Deep Processing Tanking

Part of Deep Processing

Deep processing tanking is the stretch of years spent evaluating importance, comparing, analogizing, and forming schema by instinct, which clears curricula pitched at the average without a method being demanded. It holds until the material outgrows what instinct can carry. Copying a top learner’s visible strategy does not transfer the result, because the strategy was never what produced it.

The type is a history of automatic work, not a measured trait. People who have been evaluating importance, comparing, analogizing, and forming schema without naming any of it as a method clear ordinary environments so cleanly that no deliberate system is ever demanded.

Compressed Takeaways

• Deep processing is evaluating importance, comparing, analogizing, simplifying accurately, and building schema.
• Automatic processing can carry easier environments, and that success hides a missing method.
• The break usually arrives at a transition: school to university, university to work, a familiar domain to an unfamiliar one, individual work to higher responsibility, low volume to high.
• The diagnostic is not whether the work got harder. Harder work should cost more. Tanking is a disproportionate jump in hours for the same or worse result.
• Understanding is not enough. Material can be followed while no retrievable, usable structure exists.
• More effort only helps when it creates better organization. Effort spent fighting confusion, distraction, or unsupported detail is bad load.
• The repairs are handles, not replacements: thinking on paper, schema mapping, comparison, importance ranking, micro-retrieval, early application.
• The readiness filter is whether enough is already known to make the material simpler. If not, the work steps down a layer.

What the processing is doing

Deep processing is the work of making meaning rather than repeating a surface: evaluating how important a piece is, comparing it with what is already held, finding an analogy, simplifying it accurately, and placing it in a structure that can be retrieved later. Evaluation, stated as an act rather than a feeling, is a value judgment — how important this is, whether it is more or less important than that, what order they go in if they are ranked. Most learners finish a page asking only whether they understood it.

A strong processor runs a higher bar on the same sense-making test. The material does not make sense yet until it has been simplified, connected to what is already known, given an analogy, and ranked against its neighbours. The work keeps going until it is clear what matters, why, how it compares, where it fits, what it changes, how to simplify it, and how to use it. That is what produces schema rather than isolated facts. Surface detail can be forgotten and the idea rebuilt from the structure.

The same work can feel effortless. While it stays implicit it is never noticed as a set of skills, and never built into tools that still run when the day is bad and the volume is high.

Why the method never gets built

Three levers move results: time and effort, strategy, and natural processing. Challenge rises, and most learners add time and effort first. That lever is finite. Exhaustion or weak returns end it, and they are pushed onto strategy. Strategy is what carries them upward afterwards. The average learner’s strategy improved continuously across those same years because improving it was the only thing available.

A house built with one hammer goes up faster the longer and harder the hammer swings, right up to the jobs a hammer cannot do at all. No amount of swinging converts into those. This system’s own teaching model of the same fact: a learner whose methods are limited moves from thirty hours to sixty and lifts total mastery from about 66 percent to about 76 percent. Double the hours, ten points. Time and effort is finite; more of it only helps when it creates better organization.

Some methods are incompatible with the higher levels of mastery unless the processing is already high.

Someone who can lift a hundred kilos one-handed does not need clean technique to lift five or ten. School curricula, university examinations, and grading standards are pitched at the average, so almost any method clears them for a strong processor. The strategy lever is never pulled because it is never needed. The discovery arrives a long way from shore, in deep water, with no stroke to swim.

Natural processing is powerful and unreliable while it stays implicit. The first response to the jump is usually more hours. Grinding the hours, or having someone else break the topic down, clears the instance and leaves the cause in place, so the same wall returns. Being a hard worker is a real capability. Needing maximum effort for everything is not.

Nothing was lost

The learner has not become less capable. The limit that was reached is the limit of the strategies already in use. The test is not whether the work got harder. Harder work should cost more. Tanking is the disproportionate jump: the same or worse result for far more hours. The ratio has a real numerator in this system’s teaching model: mastery and retention held per hour, not content covered per session. Coverage never measures what is still there a week later, or whether the knowledge can be used rather than recognised.

