleanly and important enough to matter on exam day.

02.

Truths and Traps expose the distinction

A Truth states the concept correctly. A Trap changes one meaningful detail. Easy falsehoods do not help. Plausible Traps reveal whether the concept is actually known.

03.

Runs keep the loop short

A Run uses a small set of Cards. If one answer fails, the Run resets and comes back shuffled. The same material returns immediately because the miss is the point.

04.

Perfect Runs are proof

A correct answer inside a failed Run is useful, but it is not proof. Drill only counts a clean Run because exams do not care that the right answer felt familiar yesterday.

05.

Freedom in creation. Strictness in verification.

Use notes, PDFs, past papers, lectures, or AI. Drill does not decide what is true for you. It gives your material a structure, a loop, and a verdict.

Readiness states

Readiness is a factual state.

The dashboard doesn't interpret. It reports. Every Card sits in one of these states — factually, not ambitiously.

  1. Untested

    The Card has not faced a Run.

  2. Weak

    The Card has shown unreliable performance.

  3. Unproven

    The Card has been seen but not confirmed.

  4. Mastered

    The Card has produced enough clean proof.

  5. Due

    The proof needs to be checked again.

Origin

Why this protocol had to exist.

Drill came from medical anatomy: dense systems, close distinctions, and MCQs where familiarity fails.

The breakthrough was a stricter loop: restart on error, shuffle the retry, and only count a full set when every answer came back correct.

That became the Perfect Run. Not punishment. Short-loop repetition with a clean verdict.

MCQ model

One concept. Several ways to fail.

Drill treats MCQ thinking as a structure, not a question format. The choices become the training surface.

One tested concept becomes a Card. A correct formulation becomes a Truth. A plausible wrong formulation becomes a Trap.

The student is not memorizing one sentence. They are proving the distinction still holds when the wording changes.

What the protocol leaves out

No streaks. No XP. No badges.

Activity rewards are a study-system problem. They optimize for the proxy — minutes, sessions, login days — and corrupt the signal the student actually needs.

A 90-day streak doesn't know what's on the exam. A badge doesn't know which Card you can produce cold. The dopamine loop feels like progress and predicts none of it.

Streaks

Reward opening the app. Don't measure retrieval.

XP

Reward minutes spent. Don't measure what's proven.

Badges

Reward activity milestones. Don't reflect Mastery.

Leaderboards

Reward comparison. Don't say whether you know the material.

The only number that matters is the Mastery state of each Card. Cold, factual, tied to actual retrieval — not to login days.

The craft

From science to Card design.

How to write effective flashcards and Drill Cards — the six rules for atomic concepts, standalone Statements, Truths, Traps, mirror distractors, single-pivot questions, and plausible wrong answers.

Five cognitive principles decide why testing works. Six rules decide whether your Cards actually test anything. Apply them to AI-generated content, hand-written Cards, imported decks — the rules don't change.

01.

Atomic concept

Desirable difficulty

One concept per Card. Not a paragraph, not a summary. If a Card holds two ideas, split it — two sharp Cards beat one bloated one every time.

Bad

The mitochondria produce ATP through oxidative phosphorylation, which requires the electron transport chain and ATP synthase located in the inner membrane.

Good

ATP is produced by ATP synthase on the inner mitochondrial membrane.

02.

Standalone Statement

Active retrieval

Each Statement stands on its own. No 'it', 'this process', 'the above'. Assume the reader has no memory of the Card before it.

Bad

This step produces two molecules of pyruvate.

Good

Glycolysis produces two pyruvate molecules from one glucose.

03.

Truth and Trap

Testing effect

At least one Truth and one Trap per Card. A Card with only Truths trains recognition — the feeling of 'I've seen that' — not discrimination.

Bad

Card with one Statement: 'Glycolysis occurs in the cytoplasm.' (Truth only)

Good

Truth: 'Glycolysis occurs in the cytoplasm.' — Trap: 'Glycolysis occurs in the mitochondrial matrix.'

04.

Mirror Trap, single pivot

Desirable difficulty

The Trap pivots exactly one factual detail from the Truth. Not two. Not an unrelated claim. The single pivot is what forces real retrieval instead of pattern-matching.

Bad

Truth: 'ATP synthase sits on the inner membrane.' — Trap: 'Mitochondria are found only in plant cells.' (unrelated)

Good

Truth: 'ATP synthase sits on the inner membrane.' — Trap: 'ATP synthase sits on the outer membrane.' (single pivot)

05.

