Problem-First Learning

Grinding practice problems before deciding what the topic is for pays one question type per question practiced. Nothing cycles back onto the map. Each attempt is guesswork against a stored procedure, hits look accidental, and in-context questions stay out of reach because they were never part of the practiced pattern. Formula-heavy subjects change that payback by locking the problem the topic exists to solve — every way it applies — and only then learning the equations.

Inventory, then why

Formula-heavy subjects invite the trap. A sliver of theory, a jump to the equations, and the grind starts. Applying early is genuinely good for learning, which is why the grind feels productive. The missing cycle is the return to the knowledge: what the result means for the map, not only whether the steps ran. Without that return, each new wrapper looks like a new procedure. Jumping to equations is the trap. Worked examples are early application with guidance, and they belong after the map, not in place of it.

The first pass is a flat list of names — main concepts, equation types, keywords — pulled from a survey, the curriculum, and a reference, as one list. The urge to dive into equations, and the overwhelm at the list’s size, are both normal. The pass starts slow and then accelerates, because early answers pre-cover later relationships. The first names on the list are slow; by the time derived terms arrive, those relations have already been met while hunting what the earlier names were for.

Every keyword is interrogated in a fixed order before any equation is opened: what it is, why it is important, how it relates. In formula subjects the second and third carry the weight. That is the same sequence Layers of Learning names as big-picture before details. One question is held while searching. A resource that answers slowly, or in the wrong grain, is skipped. Once the question is answered, the search stops. Fine detail is a later pass. A page in front of the eyes is not a reason to linger on it.

The tight loop, then stop

Aim and shoot then alternate in tight loops. Every three to five minutes — a working figure of this system, not a finding — what was just learned is asked against everything still on the list: same, different, cause, consequence. The list is a standing set of aims; the reading is the shooting. Bear Hunter System is the cycle this specialises; the loop is not re-taught here.

Relations are mapped as simply as the page allows. The artifact is a logical framework of the topic, never a transcript. Survive and Thrive is why that relation-map is the keep-signal.

A purpose-locked map for a change-and-accumulation topic looks like this. The topic splits into a change-operation and its inverse. The change-operation produces a rate, used for slopes, extrema, and smoothness. The inverse rebuilds the original quantity, used for accumulated totals, averages, and areas. No calculation has been run. Every later equation arrives as a tool for one of those jobs.

The stop is when purpose is locked: for every equation not yet learned, the job it will do is already named. Equations then install fast, and practice turns into hypothesis testing. Ease is not speed at the steps. Ease is knowing when a tool applies, because the job it does is already known.

What “problem first” is not

The problem is the purpose frame — what this tool is for, every way it applies — not an unsolved calculation opened cold. That is also why this is not classroom problem-based learning, which is a different, often minimally guided, design. The house name is easy to misread as either.

Applications filed without a purpose become more memoranda. Applications filed as uses of a tool already wanted transfer. Novices learn procedures better from worked examples than from solving first. After the map, worked examples are how the equations install. A blank worksheet attempted cold is the version the evidence contradicts.

Practice then starts to transfer in direction, not as a counted ratio: later exercises stand for a class of problems rather than adding one more routine. Formula practice produces two kinds of item. The first is the same form as the drill — plug in, run the steps. The second wraps the same math in a situation: a vehicle changing speed along a road, a quantity accumulating over time. The wall is the second kind. Success on the first is not evidence the map exists. Cave Theory is why purpose-before-information is the keep-signal: isolated facts slip, and purpose-tied connected ones stay.

Author the hard item

The wrapped item is also something that has to be written. Questions get authored, maximally hard, and always applied. Once answering found items feels comfortable, the next items wrap several concepts in a realistic combination, hard enough that the item looks like one that would rather not be sat. Combining and comparing sit a level above applying any single tool, and that level only exists if the map does.

Answering found questions while being unable to invent a hard applied one flags a conceptual gap. The repair is at the map: rebuild how everything fits and what each piece is for, then author again. Grinding more found practice leaves the gap untouched.

