12 Streaming
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. 1
| Notes in a row | Name |
|---|---|
| 2 | Double |
| 3 | Triple |
| 4 | Quad |
| 5 | Quint (five-note burst) |
| 5 to 8 | Burst |
| 9 or more | Stream |
| Long enough that stamina dominates (often around 32 or more) | Deathstream |
The subtle part is coordination. With flow aim, your cursor sits over each note only briefly, often for less time than the 300 window lasts. Time each tap to when the cursor is actually on the note, even if that means rushing or delaying slightly: a perfectly timed tap while the cursor is off the note is a miss, not just a lower judgment.
Fact check: Low bpm is harder to hit
- By the numbers. Hit windows are fixed in milliseconds, but the gap between notes shrinks as BPM rises, so the share of each gap that counts as a 300 grows with speed. At OD 8 the 300 window is 64 ms wide: that's 51% of the gap between notes at 120 BPM, 77% at 180, and 102% at 240, where the window is wider than the gap itself. Slow streams demand real timing control; fast ones forgive it, which is why accuracy on slow streams is its own skill and why it exposes finger control that speed hides. 46,67
- Adds detail. Almost's accuracy ladder follows from that: train stream accuracy on simple spacing, raise OD one step at a time from 7 upward once you hold 99%, and pick a BPM floor you actually need rather than grinding 120 BPM streams you'll never play. 67
Fact check
- Nuance. The note-count cutoffs are community conventions, not official definitions. Tonehh puts streams at about 9 to 32 notes and BTMC calls 7 the borderline; everyone who uses them calls them vague. 3,34,73
- By the numbers. 1/4 notes arrive at BPM ÷ 15 per second: 12 per second at 180 BPM and 16 at 240. Alternating splits that load between two fingers; see the stream speed table in Ch. 35.
- Research. Keyboard studies put a single index finger's maximum tapping rate at roughly 5 to 7 taps per second. A 180 BPM stream needs 12 notes per second, so fast streams are close to impossible without splitting the work between fingers. 116
"I can't stream 180" isn't a diagnosis
Asked by a viewer how to push stream consistency past 180 BPM, BTMC's first move is to refuse the question as stated: that sentence contains at least four variables, and which one is failing decides what you practice. 34
| Variable | What it changes |
|---|---|
| BPM | The tap rate: 12 notes per second at 180 BPM, 6 per finger |
| Object count | How long the run lasts, and so whether it tests control or stamina |
| Spacing | How much aim the stream needs. Stacked streams barely move the cursor; wide ones are an aim test at the same tap rate |
| Pattern complexity | Curves, direction changes, and shifts in spacing, which load reading and flow aim rather than tapping |
The four trade against each other. The same run is playable with more objects and wider spacing if the BPM drops, and at higher BPM if the count and spacing drop. Trying to raise all four at once is how a player gets stuck on a number.
Fact check
- By the numbers. At 180 BPM notes are 83 ms apart, so seven of them last about half a second and thirty last 2.4 seconds: identical tap rate, roughly five times the stamina demand.
- By the numbers. Spacing sets cursor speed, not tap speed. At 180 BPM, a stream spaced 50 osu!pixels apart needs an average cursor speed of about 600 osu!pixels per second; stacked at 10 pixels it needs 120. That's why "dense 180" and "spaced 180" are different skills. 100
BTMC’s take
- Seven notes is where a burst becomes a stream. Nuance. One more convention alongside the nine-note one above. The label matters less than his point behind it: the demand changes with the count.
- Push one variable at a time from a base that works. Agrees with the chart. If 180 BPM at seven objects is solid, try 175 at nine, or 185 to 190 at five with constant spacing. That's the challenge point framework applied to four dials instead of one, and it's the same start-from-the-bottom logic khz and Korilak use with BPM alone. 5,31,108
- The variety is what builds the underlying skills. Agrees with the chart. He names finger control, stamina, and reading as the foundation streaming sits on, which is the chart's structure in his own words. 1
Take: jesse1412's thread: Streams
- Split streams into two kinds: long and fast. Agrees with BTMC. A long stream is anything from about 20 to 200 notes; a fast stream is one that's hard to clear from the first note. They need different approaches, which is the object-count dial in Ch. 12 stated in 2012.