Between the stretch when things were easy and the point where they stopped being easy, the difficulty moved roughly from a three to a seven. Did the strategy move from a three to a seven, or from a three to a four?

The chain is short. Strong processing produces early success on ordinary methods. Early success removes the pressure to build strategy. A transition arrives. Hours go up. Returns decline. The belief forms that ability dropped.

What the jump looks like from inside

Tanking appears at transitions: school to university, university to work, a familiar domain to an unfamiliar one, individual contributor to higher responsibility, low volume to high. A programme where everyone arrived as the top of their own school, taught at a level pitched for exactly that, removes the margin the strong processor had been living on. When complexity or volume rises past what instinct carries, performance drops or effort spikes.

The signs are concrete. Hours rise without proportional gains. The material makes sense but is not usable. Notes multiply without command. The work feels slower than the self-image. Rereading and rewatching stop solving it. Performance exposes gaps that passive understanding hid.

This is not the main explanation for attention-span problems, low motivation, or misuse of generated answers. Those are adjacent mechanisms. Tanking is specifically the strategy gap when natural depth stops scaling. If a few weeks of externalizing and testing early has not moved the effort-to-result ratio, strategy was not the bottleneck. Volume, sleep, prerequisite knowledge, and attention each produce the same surface and have their own repairs.

Understanding is not organization

Understanding is not enough. The material can be followed while no retrievable, applicable, modifiable structure exists.

Understanding	Organization
Makes sense while looking at it	What matters, where it fits, how to use it
Recognition, familiarity, short-term clarity	Schema, retrieval, transfer, application

Deep processors are especially exposed because they understand quickly, and quick understanding hides weak organization until performance. The replacement question is “How can I organize this?” Knowledge Mastery owns the levels between recognizing a thing and being able to use it. Schema Construction, Assimilation, and Reorganization owns how an organized structure is actually built, extended, and rebuilt.

The repair is not easier learning

The repair is not to make learning easier. Easy learning often removes the effort that creates durable encoding. The move is into the right kind of load, not out of load.

Good load and bad load

More effort only helps when it creates better organization. Load spent decoding, copying, or holding detail that has nothing to attach to occupies working memory, and the structure has to be built with what is left. Resources that interleave fine detail with occasional logic keep relevance fluctuating, so no model can form to hold what arrives, and difficulty stays high the whole way through.

weak: copying notes without judgment; decoding technical language too early; memorizing unsupported detail; fighting distraction; accepting another person's structure wholesale; pushing into material with no foundation

strong: comparing distant concepts; judging importance; simplifying accurately; explaining from memory; asking non-obvious relationship questions; building a schema map; testing a hypothesis; applying immediately

Cognitive Load is the page that owns this split: effort that builds the model against effort that occupies working memory without building it.

Readiness, and six handles

The readiness filter is a question already half-owned: do I know enough about this to make it simpler? If not, the work steps down a layer. A layer, in this system’s teaching model, is one of four: logic — the backbone, the largest chunks and the relationships that hold them; concepts — what a thing is and how it happens; important details — the ones that make a concept concrete; arbitrary details — carried only because they are assessed. Roughly a tenth of the material is conceptual, a fifth is important detail, and the remaining seventy per cent is arbitrary — and the arbitrary share shrinks as the first two layers get better, because relevance is what converts an arbitrary detail into a placed one. Ten thousand leaves hang off five or six branches. The operative move is to improve the backbone, not to memorise harder.

If a piece cannot be simplified, the next move is a simpler explanation, the missing terms, the prerequisites, the detail marked for later, or a rough schema first. Deep processors tolerate high effort and can read it as progress; the question keeps productive depth from becoming unsupported strain. Layers of Learning owns this structure and the volume estimate behind it.

The six repairs do not replace deep processing. They give it handles.

• Thinking on paper makes the evaluation visible, so it can be inspected and revised.
• Schema mapping places each piece against the others.
• Importance ranking gives each relationship an explicit significance rating against the wider topic, with a short bounded lookup when the rating is uncertain; gaps stop hiding, and relationships that felt settled turn uncertain.
• Non-obvious questions force comparisons the source never made.
• Micro-retrieval reconstructs a piece immediately, before checking.
• Early application ships an imperfect structure on purpose and tests it.

