What to expect

Most trainers train one thing. This trains three.

Nearly every brain trainer you can download does one of two things. Either it’s a single exercise — usually n-back — with one difficulty slider. Or it’s a box of thirty small games that are enjoyable for about a week.

Tanren Sensei has fourteen exercises across three kinds of thinking, all running on one engine that finds your level and keeps you there.

Memory — holding several things in mind at once and keeping them current while more keeps arriving. Four exercises.

Reasoning — working out what actually follows from what you know, and stopping there. Seven families of puzzle, each closing off a different shortcut, plus a visual search that makes you hold a target in mind while you hunt for it.

Speed — taking in more of what’s in front of you, in less time. Two exercises: a flash you see, and a pair of quick slides you hear.

We don’t know of another trainer that covers all three properly, and the breadth is the whole idea. Train only one thing and you get very good at one thing.

Your first five minutes

You pick a door — memory, reasoning or speed — or let Guided Play choose for you, which is what we’d suggest at the start.

The first few problems will feel easy. That’s deliberate: the app is watching how you do and working out where to put you. Within a couple of minutes it has found a level, and from then on it stays just ahead of you.

There’s nothing to configure. No level to guess at, no difficulty menu, no settings you’ll wonder whether you got right. A session is short — a few minutes — and ends with your score and the sum that produced it.

About that 80% — it’s a setting, not a grade

The app aims to keep you at around 80% correct, and it’s completely fair to read that as “one in five wrong — that’s not accurate at all.” It isn’t a report card. It’s a dial, and it’s set there on purpose.

Think of a coach setting the speed on a treadmill. Too slow and you’re strolling — comfortable, and you get nothing from it. Too fast and you’re just clinging on. The pace worth training at is the one you can only just hold.

80% is that pace. If you’re getting everything right, the app has made it too easy and is wasting your time. If you’re getting half right, you’re guessing. So it keeps nudging the difficulty up and down until you’re sitting right at the edge of what you can do — and then quietly keeps making it harder as you get better.

So the number that matters isn’t your accuracy. It’s the difficulty you can hold at that accuracy. That’s what the app tracks and shows you, and it’s why this graph can’t flatter you: doing well doesn’t raise the number, it makes the next session harder.

Your first week

Expect a jump. Everyone gets one, and it’s real — it’s mostly you learning what the exercises are asking for, which is a genuine skill and worth having.

You’ll also find out quickly which door you like. Most people take to one of the three immediately and find another one frustrating for a fortnight. That’s normal and it’s a reason to keep the frustrating one in the rotation rather than drop it.

Two sessions on most days is plenty. Ten or fifteen minutes. This isn’t an app that wants an hour of your evening — and the streak forgives one missed day a week, so an ordinary bad Tuesday costs you nothing.

Your first month

The early jump flattens out. That’s not you plateauing — it’s the app having found your real level, and from here the difficulty climbs slowly rather than in leaps. The number to watch is the one that says how hard the problems you’re holding at 80% have become, not your accuracy.

Somewhere in here, if it’s going to happen for you, is when people start reporting the outside-the-app stuff: reading feels easier, a conversation across a noisy room takes less effort, you can hold a few things at once without one of them slipping. We’re not promising that. We are, unusually, giving you a way to check it.

Why it keeps working after the novelty wears off

Three things, and none of them are about making it unpleasant.

Every puzzle has exactly one right answer, and we prove it. Before a puzzle ever reaches you, a second program tries to solve it on its own. If that program can find more than one defensible answer, the puzzle is thrown away and you never see it. You will never lose a point to a question that wasn’t fair.

You can’t win by reading the pictures. Sometimes a puzzle gives itself away by accident — the right answer happens to be the biggest shape, or always sits in a certain spot. That’s a real problem in apps like this, because people then improve at spotting the tell rather than at thinking. So an automatic check hunts for those giveaways and throws out anything that leaks. It has caught five in our own app, including one that practically printed the answer on the page.

It changes shape, so you can’t settle in. Difficulty here isn’t one slider. There are twenty-three separate things the app can turn up — how much you have to hold, how fast it arrives, how tangled the reasoning gets, how convincing the wrong answers are — and only one or two move in any session. A strategy that carried you last week quietly stops working. That’s not there to frustrate you; it’s what stops the training becoming a routine you can sleepwalk through.

And you get the instruments to check it

This is the part nothing else in this category gives you.

Five tests come with the app. You never practise them, their difficulty never changes, and they never feed your training. So a score from today and a score from next month mean the same thing — which is not true of the exercises themselves, and not true of the number most apps show you.

Four of them measure the kind of thinking the training goes after. The fifth measures 3D mental rotation, and we deliberately don’t train it. It’s a control: if it climbs just as much as the other four, the honest reading is that you got better at taking tests rather than at thinking. You’ll see that yourself, in the app, without having to ask us.

The honest state of the evidence, both sides →

Start whenever you like

Sit the tests first, before any training — that first sitting is your baseline. Five short tests, about 25 minutes, and each one opens with two warm-up items that don’t count. Then train: the next check-in comes due once you’ve done 200 minutes of training and at least a week has gone by. At ten minutes a day, that’s about three weeks.

If it works for you, you’ll have the evidence in your own hands. If it doesn’t, you’ll know that too — and you’ll have spent a few weeks on something a great deal more interesting than most things on your phone.

Download for Android → · How it all works →