- Tense and hook the fingers for fast streams; relax for long ones. Contested. jesse1412 describes a claw-like hooked hand for speed and a relaxed hand for endurance. khz and ferret_irl argue the other way: tap lighter and ground the hand to relieve tension, and khz uses lighter tapping specifically to recover mid-stream. Nobody has tested these against each other; what they share is that both change technique deliberately rather than pushing harder. 5,38,66
- Spam is not streaming ability. Agrees with the chart. One contributor dropped to 120 BPM to stream accurately on keyboard, then reached clean 180 BPM long streams about ten months later. That is khz's ladder and Korilak's five-month climb, from a third player in a different decade. 5,31,66
How hard is a stream? Spacing, in order
Almost's streaming guide ranks stream difficulty by spacing rather than by BPM, which is the single most useful ordering in this book for choosing what to play next. Roughly easiest first: 67
| Step | Stream type | Note |
|---|---|---|
| 1 | Spacing of half a note or less | Barely moves the cursor: a tapping test (Ch. 12) |
| 2 | Curved streams | Flow aim proper begins here |
| 3 | Changing spacing within the stream | Acceleration, not just direction |
| 4 | Between half a note and one note of spacing | Aim load rises sharply |
| 5 | One note of spacing | Each note is its own small jump |
| 6 | Odd shapes: W-shaped, self-overlapping | Reading cost, not aim cost |
| 7 | More than one note of spacing | Spaced streams |
| 8 | A stream with jumps inside it, unbroken | Two skills at once |
| 9 | Irregular streams full of jumps | The top of the ladder |
His recommendation is to spend months with 95% of your playtime inside the first two or three categories, at lower BPM, until misreading one genuinely surprises you. Everything above that is optional until then.
Almost’s take
- Read a stream by looking directly at each circle you have to hit, and let the cursor follow your eyes. Supported. On singles you can snap your eyes to each note; on streams the eyes move more fluidly, but the target is still the individual circle rather than the shape. He adds one concrete tip: lock your eyes onto the very first circle of the stream. That is gaze anchoring, and it matches both Tonehh's individual-note reading and the eye-movement research in Ch. 10. 10,67,141
- Don't consciously manage hand tension. Contested, and the evidence cuts both ways. ScarletStory, khz and ferret_irl all say to relax and tap lightly; Almost argues that tension is a natural response to a strenuous task, that it falls away on its own as the same BPM becomes easier, and that your subconscious regulates it better than deliberate control does. Both halves have support: fingertip force raises carpal tunnel pressure, so lighter really is safer, while consciously monitoring a well-learned movement is a known route to choking. The reconciliation the book offers: build technique and speed so less tension is needed, keep the light-touch habit in slow practice, and don't police your hand mid-run. 67,106,135
- When training top-end speed, keep some low-BPM streams in the session. Supported. At your limit you alternate all-out with no conscious speed control, and players who only do that lose the ability to stream slower maps. That is the specific mechanism behind the "control before speed" rule every source here repeats. 67,114
How the skill grows
Every source that teaches streams uses the same ladder: clean bursts, then longer bursts, then short streams, long streams, and deathstreams. Tonehh places bursts from about 3 to 4 stars, longer bursts at high 4 stars, and first streams from high 4 to mid 5 stars. Three things have to come together along the way: tapping, flow aim, and stamina. 7,8
Tonehh’s take
- Your tapping must be automatic before you can work on flow aim. Supported. Skill learning moves from effortful to automatic, and only automatic skills free your attention for something else; consciously monitoring a well-learned movement can even make it worse. The rank-80 player's roadmap puts the same step, tapping without thinking about it, just before stream aim. 8,45,106,160