Thinking on Paper is the externalizing move and the likely first repair. Best-attempt Encoding is what testing the structure early means in practice.

Owning a skill takes more than understanding it, remembering it, and applying it correctly. It has to run consistently and easily when tired, stressed, or having a bad day. That fourth requirement is the one that fails, and it fails exactly where the skill was most needed. Implicit depth only runs on a good day. The handles are what change that.

Practice time is fixed. Spreading it across five new skills grows each at a fifth of the rate, and a payoff that would have arrived after ten focused hours arrives after fifty. One or two at a time, worked until each is correct and consistent.

For the average learner the eventual limiter is the processing habits themselves, and retraining those runs in months to years. A learner whose processing is already strong is limited by strategy instead, and strategy moves on a much shorter clock. That is why the repair is often fast. It is a reason, not a promise.

The tell is not that the material feels easier — feeling easier is what this page warns against. The same result arriving in fewer hours, and the topic rebuildable from memory with the notes closed, is the check.

This is a practitioner category, not a measured syndrome. The nearest public research establishes that instructional support has to change as skill grows — which supports the general shape, that what helped before stops helping, and does not establish a named condition. The reframe is only earned once the ordinary causes of the same surface are ruled out.

Making depth repeatable

The operating model is a short loop: diagnose the transition, identify the hidden strategy gap, externalize, move effort into good load, apply the readiness filter, test early, make depth repeatable. The work is turning deep processing from an unconscious strength into a conscious skill. The repair adds no processing power. It gives the processing handles, and the difference between a capability and a tendency is whether the depth still runs when the learner is tired and the volume is high.

Links into the knowledge base

• Deep Processing — the parent dimension; this page is one failure mode of processing that is strong and implicit.
• Thinking on Paper — the externalizing move, and the likely first handle.
• Schema Construction, Assimilation, and Reorganization — how an organized structure is built, extended, and rebuilt.
• Knowledge Mastery — the ladder from recognizing a thing to being able to use it.
• Best-attempt Encoding — shipping an imperfect structure on purpose and testing it.
• Cognitive Load — good load against bad load, with working-memory occupancy as the reason rather than fuel.
• Layers of Learning — the logic / concept / important-detail / arbitrary-detail structure that “step down a layer” names.

Open Questions

What are the earliest warning signs, before the ratio breaks?

Which repair comes first when more than one is missing?

Should good load and bad load have their own page?

How do the six handles map onto Aim, Shoot, and Skin?

Can tanking be partial — one domain failing while another is fine?

Sources

Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11(6), 671–684. Depth of processing as the durability mechanism.

Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention of words in episodic memory. Journal of Experimental Psychology: General, 104(3), 268–294. The experimental follow-through.

Rozenblit, L., & Keil, F. (2002). The misunderstood limits of folk science: An illusion of explanatory depth. Cognitive Science, 26(5), 521–562. Understanding while looking is not a usable structure.

Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981). Categorization and representation of physics problems by experts and novices. Cognitive Science, 5(2), 121–152. Expert representation is organized by principle, not surface.

Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In J. Metcalfe & A. P. Shimamura (Eds.), Metacognition (pp. 185–205). MIT Press. Easy-feeling study can be the weaker encoding.

Bjork, R. A., Dunlosky, J., & Kornell, N. (2013). Self-regulated learning: Beliefs, techniques, and illusions. Annual Review of Psychology, 64, 417–444. The desirable-difficulties follow-through.

Sweller, J., van Merriënboer, J. J. G., & Paas, F. (2019). Cognitive architecture and instructional design: 20 years later. Educational Psychology Review, 31, 261–292. Germane load against extraneous; working-memory occupancy rather than fuel.

Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31. Instructional support has to change as skill grows — the nearest public cousin, not a trial of this named condition.

Hagger, M. S., et al. (2016). A multilab preregistered replication of the ego-depletion effect. Perspectives on Psychological Science, 11(4), 546–573. The depletion account that would have made bad load a fuel problem did not replicate.

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. Practice testing and elaborative interrogation as the public cousins of the handles.