Plausibility

Metacognitive calibration

The Trap must be credible to someone who half-knows the material. If it's obviously wrong, the Card trains nothing. Aim for the mistake a tired student would actually make.

Bad

Trap: 'The mitochondria were invented by Isaac Newton in 1687.'

Good

Trap: 'The mitochondria were first observed by Robert Hooke in the 17th century.'

06.

Concision

Active retrieval

Short wins. If a word can be cut without losing meaning, cut it. Padding slows retrieval and dilutes the pivot the Card is actually testing.

Bad

It is worth noting that glycolysis, a metabolic process, takes place within the cytoplasm of the cell.

Good

Glycolysis occurs in the cytoplasm.

A good Card isn't harder. It's sharper.

The gap

The research said it first. The tools didn't listen.

Since 1885, memory research has been remarkably consistent about what makes knowledge stick.

Most study tools optimize for something else: time spent, cards seen, streaks kept, reviews scheduled.

Those are proxies. None of them measure whether you know.

Drill is built the other way around — the mechanic serves the evidence.

The science

Five principles. One protocol.

Each principle below is a well-documented finding in cognitive science. Each maps to a specific mechanic in Drill. No ornament. No appeal to novelty.

01.

The Testing Effect

What the research shows

Retrieving information from memory strengthens it more than re-exposing yourself to it. Students who test themselves once retain more after a week than students who re-read the same material four times.

The evidence
  • Roediger & Karpicke (2006) — Psychological Science
  • Dunlosky et al. (2013) — Psychological Science in the Public Interest — meta-review classifying testing among the two most effective study techniques
What Drill does

A Card is not a thing to re-read. It's a thing to be tested on. Every interaction in Drill is a retrieval event, not an exposure event.

02.

Active Retrieval

What the research shows

Recognizing an answer (seeing it and saying "yes, I knew that") is not the same as generating it. Forced retrieval — even in a binary format — produces deeper encoding than recognition alone.

The evidence
  • Karpicke & Blunt (2011) — Science — retrieval practice outperformed elaborate concept mapping for long-term learning
  • Pyc & Rawson (2010) — Journal of Memory and Language — the retrieval effort hypothesis
What Drill does

Statements force a decision. True or false. No half-answer, no "I kind of remember." The Trap — a plausible falsehood — makes recognition insufficient. You have to retrieve the distinction, not just the gist.

03.

The Spacing Effect

What the research shows

The same total study time, distributed across multiple sessions, produces far better long-term retention than the same time massed into one session. The optimal gap scales with the target retention interval.

The evidence
  • Ebbinghaus (1885) — the original forgetting curve
  • Cepeda, Pashler, Vul, Wixted & Rohrer (2006) — Psychological Bulletin — meta-analysis of 317 experiments confirming the spacing effect
What Drill does

Drill's selection algorithm combines three objectives — introduce new Cards, bring back due Cards at the right time, prioritize weak Cards that have shown unreliable performance. Scheduling is not a feature. It's the default.

04.

Desirable Difficulty

What the research shows

Conditions that feel harder in the moment — spaced practice, interleaving, retrieval under pressure — produce stronger long-term learning than conditions that feel smoother. Ease of study is not a signal of learning.

The evidence
  • Bjork (1994) — introduced desirable difficulties in Metacognition: Knowing about Knowing
  • Bjork & Bjork (2011) — Psychology and the Real World (APS)
What Drill does

The Perfect Run is not punishment. It's a deliberately harder condition that produces cleaner proof. A Run with one mistake resets — same Cards, different order — until the student clears it. The difficulty is the mechanism.

05.

Metacognitive Calibration

What the research shows

Students are systematically overconfident about what they've learned through passive review. Re-reading increases familiarity with the material, which feels like knowing — but predicts retention poorly.

The evidence
  • Koriat & Bjork (2005) — Journal of Experimental Psychology — the illusion of competence
  • Dunlosky & Rawson (2012) — Learning and Instruction — overconfidence in self-assessment predicts exam failure
What Drill does

Drill's Mastery states — Untested, Weak, Unproven, Mastered — are factual, not interpretive. Tracking is asymmetric: errors always count, successes count only inside a Perfect Run. The dashboard doesn't reward effort. It reports the signal.

Signature mechanic

The Perfect Run, in detail.

Every decision in the Perfect Run traces back to a specific cognitive principle. Nothing here is arbitrary.

