Hi, this is Ray.

I want to start with a specific experience that changed how I think about repetition in learning. Years ago, I was trying to memorize vocabulary for a language I was studying. I was using a specific method that seemed obviously effective… write each word 20 times, then move to the next word. It felt disciplined. It looked like serious effort. I spent hours doing this. And when I tested myself a week later, my retention was appalling. Not just imperfect. Appalling. I remembered fewer words than I would have if I'd just glanced at each one once.

At first I assumed I hadn't done enough repetitions. So I tried again with 40 repetitions per word. Same result. Then 60. My retention actually got worse, not better. Something was clearly wrong with what I was doing. But everything about the method aligned with what I'd been told about how memory works. Repetition strengthens memory, right? Wrong practice makes perfect. Right? Except my results said something different.

What I didn't know at the time is that I had stumbled into what researchers call the repetition paradox. Not all repetition helps memory. Some repetition actively hurts it. And the difference between productive repetition and destructive repetition isn't about how many times you repeat… it's about how you repeat, when you repeat, and what your brain is actually doing during the repetition. Once I understood this, my learning transformed in ways that felt almost unfair. Same effort, dramatically better outcomes, just because I'd learned what kind of repetition actually works.

Today's newsletter is about that. What the research actually shows about repetition and learning, why the popular framing of "just repeat it more" is genuinely misleading, and how to actually structure repetition so it produces the retention it's supposed to produce. This is one of those topics where getting the specifics right matters enormously, because most people are using repetition in ways that produce far less learning than the same effort applied correctly would produce. Let's get into it.

Most CX platforms don't own the voice. They orchestrate a workflow, then call a third party for speech and transcription. Every hop adds latency, and latency is what turns a frustrated customer into a churned one.

ElevenAgents is the opposite. Built on the voice models the market already builds on, it runs voice, transcription, chat, and reasoning in one vertically integrated pipeline. Responses come back in under 400 milliseconds and sound human, not synthetic. When a caller gets frustrated, the agent detects it and shifts tone in real time: calm, reassuring, patient.

You keep full control. Plug in any LLM, connect tools, webhooks, and MCP servers, and ground every answer in your knowledge base. Launch in minutes, A/B test with Experiments, enforce Guardrails, and version every change.

More resolved conversations, less infrastructure stitching. Pricing is transparent and flat at $0.08 per minute.

Start building free

Let me start with what the science actually shows, because the picture is more complicated than either "repetition helps" or "repetition doesn't help."

The foundational finding on repetition and time comes from Ebbinghaus's 1885 work. According to research on spacing effects, as early as 1885 the German psychologist Herrmann Ebbinghaus observed in one of his numerous memory experiments that 38 learning repetitions distributed over three days resulted in the same memory performance as with 68 massed repetitions. Read that carefully. Thirty-eight distributed repetitions produced the same memory as sixty-eight massed repetitions. Same brain. Same material. But the distributed repetitions were nearly twice as efficient per repetition. This is the specific finding that started the whole field of research on spacing effects, and it points at something genuinely important. It's not the number of repetitions that matters. It's the timing between them.

Modern research has quantified this dramatically. According to a recent study on multi-session learning, when repetitions were distributed across five sessions rather than massed within one session, participants showed a 1.9 times improvement on retention after 12 h, and a 5.1 times improvement after 36 h. Five point one times better retention at 36 hours. From the same total number of repetitions. Just distributed differently. This is what "practice makes perfect" gets wrong when applied naively. Same practice. Different distribution. Wildly different outcomes.

But here's where it gets genuinely surprising. According to research on the repetition paradox, in results reminiscent of those from semantic-satiation studies, participants who repeated the words for 5 or 10 s were significantly more likely to use them in the recall task, relative to chance level (no repetition). However, words repeated for 20 or 40 s were reported no more often than chance level. Read this carefully. Words repeated for too long, without meaningful engagement, were remembered no better than words never seen at all. Prolonged mechanical repetition didn't just fail to help… it produced results indistinguishable from having done nothing.

The specific finding that resolves the paradox is worth understanding. According to the same research, replicating previous findings, memory for repeated items declined with increased repetition under incidental-learning conditions. However, increased repetition had the opposite effect under intentional-learning conditions. Taken together, these results provide evidence for distinctive mechanisms of memory acquisition during repetition that vary depending on learning context. The distinction is intentional versus incidental. Repetition with active engagement, deliberate encoding, and meaningful attention helps. Repetition without these (mechanical, mindless, going through the motions) can actually hurt memory. This is why my vocabulary approach failed. I was repeating without the specific mental engagement that makes repetition productive.

