Interleaved spaced retrieval

Practise mixed topics on a schedule that also spaces each item across days, selection and timing together.

When to use this: While you practise
Learning impact: Highest-leverage move
Learning evidence: Strong research tradition
Where you are with the topic: Strengthening memory
How often to use it: Build and keep as a habit

When to use it

You already space items, but each day’s pile is still one topic blocked together.
You already mix types, but only inside a single cram evening.
You want both constraints at once: mixed and spaced.
Exams mix topics across weeks of content, and your practice should too.

What to ask AI

Type something like: “From my tracker, pull what’s due today and shuffle topics into one mixed retrieval set. Cues only. No answers until I finish. Then propose updated intervals.”

Ask it not to group all biology then all maths, truly shuffle.

Feed outcomes so spacing and mixing both update.

How it helps

Spaced schedule design decides when an item returns. Interleaving decides how items sit beside neighbours.

Doing both at once trains timing and selection together, closer to real exams.

It feels harder than blocked weekly revision. That difficulty is often the point.

Without outcome logging, it collapses into a random mixed quiz.

How to run it

Keep a spaced tracker per item (last outcome, next due, interval).
Each session, pull what is due and shuffle across topics.
Retrieve with cues or free prompts, no type labels.
Log outcomes; update intervals.
Refuse to re-block the set by subject for comfort.

Example exchange

You

Here is my tracker. Can you build today’s due set mixed across topics, give cues only, and wait until I submit everything before updating intervals?

Four due cues, mixed. 1) Derivative of sin(3x). 2) What NADPH carries to the Calvin cycle. 3) Necessary versus sufficient in one sentence. 4) Limit of (sin x)/x as x approaches 0. Send answers in order.

You

1) 3 cos(3x). 2) Reducing power. 3) I said “needed” and “enough” but muddled which is which. 4) The limit is 1.

Logged. Necessary versus sufficient failed, so that item is due tomorrow. Chain rule was easy, so we can stretch its interval. Keep tomorrow’s set mixed across topics too?

Copyable prompt

Here is my spaced tracker: [PASTE].
Pull items due on [DATE], shuffle across topics (no blocking by subject), and
give retrieval cues only. After I submit outcomes, propose updated intervals
and next dues. I approve before locking.

The Tell

You will know the method has flipped when you take the mixed due list and quietly re-sort it back into subject blocks before working.

Then you kept spacing maybe, but you threw away interleaving.

Leave the shuffle intact and retrieve in the uncomfortable order.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

Combining spacing and interleaving applies two well-supported practice structures at once. Each has evidence alone; together they match many authentic test conditions.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI assembly of due mixed sets is a convenience layer. Schedule optimality and shuffle quality remain unproven beyond the underlying spacing and interleaving principles.