Parameter 01

7 to 13 Cards

Working memory holds roughly 7±2 chunks (Miller, 1956). Cowan's more recent estimate is 4±1 active chunks. A Run fits inside that window — no external recall aids needed, no overload.

Miller (1956) · Cowan (2001)

Parameter 02

Binary verdict

True-or-false forces a decision. No "almost." No graded confidence. The forced choice produces a cleaner retrieval signal than open-ended self-assessment.

Active retrieval · 2AFC paradigms

Parameter 03

Restart on error

The restart is a short-loop repetition mechanism, not a penalty. Same Cards, different order — dense retrieval, short cycle. The retrieval effort hypothesis (Pyc & Rawson, 2010) predicts that repeated effortful retrieval consolidates the trace faster than easy success. Then the loop closes. Most schedulers would still be sending you that Card three weeks later.

Desirable difficulty · Retrieval effort

Parameter 04

Asymmetric tracking

Errors count permanently; successes only count inside a clean Run. The asymmetry corrects the metacognitive bias that makes partial successes feel like mastery.

Metacognitive calibration

A Perfect Run isn't hard to be hard. It's hard to be honest.

Honest scope

What the method doesn't claim.

Drill is not a tutor.

It does not explain, teach, or generate authoritative content. You bring the material — or generate it with your own tools. Drill verifies whether you can produce it on demand.

Drill does not own truth.

Cards can be wrong. Traps can be misleading. You are responsible for reviewing and validating what goes in. Drill provides the framework to inspect, edit, and re-test.

Drill is not a substitute for understanding.

Testing works on content the student has first encountered. It reveals retention. It does not replace the initial learning. Use Drill after the lesson, the reading, or the LLM conversation — not instead of them.

Drill doesn't claim to know. It's built so you can.

Sources

Every claim traces back to the research.

The method is auditable. Below are the primary sources behind the five principles and the Perfect Run.

  1. 01
    Ebbinghaus, H. · 1885

    Über das Gedächtnis: Untersuchungen zur experimentellen Psychologie.

    Duncker & Humblot, Leipzig.

  2. 02
    Miller, G. A. · 1956

    The magical number seven, plus or minus two.

    Psychological Review, 63(2), 81–97.

  3. 03
    Bjork, R. A. · 1994

    Memory and metamemory considerations in the training of human beings.

    In Metacognition: Knowing about Knowing. MIT Press.

  4. 04
    Koriat, A. & Bjork, R. A. · 2005

    Illusions of competence in monitoring one's knowledge during study.

    Journal of Experimental Psychology: Learning, Memory, and Cognition, 31(2).

  5. 05
    Roediger, H. L. & Karpicke, J. D. · 2006

    Test-enhanced learning: Taking memory tests improves long-term retention.

    Psychological Science, 17(3), 249–255.

  6. 06
    Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T. & Rohrer, D. · 2006

    Distributed practice in verbal recall tasks: A review and quantitative synthesis.

    Psychological Bulletin, 132(3), 354–380.

  7. 07
    Pyc, M. A. & Rawson, K. A. · 2010

    Why testing improves memory: Mediator effectiveness hypothesis.

    Journal of Memory and Language, 60(4).

  8. 08
    Karpicke, J. D. & Blunt, J. R. · 2011

    Retrieval practice produces more learning than elaborative studying with concept mapping.

    Science, 331(6018), 772–775.

  9. 09
    Bjork, E. L. & Bjork, R. A. · 2011

    Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning.

    In Psychology and the Real World. Worth Publishers.

  10. 10
    Dunlosky, J. & Rawson, K. A. · 2012

    Overconfidence produces underachievement.

    Learning and Instruction, 22(4).

  11. 11
    Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J. & Willingham, D. T. · 2013

    Improving students' learning with effective learning techniques.

    Psychological Science in the Public Interest, 14(1), 4–58.

Go deeper

The protocol, applied.

Four guides for the study patterns Drill is built to enforce.

Past exam MCQ practice

Turn annales into Cards, keep distractors as Traps, and prove what still breaks.

Read the guide

Active recall app

Why testing beats re-reading, and how Drill enforces retrieval on every Card.

Read the guide

Spaced repetition app

How Drill schedules due Cards over time so retrieval stays effortful enough to stick.

Read the guide

Exam revision app

Using Mastery and Readiness states to know what's safe and what still needs work before the exam.

Read the guide
Ready?

The method is clear. The tool is ready.

One Run is all it takes to see how it feels — and what it reveals.

Try a short RunSee pricing