Should You Finish One Deck Before Starting the Next?

Mixing several decks usually beats finishing them one at a time, and the reason is that mixing forces you to work out which word is being asked for rather than riding a pattern. But the benefit has conditions. It shrinks or reverses when the material is brand new, and it goes the wrong way for pronunciation.

The question underneath

Anyone with more than one deck has run into this. You are halfway through a business-English set, an exam list is waiting, and something in you wants to finish the first one cleanly before opening the second. Starting both feels like a failure of discipline.

It usually is not. It is closer to the better default, but the reasoning matters more than the verdict here, because the exceptions are not exotic.

Two terms make the rest of this easier. Blocking is studying one category at a time: all of deck A, then all of deck B. Interleaving is mixing them, so consecutive cards come from different sets. Cognitive psychology writes these as AAA/BBB versus ABC/ABC.

What each one actually trains

The clearest framing I have found is that the two do different jobs rather than one being a better version of the other.

Blocking helps you notice what a group of items has in common. Interleaving helps you tell items apart. Both are things you need, but they are not the same thing and they are not needed at the same moment.

This explains the part that confuses people. When you study one deck straight through, you are not really retrieving each word from scratch. You are retrieving it from a context you already have loaded: you know the theme, you know roughly what kind of answer is coming, and the previous card primed you for this one. Performance inside the block looks good.

Then the exam or the conversation arrives with no context attached, and the support you were leaning on is not there.

Where mixing wins

Interleaving generally produces better long-term retention and better transfer, across a lot of learning domains, not just vocabulary. The effect shows up immediately and it holds at a delay.

For vocabulary specifically, the win is concentrated somewhere useful: words you are likely to confuse with each other. If two words occupy nearby territory, seeing them in the same session forces the distinction to be made rather than assumed. Studying them in separate blocks lets you keep a fuzzy sense of each and never test whether you can actually tell them apart.

That is the practical version of the finding. Mixing is most valuable exactly where your errors are going to come from.

A concrete case. Suppose you have met imply and infer in the same week. Studied in separate blocks, each one gets its own moment of attention and each one feels learned, because nothing in the session ever asks you to choose between them. The distinction is the whole difficulty of that pair, and a blocked session never poses it.

Put them in the same session and the question changes from “what does this word mean” to “which of these two is this.” That second question is the one real usage asks. The same applies to any set that occupies neighbouring ground: discreet and discrete, affect and effect, continual and continuous.

The general rule that falls out of this: if you can imagine yourself hesitating between two words, they belong in the same session, not in tidy separate ones.

Where blocking wins

Two exceptions are worth knowing because they are common rather than edge cases.

The first is pronunciation. One study on second-language pronunciation found the opposite pattern: blocked study outperformed interleaved practice, which the authors put down to the task depending on noticing similarity rather than contrast. That is a single finding rather than a literature, so hold it loosely, but it points the same way as the reasoning does. If you are drilling a sound, hearing the same sound repeatedly is the point. Mixing dilutes it. Our notes on improving pronunciation sit on this side of the line.

The second is material that is genuinely new. A 2025 study in the journal Language Learning looked at interleaving with lower-achieving adolescent learners and framed it as an undesirable difficulty rather than a desirable one, arguing that an initial blocked phase matters for building declarative knowledge before mixing helps.

That second one deserves emphasis, because it is the case most people are actually in when they meet a new deck. You cannot discriminate between things you do not yet recognise individually. Mixing too early just produces confusion with no learning attached to it.

The order matters more than the choice. Interleaving asks you to tell items apart, which presupposes that each item exists in your memory as a separate thing to begin with. On a genuinely first pass, it does not yet. What you get instead is a session where every card is unfamiliar and nothing accumulates, which is discouraging in a way that has nothing to do with the method being wrong later.

So the two are better understood as a sequence than as a choice. Block to get items into memory, then mix to make them distinguishable. Choosing one for all time is the actual mistake.

The condition people skip

There is a third finding that changes how you should organise decks, and it gets left out of most summaries.

