I’ve been trying to get better at Osu. Here’s what I’ve learned:
- There are (extremely broadly) 4 key skills to any game: observing the environment, world-modelling, planning next-actions, and executing those actions. Getting good at Osu is about eliminating the middle steps and streamlining the outer steps - there is no uncertainty to the play: it is just faster and faster clicking of targets. You are trying to Maximise Task Throughput without end.
- There does not seem to be any concrete skill cap - the difficulty continues to rise on a continuum, where legible progress can always be made. The limitations are often the frame rate of your computer and the input delay on your peripherals.
- Developing expertise results in not registering individual targets, but collections of targets as chained motions, similar to chords or movements, or reading words rather than individual letters. The base unit ceases to be the Object and becomes the Pattern; then the Patterns are chained together into increasingly large or complex combinations. There is a fractal complexity depending on your minimum unit - migrating to larger base units can facilitate step-changes in play capability.
- There are Patterns which have no parallel at lower levels of play because they hinge on rapid timings or precise motions which, when slowed down, are qualitatively “less Pattern-ish”. A rapid star pattern forces you to register it As Gestalt, because it’s too rapid to register sequentially as “5 targets”. A burst is a rapid chain of targets at a specific rhythm; again, these structures become Patterns when they are registered as a single motion.
- Most higher-level play is two-handed: one hand on a pointing peripheral like a mouse or tablet, the other hand on a beat-tapping peripheral like a keyboard. This is interesting since it’s a kind of parallel automaticity that you can get in split-brain cases - at high-level play there’s likely parallel subsystems at work.
- Analog keyboards (which I’d mostly seen as a gimmick previously, or purely used for analog input) allow for dynamic triggers instead of fixed-depth ones. As such, rather than a keypress being actuated at a specific depth (as is usually the case with mechanical keyboards), the keypress can start immediately as the key moves downwards, and end immediately when the key moves upwards. This allows for a higher maximum speed of rapid actuation, since keypresses are velocity-based rather than displacement-based. (Which makes me wonder a bit about the prospect of an acceleration-based keyboard.)
- It is difficult to balance intentional, diagnostic practice with the end goal of automaticity. When proper automaticity is reached, performance moves from a System 2 to a System 1 task: the pattern is executed subconsciously faster than the conscious mind can feasibly process the input. The downside is that improvement at a System 1 task is difficult, since you’ve driven it into a more intuitive than analytic/diagnostic task performance space.
It has been incredibly interesting to observe from-the-inside how my brain is improving at the game: a lot of the fun I’m having isn’t so much the “game in itself” but using it as a case study in skill acquisition.