How I Learned English Using Spaced Repetition

I could read English for years before I could speak it.

I grew up in Ukraine speaking Ukrainian and Russian. English came later, through school, then through code. By my twenties I could read technical documentation, follow long Hacker News threads, and get through entire books in the original without much friction. Then I would get on a call with someone from another country, and the words I needed would sit just out of reach. I knew them. I had read them a hundred times. I just could not pull them up fast enough to finish the sentence I had already started.

That gap, between recognising a word on a page and producing it in conversation, is where most language learning quietly stalls. It is not a motivation problem and it is not an intelligence problem. It is a memory problem, and for a long time I was treating it with the wrong tools.

What eventually closed the gap was not a new course or a better app full of streaks and confetti. It was a boring, repeatable idea borrowed from cognitive psychology, and a schedule I could keep on a bad week. This is the story of how I actually did it, and the planner I built around it, Memset, came directly out of this experience.

The years of input that went nowhere

For most of my adult life my English strategy was simply more exposure. Read more, watch more, listen to more. The logic felt sound. If I just kept consuming the language, it would seep in.

A lot of it did. My reading got fast and my comprehension got deep. But comprehension and production are two different skills, and exposure alone kept feeding the first while starving the second. I was building an enormous passive vocabulary, a wall of words I could understand instantly when I saw them and could almost never summon when I needed them.

This is real and well documented. The linguist Stephen Krashen built a whole theory around comprehensible input, the idea that we acquire language by understanding messages slightly above our current level. Input matters enormously. But input is what gets a word into recognition range. It does not reliably move that word into the part of memory you can reach under the time pressure of a live conversation.

The other thing I tried was vocabulary lists. I would collect twenty good words, write them down, feel productive, and forget seventeen of them within the week. The list was not the problem. The timing was. I reviewed everything once, at the moment I wrote it down, and then never again at the moment it was actually slipping away.

What spaced repetition actually fixed for me

The shift came when I stopped thinking about how much English I was exposed to and started thinking about when I was reviewing it.

The relevant idea is the forgetting curve, first sketched by Hermann Ebbinghaus in the 1880s and since replicated experimentally. Learn something today and within a day most of the detail is gone. Within a week you are down to a sparse outline. The brain discards what does not seem to matter, and it decides what matters mostly by how often the thing comes back.

Spaced repetition is the deliberate countermove. You review a word a day after meeting it, then a few days later, then a week, then a few weeks, then months. Each review is short, and each one resets the forgetting clock from a higher baseline. The word stops being something you recognise and becomes something you own.

When I framed my problem that way, the diagnosis was almost insulting in its simplicity.

I didn’t have a vocabulary problem. I had a review problem.

I already knew thousands of English words well enough to recognise them. What I lacked was a system that brought the right ones back to me at the right moment, just before I lost them, often enough that they crossed over from passive to active. Spacing was the missing piece, and it was a scheduling piece, not a learning piece.

How I set it up for English

The setup that worked was less clever than the failed ones, which is probably why it lasted.

First, I stopped reviewing isolated words and started reviewing things I had actually encountered. When a phrase tripped me up in a meeting or a sentence in a book did something I could not have produced myself, that became a review item. Not “ubiquitous = present everywhere,” but the actual sentence where I met it. Real context survives in memory far better than a dictionary gloss, and it teaches you how a word behaves, not just what it means.

Second, I kept each item small. A short phrase, a single collocation, one sentence. The unit you choose decides how heavy a review feels, and heavy reviews are the ones you skip. Small units make the habit cheap. A review I could finish in fifteen seconds was a review I would still do on a tired Tuesday.

Third, I let the schedule do the thinking. I was not going to manually track when each of a few hundred phrases was due. That bookkeeping is exactly what kills these systems. I needed something that just told me, today, here are the eleven things to look at, and otherwise stayed quiet.

That is the entire mental model I eventually hard-wired into a planner: a reference to the thing you are learning, a rhythm of growing intervals, and a daily list that surfaces on its own. Everything past that is convenience.

The rule that kept it alive

There is one detail I want to dwell on, because it is the reason the system survived past month one when every previous attempt had died.

The danger with spaced repetition over time is that the reviews accumulate. You add new material every week, and the older material keeps coming back for its scheduled passes. A few months in, you can open your reviews and find sixty items waiting. Sixty short reviews is still half an hour, and the first time that happens on a busy day, you bounce off it and quietly quit.

The fix was a daily cap. I set a hard ceiling on how many items I would review per day. If more than that came due, the overflow simply rolled forward to the next day instead of dumping everything on me at once. The schedule flexed around my life instead of punishing me for having one.

This sounds like a small thing. It is the whole thing. The science of spacing has been settled for over a century. The reason people fail at it is never the evidence, it is that the workload spikes and the guilt follows and the system gets abandoned. Capping the daily load is what turns a correct method into a sustainable one.

It also reframed missed days entirely. A skipped day no longer reset anything or buried me. It just meant the next day had a few more items, still under control, still finishable. The drama that usually ends these habits never got a foothold.

What it did not solve

I want to be honest about the limits, because spaced repetition gets oversold as a cure for everything.

It is excellent at moving discrete pieces of language into reach: vocabulary, phrases, fixed expressions, irregular forms. It is mediocre on its own at the things that are really motor skills. Pronunciation needs your mouth, not your memory. Listening to fast natural speech needs hours of actual listening. The rhythm of unscripted conversation needs unscripted conversation. No review schedule substitutes for those.

What spacing did was clear the bottleneck so the rest could happen. Once the words were reliably available, the cognitive load of speaking dropped, and I had spare attention for pronunciation, for listening, for the social timing of a real exchange. The recall layer is necessary but not sufficient, and that is not a flaw of the method. It is just an honest description of what it does.

So I kept the input. I kept talking to people. I just stopped pretending that exposure alone would do the part that only deliberate, well-timed review can do.

What a year of small reviews actually changed

People underestimate what ten quiet minutes a day compound into over a year.

The vocabulary I would have forgotten in six weeks became part of my working speech. Phrases I used to grope for arrived on time. The freeze on calls, the half-second too long that wrecks your confidence and makes you sound less capable than you are, mostly went away. As a software engineer working with people across the world, that change was not cosmetic. It showed up in how I ran meetings, how I wrote, how readily I spoke up.

The strange part is how invisible it is at the start. The first two weeks feel like nothing. The compounding takes a few months to surface, and then one day you notice you said the exact right word without reaching for it, and you realise it has been happening for a while.

If you already read a language well but stall when you have to produce it, you probably do not have a vocabulary problem either. You have a review problem, and that is a solvable one. Keeping that review sustainable, with small units, growing intervals, and a daily cap so a single bad day never breaks the streak, is the entire job Memset was built to do.