Interleaving works best when the mixed topics are similar to each other and the individual examples differ. When the topics are too dissimilar, the benefit can reverse, and vocabulary is specifically named as a place where this happens.

The reason follows from what interleaving trains. If it works by forcing discrimination, then it needs things that are actually worth discriminating. Mixing medical terminology with phrasal verbs does not train anything, because no one was going to confuse those. You have just made the session feel harder without buying anything.

So the instruction is not “mix everything.” It is closer to “mix things that could be mistaken for each other.”

Two examples, one on each side. A deck of formal writing verbs mixed with a deck of everyday spoken equivalents is a good pairing, because the sets overlap in meaning and differ in register, which is precisely the judgement you have to make when writing. A deck of medical terminology mixed with a deck of travel phrases is a poor pairing, because nothing in either set competes with anything in the other.

Both sessions feel similarly effortful. Only one of them is buying you anything, and effort is not the thing to optimise for.

Which to use when

Situation	Use	Why
First pass through unfamiliar material	Blocking	You cannot tell items apart before you recognise them
Words you keep confusing with each other	Interleaving	The confusion is the thing being trained away
Drilling a sound or a stress pattern	Blocking	The task depends on hearing similarity, not contrast
Two decks in unrelated fields	Either	Mixing them buys little, so pick by convenience
Reviewing anything you have seen before	Interleaving	Removes the context you were leaning on
A week before an exam	Interleaving	Exam conditions supply no context either

The middle row is the one that gets misapplied most often. Unrelated decks are exactly where mixing has the least to offer, and they are usually the decks people feel guiltiest about running in parallel.

What this means for your decks

The practical shape:

1. Take a new deck through one blocked pass. Enough to recognise the words on sight, not to know them well
2. Then let it join the rotation. From this point the words should be arriving mixed with everything else
3. Deliberately pair decks that overlap. If two sets contain words that live near each other, running them together is the highest-value mixing you can do
4. Keep pronunciation drills separate. Different task, different rule
5. Do not reshuffle daily. Interleaving is about what sits next to what, not about constant reorganisation

Step 3 is the one worth actually planning. Most people mix decks by accident, based on what they happened to start. Choosing to run two confusable sets together is a decision with a real payoff.

If you want a single change to make today, it is this: look at the last ten words you got wrong and check whether any of them were wrong because you confused them with a specific other word. Those pairs are the ones to deliberately put together. Errors of that kind are the visible edge of a distinction you have not made yet, and they will keep recurring until something forces the comparison.

Why the better method feels worse

Worth naming, because it is the reason people abandon this.

Blocked study feels good. You get answers right, the session moves, and you finish feeling like the deck is going in. Interleaved study feels choppy and you make more mistakes in the moment.

Both of those impressions are accurate about the session and wrong about the outcome. The fluency inside a block comes partly from the block itself, so it does not survive contact with a context that is not the block. This is the same trap as rereading notes, which feels productive precisely because it is easy.

There is a measurement problem inside this. The signal you naturally use to judge a study method is how well the session went, and that signal is systematically biased toward blocking. Blocking improves performance now at the cost of performance later, so the method that looks best while you are using it is the one that leaves you with less afterwards.

The honest correction is to judge by something other than the session. How well did you do on these words a week later, in a context you did not choose? That is slower feedback and much less satisfying, which is why almost nobody uses it.

I find this genuinely hard to act on. Knowing that the harder session is the better one does not make it pleasant, and on a tired evening I still catch myself wanting to run one deck straight through because it will feel like progress.

What your app is probably already doing

If you use spaced repetition, you may be interleaving without having decided to. Due cards arrive from every deck at once, which mixes them as a side effect of scheduling. Spaced repetition answers when a card returns; interleaving is about what it returns alongside.

Vocaby works this way. Its FSRS scheduler pulls whatever is due, so a review session is mixed by default, while the curated decks give you the blocked unit for a first pass. Both halves are available and the choice of when to use which is yours. You can see how the words are grouped on the word list.

The one thing worth changing deliberately is step 3 above: put confusable decks together on purpose instead of letting the mix happen by accident.

Get Vocaby on the App Store