I used to watch a lot of speedrunning videos, and I wondered whether any lessons can be translated from speedrunning to Go.
First of all, what is speedrunning? Speedrunning is, classically, an attempt to complete a video game, or a subsection of one, or a self-imposed goal in one, in the shortest possible time. As a mature sport it’s arguably existed for around twenty years, having its roots in early video games like Super Mario Bros (1985) and Doom (1993).
I think speedrunning strategy can be described as the product of five parts:
general good play
The first top speedruns of any game are almost purely general good play. General good play is a way of play that isn’t bad in any respect – movement and routing are fairly efficient and no large mistakes are made. General good play is the foundation of any competent speedrun.
The meta is the setup, akin to an opening in Go. It consists of one’s choice of character, vehicle, weaponry, etc. etc. – for instance, a certain car in a Mario Kart game may be quick but poor at handling, allowing better results than a slower and more stable one but only with more skill.
The patterns are definitions of RNG (random generation). They mainly include where enemies will be placed in the level, and where they will be moving.
Tricks are high-skill manoeuvres, often exploiting glitches. Tricks are risky, often providing advantage if successfully pulled off but a disproportionately large disadvantage if failed.
Grinding is practice: attempting the speedrun again and again and again until the desired time is achieved. Specialists can spend thousands of hours practicing a single game or even level.
So, what can be learnt?
The underpinning of strong Go is general good play, an undistinguished but consistently competent flow of decent moves.
Tesuji, difficult joseki, and tactical flourishes tend to produce a small advantage if successful and a large disadvantage if unsuccessful, which is not to say that this should put one off playing them.
Some element of pattern understanding, of joseki and certain shapes, is necessary.
No matter how many resources one has, it may be impossible to escape doing some grinding – recall the proverb “lose your first hundred games as quickly as possible”.
At first, not knowing what you meant by speed-running, I thought, “isn’t all running supposed to be speedy, by definition!?”
Then I realised, I used to do speed-running! I just didn’t know it was called that!
My speed-running favourites:
Super Mario land on the game boy. There’s some really fast ones on YouTube. I wasn’t that fast but pretty chuffed that I could regularly beat the game in one sitting. If you immediately play it again without restarting is different and harder the second time!
Tony Hawk’s underground, Moscow level (online play) exploiting the darkslide bug (grind balance resets whenever you do a darkslide) to pull insane 10 minute combos! Land it and rack up a score of billions or try to graffiti tag every obstacle in the level. Bail and you get nothing. Eventually I had to give up because it was too frustrating
Guitar hero on expert in the end, I decided real guitar was more fun
As for how this relates to Go…
Practice is obviously crucial.
Other than that, I find it a little hard to apply the analogy because those games are largely the same every time but Go has the problem of human opponents and the fact that every game is different…
Maybe you could play the same opening / joseki for 50 games in a row to grind that
Haha! Yeah. Ok then, ultimate OGS speedrunning challenge: how fast can you get to the top of the 19x19 ladder but you can only challenge the player immediately above you, i.e. you have to beat EVERYONE!