Let me name the specific mechanisms by which repetition can fail to help learning.

Semantic satiation. When you repeat something too many times without meaningful engagement, the meaning can literally fade from awareness. Say a familiar word 30 times in a row. It starts sounding strange, meaningless, foreign. Your brain has been processing the sound without engaging with the meaning, and eventually the meaning disconnects from the sound entirely. This is bad for memory because you're strengthening the sound while weakening the meaning connection… which is the specific thing you need to retrieve the meaning later.

Automatic deficient processing. According to research, when repetition happens too close together, an automatic deficient-processing mechanism may contribute to the spacing effect. Your brain notices "I just saw this" and reduces engagement with the second exposure. The second exposure doesn't get the deep processing it needs because your brain has essentially decided the item is already handled. This is why massed practice produces the illusion of learning (you feel like you know it because you just saw it) without producing the actual encoding.

Fluency illusion. Repeated exposure produces a sense of familiarity that feels like knowledge but isn't. You've seen the material so many times that it feels obvious, so you feel like you know it. But feeling like you know it is different from actually being able to retrieve it under different conditions. This is one of the specific failure modes of rereading textbooks… you feel like you're learning because the material feels familiar, but you can't reproduce it on an exam because familiarity isn't retrieval.

Attention drift. During long mechanical repetition, attention wanders. You're technically still repeating, but your mind is elsewhere. The repetitions during attention drift produce minimal encoding. This is what my vocabulary method was producing… my hand was writing the words but my brain had checked out.

Missed retrieval practice. Repetition without retrieval doesn't train the specific skill of pulling information from memory. You're re-encoding the input pathway repeatedly without exercising the output pathway. When you need the output later, you haven't trained it. In Metal Gear Solid terms, you can't fight enemies effectively if you've only practiced watching combat, not doing it. Same principle applies to memory. Passive exposure and active retrieval are different skills.

Now the specific features that separate productive repetition from destructive repetition.

Spacing between repetitions. Not zero seconds, not one hour, not one day for everything. Spacing appropriate to the material and your current learning stage. The spacing effect is real, well-documented, and one of the most robust findings in cognitive science. According to research on the mechanism, Increasing the number of learning repetitions improves memory, which is well known. Less well known may be, however, that massed learning is less efficient than spaced learning: two spaced learning sessions can be twice as efficient as two learning sessions without spacing. Same total practice. Different distribution. Substantially different outcomes.

Intentional engagement. Each repetition needs to involve your mind, not just your mechanical actions. What does this mean? Why is this true? How does this connect to what I already know? What example would illustrate this? These questions during repetition transform mechanical review into productive encoding. Repetition without intentional engagement is often no better than no repetition at all.

Retrieval practice, not just re-exposure. As I've covered in the retention techniques articles, active recall dramatically outperforms passive review. Repetition that involves trying to retrieve information before checking it produces different outcomes than repetition that just re-exposes you to the information. The retrieval attempt is what strengthens the pathway back to the memory.

Variation. Repeating the same thing in slightly different contexts, with slightly different framings, or applying it to slightly different examples produces different (better) results than identical repetition. Your brain notices the variations, which forces engagement rather than allowing automatic processing to take over.

Deep processing, not surface features. As I've covered in the deep processing newsletter, engaging with the meaning of material during repetition produces dramatically better retention than engaging with surface features. Repeating "the mitochondrion is the powerhouse of the cell" as a sound pattern is different from actually thinking about what mitochondria do and why they're important. Same words. Different processing. Different outcomes.

Distributed intervals. According to the specific research on this, the pattern that works is expanding intervals… review, then again in a day, then again in three days, then again in a week, then again in two weeks. This pattern captures the specific benefits of spacing while also including enough repetition to actually consolidate the memory.

The core insight worth internalizing is that repetition is a tool, not a substitute for thinking. When you're using the repetition as a container for genuine cognitive work… retrieving information, considering meaning, making connections, generating examples… the repetition helps. When you're using the repetition as a substitute for cognitive work (hoping that mechanical exposure will produce learning without you having to actually engage), the repetition often hurts.

This distinction is dramatic in its practical implications. In Fire Emblem, you don't level up units by having them stand at the back of the battlefield while others fight. They have to actually engage in the combat to gain experience. Your memory works the same way. It levels up through actual cognitive engagement, not through mechanical exposure. Repetition can be the framework for the engagement, but it can't be the engagement itself.