After purpose is locked, a practice item is a hypothesis: this job should take this tool, perhaps combined with that one. A miss updates the map — a third element was needed, or this pairing does not apply. A hit confirms it. That is a different activity from memorising a fresh procedure for each wrapper.

Two imbalances, two repairs. Cannot run the steps on a found item: the procedure is thin, and practice against the map is the repair. Can run found items but cannot invent a hard applied one that combines several tools: the map is thin, and purpose and relations get rebuilt, then authoring happens again. Found questions plus authored ones are protective as a pair. Swapping hard items with a partner puts both maps under test in the same sitting.

Research Foundations

Routine practice without a principle map transfers poorly. Why and relate questions before procedures help the concept layer; the house order (what, why, how it relates) is a default, not a finding. The papers sit in Sources.

The same split now governs coding and agent work. An agent can hold syntax, API details, and step-by-step execution — the equivalent of the equations — leaving the human the map of what each tool is, why it exists, and how the pieces relate, with purpose locked before implementation. Someone who can operate the tools but cannot author a maximally hard, realistic problem for the system being built is missing the map. The repair is the conceptual layer, rebuilt before more reps through the tool. Agentic Engineering is where that migration goes. The map-versus-equations split is the one Declarative, Procedural, and Conditional Knowledge defines.

The first pass is slow and overwhelming: a flat list of names, no equations yet, and the urge to dive into calculation is constant. Re-chunking a topic to lock purpose is a real second pass. The benefit in the same breath: later practice stops paying one type per item, and applied wrappers stop arriving as new procedures.

This is not unguided solving before instruction, and it is not classroom problem-based learning. A reader whose subject is weakly interdependent drills should not reach for this first. If after the inventory-and-why pass an unlearned equation still has no named job, the map is not locked and the drill book does not start. If the found set can be finished and one hard applied item still cannot be invented, grinding stops and the map is rebuilt. Two sessions in which each wrapper needs its own new steps mean the purpose pass did not happen.

Purpose locked means every unlearned tool has a named job. A first calculation pass after that should feel like testing a briefing, not like meeting a new subject.

Practice, once purpose is locked, is a test of that briefing. A miss updates the map. The item no longer buys only itself.

Open Questions

How does the purpose pass run on a subject that is almost entirely procedure — a syntax, a technique catalog — where there is no equation to delay?

Links

• Bear Hunter System — the aim–shoot–skin cycle this specialises
• Cave Theory — why purpose-before-information is the keep-signal
• Survive and Thrive — the relation-map that memory actually keeps
• Layers of Learning — big-picture-before-details as the same sequence
• Declarative, Procedural, and Conditional Knowledge — the split this page spends: map vs equations
• Agentic Engineering — where the AI-absorbs-procedure angle goes

Sources

• Hatano, G., & Inagaki, K. (1986). Two courses of expertise. In Child Development and Education in Japan. Routine versus adaptive expertise: practice without a principle map stays on the practiced path.
• Singley, M. K., & Anderson, J. R. (1989). The Transfer of Cognitive Skill. Harvard University Press. Single-path productions transfer poorly off the practiced item type.
• 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. Elaborative interrogation — why and how-it-relates — supports the concept layer.
• Chi, M. T. H., Bassok, M., Lewis, M. W., Reimann, P., & Glaser, R. (1989). Self-explanations: How students study and use examples in learning to solve problems. Cognitive Science. Generation and self-explanation as the class this page’s authoring step belongs to.
• Sweller, J., van Merriënboer, J. J. G., & Paas, F. (1998/2019). Cognitive architecture and instructional design. Educational Psychology Review. Worked-example effect: novices learn procedures better from worked examples than from solving first.
• Kapur, M. (2014). Productive failure in learning math. Cognition and Instruction. A bounded unsolved problem before instruction can help conceptual transfer. Neighbour, not identity.
• Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal guidance during instruction does not work. Educational Psychologist. Classroom problem-based learning is a different, often minimally guided, design.