Blocked practice feels more effective than it is

If your accuracy falls when you move from Classic to Rating, that does not by itself mean you got worse. The practice format changed. Classic uses a blocked curriculum; the competitive bank is mixed; and there is a body of research saying the mixed version tends to feel worse while you are doing it.

What the two words mean#

Blocked practice groups similar problems together. Ten questions about the opening range, then ten about VWAP. You know what kind of problem you are looking at before you look at it, because the chapter heading already told you.

Interleaved practice mixes them. Opening range, then VWAP, then a mid-session drift, then another opening range. Nothing tells you which kind you are facing.

That difference sounds administrative. It is not. In blocked practice you are executing an approach you have already been handed. In interleaved practice you have to pick one, and working out what kind of problem it is becomes part of the problem.

What the studies actually found#

⚠️ Two vocabularies get used for closely related ideas. Kornell and Bjork called their conditions massed and spaced — and in their design the spaced examples were interleaved among other artists, which is why the study is cited in both literatures. Rohrer and colleagues compared blocked and interleaved practice directly. Spacing and interleaving are related but not synonyms, and the studies below are not testing an identical manipulation.

Kornell and Bjork gave people six paintings by each of twelve artists, then showed them new paintings and asked who painted each one. Spacing won in both versions of the experiment: 61% against 35% in Experiment 1a, and 59% against 36% in Experiment 1b.

Rohrer, Dedrick and Burgess ran a blocked-versus-interleaved comparison through a real classroom for nine weeks, with 140 seventh-graders and an unannounced test two weeks after the last assignment. Interleaved practice scored 72% against blocked practice's 38%.

The authors of the painting study are careful about how far this goes, and so should we be. Their own conclusion:

Our results cannot necessarily be generalized to all of these situations, of course, but they do suggest that in inductive-learning situations, spacing may often be more effective than massing, even when intuition suggests the opposite.

"May often", not "always". The honest version of the claim is narrow: when the eventual task requires you to work out which kind of problem you are facing, mixed practice has good support. That is narrow enough to be defensible and still exactly the situation a chart puts you in.

The finding that matters most for a practice app#

Here is the sentence from the painting study worth pinning to the wall. It describes Experiment 1a:

78% of the participants did better with spaced presentations than they did with massed presentations, but 78% of the participants said that massing was as good as or better than spacing.

Same number, opposite direction. Nearly four in five learned more from the harder format, and nearly four in five walked out believing the easier one had worked better — after taking the test that proved otherwise.

That is not a curiosity. It is why blocked practice is hard to give up: it produces a feeling of fluency while you are inside it, and that feeling is what people use to judge whether practice is working. The feeling and the result came apart, in the same room, on the same afternoon.

Where this lives in SwipeTA#

Classic uses a blocked curriculum structure. It is 500 levels of three questions each, arranged in eight chapters, and each chapter leans on one idea:

"Trap share" is the proportion of questions in that chapter where the naive rule — fade the last candle — gets the answer wrong. It climbs on purpose.

To be precise about it: this is not textbook blocking. A chapter is not the same chart forty times. Every level still draws different symbols, different sessions and different setups; what a chapter holds constant is the idea being emphasised. It is much more blocked than the competitive bank, which is the comparison that matters here.

Chapter eight removes the hint#

The important change in chapter eight is not difficulty — its trap share is 50%, lower than chapter seven's 71%. It is identification. Chapters one to seven tell you what kind of thing to look for; the graduation chapter has no topic preference at all, so what was taught separately comes back mixed and unlabelled.

That is the blocked-to-interleaved transition, sitting in a file we generate. It exists because a chapter-seven graduate can spot a trap when the chapter is called "Traps", and that is not the skill anybody needs.

The competitive bank is mixed from the start#

Of the 10,000 published questions:

Add both sides of VWAP — 5,171 above and 4,829 below — three range zones (4,236 mid-range, 3,089 near the day's high, 2,675 near its low) and 22 symbols from SPY to PLTR, none of them more than 479 questions. No single one of those axes is meant to give you a reliable shortcut instead of reading the chart: "morning, mega-cap, near the high" does not narrow the answer, because the next question will not be any of those things.

So which one should you open#

Blocked first, on anything genuinely new to you. If you do not yet know what VWAP does to a chart, meeting it repeatedly in one chapter is how you find out. That is what chapters one to six are for.

Then mix, and expect the number to fall. A chapter-three accuracy was earned in a setting that told you what you were looking at, and it is not a like-for-like comparison with a Rating round.

The failure mode is staying in blocked practice because the number is nicer there. Kornell and Bjork's participants would have chosen exactly that, and they had just taken a test proving it was the wrong call.

Interleaved practice is not random practice#

Mixing only helps once there is something to mix. A shuffled stream of questions you have no framework for is not interleaved practice — it is noise, and it teaches at the pace of trial and error.

The order is the point: blocked long enough to build the categories, mixed to make you choose between them. That is why Classic exists at all, and why it has chapters rather than a difficulty slider.

What this does not mean#

These studies are about painting styles and mathematics problems. Nobody has run them on chart-reading, including us. We have not measured whether mixed practice transfers better for our own players, and until we have, the honest position is that we are applying a robust finding from adjacent domains — not reporting one from this one.

Nor is any of it a claim about trading. What the research supports is better discrimination and retention on a delayed test. What happens with money at risk is a different question, in a setting none of these studies touched.

How the questions themselves are built, filtered and judged is written out in full on the methodology page.