Okay, the practical part. Based on the research and my experience, here's how to actually structure repetition to produce the learning it's supposed to produce.

Space your repetitions. Not "review twice today." Review today, tomorrow, three days later, a week later, two weeks later, a month later. The spacing captures the specific benefits the research documents. This is exactly what spaced repetition software like Anki does automatically, which is why it's such an effective tool for long-term retention.

Use retrieval, not re-exposure. Each repetition should involve trying to retrieve the information before checking it. Not staring at the answer. Trying to produce it, then verifying. The retrieval attempt is what makes the repetition productive.

Engage meaningfully every time. During each repetition, actually think about the material. What does it mean? Why is it true? How does it connect? Even brief moments of intentional engagement transform mechanical review into productive learning.

Vary the context. Don't always review flashcards in the same order, in the same place, in the same way. Vary the conditions of your review. This prevents your brain from encoding based on surface features and forces engagement with the actual content.

Stop when repetition becomes mechanical. The moment you notice you're just going through the motions (turning cards without engaging, repeating words without thinking) stop. Take a break. Come back when you can actually engage. Mechanical repetition often produces worse results than no repetition at all, so continuing when you can't engage is actively counterproductive.

Use fewer high-quality repetitions rather than more low-quality ones. Ten repetitions with genuine engagement produce dramatically better results than fifty repetitions on autopilot. When forced to choose, choose quality over quantity. The efficient learner isn't the one who repeats the most times… it's the one who makes each repetition count.

Match the pattern to the material. Highly meaningful material (concepts you can connect to what you know) needs less mechanical repetition and more thoughtful engagement. Purely arbitrary material (random facts, vocabulary without cognates, phone numbers) might need more repetition but still requires the intentional engagement to produce actual encoding.

Test yourself under different conditions. Not just reviewing in your usual study spot. Test yourself while walking. While explaining to someone. In writing. Verbally. Different retrieval conditions produce different memory pathways, which produces more flexible retention that isn't tied to specific study contexts.

Some honest caveats.

Repetition still matters. The point isn't to avoid repetition… it's to make repetition productive. Zero repetitions produces zero retention. The question isn't whether to repeat but how to repeat well.

Not all learning requires spaced repetition. For material you only need briefly, or for skills that develop through continuous practice, the specific spaced-repetition pattern may not be the best fit. Match the technique to the type of learning.

Individual variation is real. Different learners respond differently to different repetition patterns. The general principles apply, but your specific optimal timing may vary. Pay attention to what actually works for you.

Cramming isn't universally bad. For very short-term learning where you need to perform in the next 24 hours, massed practice can be adequate. It's just terrible for long-term retention. If you actually need to remember material for years, massed practice fails you. If you just need it for a specific short-term application, it can suffice.

Boredom isn't necessarily bad. Some level of tedium during productive repetition is normal. The line isn't between "fun and effective" and "boring and ineffective." It's between "engaged and productive" and "mechanical and unproductive." Boring can still be engaged if you're actually thinking during the boring.

Here's what I want you to take from all this. The popular framing of "repetition is the key to learning" is genuinely misleading when taken naively. Some repetition dramatically helps. Some repetition actively hurts. The specific features that separate productive from destructive repetition matter enormously, and most learners are using repetition patterns that produce far less learning than the same effort applied correctly would produce.

If you've been repeating material and getting frustrated by poor retention, please consider that the issue might not be that you're not repeating enough. It might be that you're repeating in ways that don't work. The intervention isn't more repetitions… it's better ones. Space them out. Add active retrieval. Engage meaningfully. Vary the conditions. Stop when it becomes mechanical. These modest adjustments transform repetition from potentially destructive to reliably productive.

The specific practical move I'd suggest: for whatever you're currently trying to memorize or retain, examine how you're currently repeating it. Is the timing spaced or massed? Are you actively retrieving or just re-exposing? Are you engaging meaningfully or going through motions? Adjust one dimension. Notice how your retention responds over two weeks. Most learners who make these adjustments discover that they'd been working hard for years without getting the returns their effort deserved.

In Chrono Trigger, the party doesn't get stronger by fighting the same enemies over and over in the same spot. They get stronger by fighting varied enemies in varied conditions, letting the experience actually register between battles. Your memory works the same way. Vary the conditions. Space the exposures. Engage each time. The learning that emerges from good repetition is durable in ways the learning from mechanical repetition never becomes.

Keep learning (and keep repeating wisely),

Ray