The osu! Codex

The book has six parts. Part I is the map: the skills, how they feed each other, and where the coaches agree and disagree. Part II takes the skills one at a time, starting from your first hundred hours. Part III covers setup: tablet technique, hardware and accessories, drivers an

DigitalHypno's chart sorts osu! into five umbrella terms: tapping, aim, reading, endurance, and speed . Each branches into narrower skills, and those skills feed one another. Tapping, aim, and reading are the three core inputs of gameplay: clicking, moving the cursor, and making

Read side by side, the sources agree on more than they disagree. Where they split, it's usually over thresholds and routes, not principles.

Everything after this chapter assumes you already play. This one doesn't. It is also the chapter this book exists for: almost every problem in Ch. 32 starts in the first month, with a player skipping the boring part.

Every pattern in the game is built from circles and sliders, so those need to feel natural before harder maps make sense. DigitalHypno compares it to learning a language: nobody tackles complicated sentences until simple ones are automatic. "Mastery" here means comfortable and na

Doing about the same on every attempt. It's the combined result of the three control skills, so the way to improve it is to firm those up rather than chase "consistency" directly. Unreliable control can't produce reliable results.

Staying locked to the beat. It can be built outside osu! too, through an instrument or music theory, and it feeds both accuracy and sightreading.

bloo calls reading the most important skill in the game, and Tonehh's aim guide goes further, putting aim at 99% reading and 1% raw aim. The chart splits reading into pattern processing and rhythm sense, and both feed almost everything else. 1 , 10 ,

Sightreading is playing a map correctly the first time you see it. The chart treats it as a skill built from pattern processing and rhythm sense, and it is the one skill you cannot train by repetition, because the second attempt is no longer a sightread.

"My aim sucks" is rarely a diagnosis. Walk the miss backwards through the steps it passed through, which is DigitalHypno's self-test made explicit:

Skills are what you train. Patterns are what you search for. A player who can say "I lose spaced streams and squares" can find practice maps in five minutes; a player who says "my aim is bad" cannot. The vocabulary below is the community's, stabilized across more than a decade of

A stream is a run of notes in 1/4 rhythm, twice as fast as 1/2 singletaps. On the chart it builds on finger control, tapping stamina, and flow aim.

Nerve control is staying steady under pressure, built from focus and endurance. The core idea, which DigitalHypno credits to a tweet by Bubbleman, is that nerves appear when your attention moves from the task in front of you to the outcome, like realizing mid-run that you're on p

General rule: ease in with simpler maps before stepping up, and never add a mod to make a map easier to read ( Ch. 8 ).

Tonehh's Tablet Beginners series makes one argument three times: grip, area, and hover versus drag are preferences, and comfort decides. His later video on tablet problems adds the part beginners don't expect: finding and keeping a comfortable setup is a struggle for everyone, in

Hardware doesn't play the game for you. Korilak says you need neither tablet nor mechanical keyboard through the 100K to 300K ranks, Tonehh points to top scores set on laptop keys, and switching to a tablet doesn't improve aim by itself. What gear buys is comfort and consistency:

In 2026, community developers released custom firmware that raises the report rate of several Wacom tablets, reportedly with AI help in the reverse engineering. BTMC flashed four tablets on stream. Using OpenTabletDriver's debugger, he measured a stock CTL-472 at 133 Hz and flash

This is the one chapter where being wrong costs more than a rank. Everything here appears elsewhere in the book next to the skill it belongs to; it is collected here so nobody has to assemble it from five chapters while their wrist hurts.

The book's values are for stable unless a box says otherwise, because stable is what most of the sources played on. lazer is where the game is going, and several of its differences change training rather than just presentation.

"Play more" is the stock answer to every improvement question. Korilak's 2016 guide opens by attacking it: it doesn't tell you what to play, and it doesn't separate good practice from bad. Tonehh's answer, years later, is that the phrase is right but compressed: it stands for the

Almost everyone's rank graph flattens at some point. The plateau video argues that raw hours matter less at that stage than mindset, and offers four fixes: play with confidence, keep a collection of maps you have even a 1% chance to full combo, count any score that enters your to

The July 2026 update is worth knowing in outline, because it changed which skills pp rewards and it moved the model closer to the chart this book is built on.

osu!'s own news team runs an interview series, The Followpoint, which has put the same question to several of the game's best players: what would you tell someone starting out? The answers are shorter than you'd expect, and they don't agree with the internet's idea of what top pl

Tonehh now works a nine-to-five, plays four to six hours a week, and still sets top plays, often by accident. His system:

The chart's real use is diagnostic: find the symptom, check the skills that feed it, and train the weakest one. Run the test in the third column before committing weeks to a program.

Diagnosis tells you what to work on. This chapter is the loop that tells you whether the work did anything, which is the part most practice skips.

Ch. 28 asks you to record where the misses happened. This chapter is how you find out, because the results screen tells you what went wrong and only the replay tells you why.

Every source structures a session the same way: warm up, work on one thing, then play for fun or for scores. What they argue about is length.

Each program below is one creator's routine, condensed. Run one at a time. Where two programs target the same skill, pick by temperament: some people like ladders, some like maps they enjoy.

Tonehh's most-watched warning targets newer players whose profiles are full of short, pp-oriented jump maps played at low accuracy. In his experience reviewing profiles, those players hit the same three walls: accuracy that won't improve, an inability to stream or hold prolonged

Tournaments are where the rest of this book gets tested: a mappool you didn't choose, one attempt per map, a clock, and an audience. Most players never enter one. The preparation is still worth knowing, because it is the clearest example of training for a performance rather than

How far off, early or late, a hit can be and still earn each judgement.

Community picks from a 2026 thread, useful mainly as a map of who specializes in what. Read the caveat underneath before spending an evening on replays instead of playing.

The questions people actually ask, and where the answer lives.

September 2026. Built from the creators named in Ch. 1 , community guides going back to 2012, the osu! wiki, and published research. The citation balance and how claims are weighted are described in How to Use This Book.

From the chart. Diagnose weaknesses by checking the skills that feed them. Fundamentals first; low-star maps are the best isolation tool. Memorization is a legitimate tool, but keep actually reading. "Just play more" works because strong players absorb these ideas without noticin

Every source carries a letter showing how far it can be trusted on its own. The letter rates the source , not whether a claim is true: a Tier D creator can be right and a Tier B study can be irrelevant to osu!. What the letter tells you is how much weight a claim carries when not