Interleaving and varied practice

By Better Wise

You have a spaced schedule and clean flashcards, so the review system exists. Now the question is what happens inside each session: do you drill one thing until it feels smooth, or mix? This lesson turns a blocked practice session into an interleaved one, and it shows where blocking still earns its place.

By the end you can:

• Explain the difference between blocked and interleaved practice
• Convert a blocked session into an interleaved one
• Say when blocking is the right first step
• Vary the conditions of practice, not just the order of items
• Predict why interleaving feels worse while it works better

Quick recall from the flashcards lesson: a good card holds one atomic fact and gets rescheduled by how hard it was to recall. Interleaving applies the same logic at the session level, mixing items so each recall starts from a colder place.

Blocked vs interleaved practice

Blocked practice repeats one type of problem many times in a row. Interleaved practice mixes types or topics within the same session.

Consider a learner preparing a 4-week project on spreadsheet functions. A blocked session looks like this: ten SUM exercises, then ten IF exercises, then ten VLOOKUP exercises. In an interleaved session the same fifteen problems arrive shuffled, so the learner must first decide which tool applies.

That decision is the extra work. In blocked practice the method is already given by the block, so the learner only executes. In interleaved practice the learner must discriminate: is this a lookup problem, a conditional problem, or an aggregation problem? Interleaving trains selection, and selection is the part that fails in real use.

flowchart TD
    A["New skill arrives"] --> B["Blocked practice<br/>get the steps right"]
    B --> C{"Can you do it<br/>without a model?"}
    C -- "No" --> B
    C -- "Yes" --> D["Interleaved practice<br/>mix types and conditions"]
    D --> E["Delayed check<br/>can you pick the right method?"]
    E -- "Confusions remain" --> F["Short re-block<br/>on the confused pair"]
    F --> D

Exercise 1: convert one session

Take this blocked plan for your project and rewrite it as an interleaved one.

20 minutes: 12 flashcards on Module 1 memory terms, one after another. 20 minutes: 8 flashcards on spaced repetition, one after another.

Guidance: pool the 20 cards, shuffle, and run them as one set. Keep a short blocked pass only for cards you marked as leeches. Write your new plan in two lines, then note which card you expect to miss after the mix.

Why the mixed version feels worse

Interleaving produces lower performance during practice and often better retention later. This is the same pattern you met with desirable difficulty: fluency during study is a poor predictor of what survives a delay.

The trap is that mixed practice feels like confusion, so learners abandon it. Two guardrails help:

• Judge the session by a delayed check, not by today's smoothness.
• Keep items related. Mixing statistics with French vocabulary is just a random list. Mixing related problem types, such as mean vs median vs standard deviation, is productive because the items compete.

A useful rule of thumb: interleave within a subject family, not across unrelated ones.

In blocked practice the method is given by the _____, so the learner mainly executes. Interleaved practice adds the demand of _____ which method applies. Interleaving usually feels _____ during practice while improving delayed retention. Interleaving works best when the mixed items are _____, competing for the same decision.

Varying the conditions, not just the order

Interleaving is one form of varied practice. You can also vary the conditions around the same content:

• Different examples: test the same rule on a workplace example and a household example.
• Different formats: recall a concept by writing it, drawing it, or explaining it aloud.
• Different prompts: answer a direct question, then a scenario question that hides the same fact.
• Different contexts: study in two rooms or at two times of day so recall is not tied to one setting.

Variation prevents context dependency, where a fact is only retrievable under the exact conditions of study. If you always review at the same desk with the same notes open, you may be training the setting rather than the memory.

Exercise 2: build a varied mini-set

For one concept in your project, say "the forgetting curve," write four prompts: a definition question, a scenario question, a comparison question against rereading, and one visual task (a sketch or a small chart from memory).

Guidance: do not answer them in a row. Space them across three days, and let at least one prompt arrive without a heading that names the topic. Check which prompt exposed a gap.

When blocking still comes first

Interleaving is not the starting point for everything.

1. Ask: can I perform this procedure once, slowly, with the model open?
2. If no, block it. Do 3-5 near-identical examples to lay down the basic steps.
3. Test: can you do one unaided, correct example?
4. If yes, interleave it with a neighboring type and keep it mixed from then on.
5. If two types keep colliding, re-block that pair for ten minutes, then mix again.

Blocking is the right first step for genuinely new procedures, unfamiliar notation, and safety-critical sequences where a wrong step is costly. It is the wrong long-term default, because it never asks you to choose.

Field notes on mixing

• Common mistake: mixing everything at once, producing a random list with no shared decision. Mix items that compete.
• Common mistake: judging the mix in the moment. Schedule a check two days later instead.
• Common mistake: interleaving before any single type is workable. Block briefly first.
• Rule of thumb: block to learn the steps, interleave to learn the choice.
• Rule of thumb: keep blocked passes short (about ten minutes) and always return to the mix.
• Watch for: confidence rising without delayed accuracy rising. That is fluency, not learning.
• Project tip: label each session in your plan as blocked or interleaved so the mix is deliberate rather than accidental.

What to carry forward

Interleaving trades a comfortable session for a more discriminating memory. Mix related problem types, vary examples, formats, and contexts, and use short blocked passes only to get a new procedure off the ground. Judge the result with a delayed check.

Next up is the checkpoint for this module, where you will audit your review system: schedule, card quality, and how much of your practice is genuinely mixed.

Front: Blocked practice Back: Repeating one type of problem or topic many times in a row.

Front: Interleaved practice Back: Mixing related types or topics within the same session.

Front: What extra demand does interleaving add? Back: Choosing or discriminating which method applies before executing it.

Front: How does interleaved practice usually feel during study? Back: Harder, slower, and less fluent than blocked practice.

Front: What is the typical retention outcome of interleaving? Back: Better delayed recall and better transfer, despite weaker practice performance.

Front: Which items should be interleaved? Back: Related items that compete for the same decision, not unrelated subjects.

Front: When is blocking the right first step? Back: For brand-new procedures, unfamiliar notation, or when no unaided correct attempt is possible yet.

Front: What decision should end a blocked phase? Back: Whether you can produce one correct unaided example, which then unlocks interleaving.

Front: Varied practice Back: Changing examples, formats, prompts, or contexts around the same content.

Front: Context dependency Back: When recall only works under the exact conditions of study, such as one desk or one time of day.

Front: How should an interleaved session be judged? Back: By a delayed check a day or more later, not by how smooth it felt.

Front: What should you do if two types keep colliding? Back: Re-block that pair briefly, then return to mixed practice.

Front: Typical length of a useful blocked pass Back: About ten minutes, followed by a return to the mix.

Front: Warning sign during mixed practice Back: Rising confidence without rising delayed accuracy, which is fluency rather than learning.