- For flow aim, trace the stream's shape, matching its speed changes and curves, or read it in chunks. Experience-based. He found picturing the whole stream as a slider made him over-aim, and reading it a few notes at a time worked better, which fits the chunking research in Ch. 6. 8,101
- Start around 160 to 180 BPM; lower BPM streams demand more finger control, not less. Experience-based. khz (140 BPM) and Korilak (150 BPM) start lower but on long practice maps with a different goal: building control and stamina together. 7
- Deathstream practice maps are benchmarks, not a way to learn streams. Dated. In his 2020 guide Tonehh said to avoid them for learning; in his later video with khz he conceded they work if you have the mindset to grind. 5,7
- Don't use Relax to practice streams. Supported. Relax taps for you, and its leniency lets a flick land notes a real tap would miss, so the one skill you most need to train isn't trained. 8,62
- Keyboards aren't the problem. Supported. Mechanical switches help consistency but aren't required, and polling rate only matters at extreme levels. The numbers are in Ch. 17. 7
ScarletStory’s take
- Freezing mid-stream is anticipatory muscle tension, not a speed problem. Agrees with khz and ferret_irl. You see the stream coming, your body braces for a stressful event, your fingers stiffen, one of them holds a key half a second too long, and you speed up to compensate and mash. That chain explains why the freeze usually happens on the stream you were dreading rather than the harder one you didn't notice (Ch. 14). 5,38,72
- Press the keys lightly; hard presses cost stamina and control. Supported. His test is to type as hard as you can, then as lightly as you can, and compare how long each lasts. khz uses lighter tapping to recover mid-stream, and fingertip force measurably raises carpal tunnel pressure, so the light version is easier on the hand as well as faster. 5,72,135
- The fix is slower streams, not harder ones. Agrees with every stream route in this book. Low BPM teaches how your fingers hand off to each other, which is what stops the lock-up at high BPM (Ch. 31).
Drill: One key, two fingers (scarletstory)
- Bind one key to something wide, a spacebar or shift, and stream a slow section using both fingers on that single key.
- Because the key won't register a second press until the first finger lifts, sloppy overlap simply doesn't sound. You can hear exactly where your hand-off fails.
- Treat it as a diagnostic and a feel exercise, not a playstyle. Return to two keys once the transition feels clean. 72
Take: Rank-80 player
- Don't play stream maps at 4 stars; first learn to alternate, and to stop and restart bursts. Agrees with the chart. 45
- A roadmap by rank: alternation in bursts at 6 digits; long runs of 50 to 90 notes with good UR by about 30K to 40K; tapping without thinking at 4 digits; then stream aim, where most players move too smoothly and should follow each curve sharply; finally speed (Double Time at 260 BPM and up) and high circle size. Experience-based. One top player's path, and rank thresholds shift with every pp rework. 45
Four complete stream routes, bursts up the star ratings, khz's deathstream ladder, Korilak's sets, and the We Luv Lama cuts, are laid out as drills in Ch. 31.
Drill: One variable, one session
- Write your current clean stream as three numbers: BPM, note count, spacing (Ch. 12).
- This session, raise exactly one and hold the other two: 5 BPM up with a shorter run, or two more notes at 5 BPM lower, or the same run with wider spacing.
- Full combo it, then stop. Next session, push a different number. The full routes are in Ch. 31. 34,108
Train this with your own data
This chapter trains speed. Numbers come from your real top plays (see how on your profile).
Finished this chapter?
Log in to track progressMark it complete for 🍪 100 and track your way to Codex Graduate.
Community discussion
Discussion is not evidence: it never changes the chapter directly. Corrections go through Codex suggestions, where they're checked against sources.
Forum threads about this chapterStart one
None yet — longer questions and write-ups belong in the forums, where they stay findable.