Way to Go
So now I'm learning the ancient game of Go - not how to play (the rules are really simple) but how to play at a decent standard (okay, I'll come back to that). It's an unexpected turn of events because I left Chess and other two-player "combinatorial abstracts" behind a long time ago - in my mind they were too long, serious, competitive, and unforgiving (a mistake loses you the game). By contrast, Race for the Galaxy, Dominion and so on are shorter, less in-your-face, and have smaller swings.
So what happened? After getting bored with Cribbage on my phone, I tried my hand at Go. What attracted me was the idea that the very best AIs on heavy-duty hardware, get wiped out by so-so human players, despite searching millions of positions a second and simulating the consequences of countless possible futures. In other words, the game is more friendly to human ways of thinking (pattern recognition, fuzzy judgement, and longer-term planning) than automated approaches (exhaustively reading five turns ahead, and analysing according to fixed rules). I'd rather study a game that computers don't dominate, and that human senses excel. Go is also a more strategic game than Chess (which is mostly tactical), in that your moves have much longer-term effects, and their value usually cannot be judged immediately or easily.
"Lose your first hundred games as quickly as possible," is a Go proverb, and I decided that I could do just that on my phone. I expected the annoying tactical traps that lose you lots of stones if you overlook them, but that's what "undo" was for. By repeatedly playing very quick (2 minute) games on a small board with a large handicap (Go provides for clean handicaps), I'd build up a sense of the game without having to devote effort to the task. This I could do in free moments travelling or waiting. What I quickly found was that fighting the computer head-on was futile; AIs excel at the kind of direct tactical play that ends with you losing a tonne of stones. Instead, by building walls and not being drawn into fights, I found myself winning (with handicaps); the kind of AIs used by Chess and Go programs do not perform strategic or goal-based planning.
As I reduced my handicap, I found that some walls worked and some didn't, and had to look up the rules to learn why (Sensei's Library is a fantastic resource for all things Go). That's how I found out about the invincibility of "two-eyed" structures. I bumped up the size of the board (from 9x9 to 13x13) and tried to focus on avoiding tactical pitfalls while making those winning strategic moves, which I thought would be easier when there's more space. After a while, I started winning consistently with a 3 stone handicap and got a bit addicted.
I usually avoid studying games, because I see learning as part of the experience, but there's so phenomenally much to learn in Go that I'm perfectly happy to read books about it. Most boardgames are far shallower, so reading a book would almost solve the game for you. As a result, I learned how to approach indirectly, and to seal off my areas, and I started winning with a 2-stone handicap. I'm currently playing on a 13x13 board with no handicap, losing the majority of games but working my way through beginner Go texts. I don't know any openings ("fuseki") or locally optimal sequences ("joseki"), and don't plan to spend any time learning them until I have a solid grasp of the fundamentals (which is still some time away). Apparently you can become a very strong amateur-level player knowing none of either.
I've played a couple of games online against real humans, and expect that I'll play more.
So what about my criticisms of combinatorial abstracts, those luckless perfect information games? Against an AI, I can play a game quickly, undo bad short-term moves, and ponder when I need to. Go is taken seriously, but as handicapping recognises a pre-existing skill difference, there seems to be less of a competitive air.
However, I think the real difference is that I have less of an ego-attachment to winning the game. Most modern boardgames make losing less painful:
* you have excuses - luck was against you, you were ganged up on, the winner had help, or you were still learning the system
* you tend not to lose things or be attacked - most Eurogames don't allow you to be eliminated or wiped out, and limit direct attacks
* you have a score rather than just a win or loss
* there are lots of colorful things that happened even if you lost - you built up a farm or an economic engine, or led many fantasy races to their fate
* they are shallower - with a few plays, I'm usually competitive against veterans, or know what I need to do to get there
I usually feel like I'm learning rather than just trying to win, which is a much healthier mindset to approach the game.
The standard criticisms of Go boil down to variations of being dry, or requiring study to be competitive. The first is pretty subjective, but tends towards lacking a narrative and requiring brain-work. I'd argue that once you know the game these actually lessen - there are peaceful and combative styles, traps, attacks, invasions, evasion and defence, and Go is more intuitive and less calculating than many other games. The second is not really true - like any other game you can improve just by playing, though given the depth of the game the possibility of walkover matchups is inevitable.
So what happened? After getting bored with Cribbage on my phone, I tried my hand at Go. What attracted me was the idea that the very best AIs on heavy-duty hardware, get wiped out by so-so human players, despite searching millions of positions a second and simulating the consequences of countless possible futures. In other words, the game is more friendly to human ways of thinking (pattern recognition, fuzzy judgement, and longer-term planning) than automated approaches (exhaustively reading five turns ahead, and analysing according to fixed rules). I'd rather study a game that computers don't dominate, and that human senses excel. Go is also a more strategic game than Chess (which is mostly tactical), in that your moves have much longer-term effects, and their value usually cannot be judged immediately or easily.
"Lose your first hundred games as quickly as possible," is a Go proverb, and I decided that I could do just that on my phone. I expected the annoying tactical traps that lose you lots of stones if you overlook them, but that's what "undo" was for. By repeatedly playing very quick (2 minute) games on a small board with a large handicap (Go provides for clean handicaps), I'd build up a sense of the game without having to devote effort to the task. This I could do in free moments travelling or waiting. What I quickly found was that fighting the computer head-on was futile; AIs excel at the kind of direct tactical play that ends with you losing a tonne of stones. Instead, by building walls and not being drawn into fights, I found myself winning (with handicaps); the kind of AIs used by Chess and Go programs do not perform strategic or goal-based planning.
As I reduced my handicap, I found that some walls worked and some didn't, and had to look up the rules to learn why (Sensei's Library is a fantastic resource for all things Go). That's how I found out about the invincibility of "two-eyed" structures. I bumped up the size of the board (from 9x9 to 13x13) and tried to focus on avoiding tactical pitfalls while making those winning strategic moves, which I thought would be easier when there's more space. After a while, I started winning consistently with a 3 stone handicap and got a bit addicted.
I usually avoid studying games, because I see learning as part of the experience, but there's so phenomenally much to learn in Go that I'm perfectly happy to read books about it. Most boardgames are far shallower, so reading a book would almost solve the game for you. As a result, I learned how to approach indirectly, and to seal off my areas, and I started winning with a 2-stone handicap. I'm currently playing on a 13x13 board with no handicap, losing the majority of games but working my way through beginner Go texts. I don't know any openings ("fuseki") or locally optimal sequences ("joseki"), and don't plan to spend any time learning them until I have a solid grasp of the fundamentals (which is still some time away). Apparently you can become a very strong amateur-level player knowing none of either.
I've played a couple of games online against real humans, and expect that I'll play more.
So what about my criticisms of combinatorial abstracts, those luckless perfect information games? Against an AI, I can play a game quickly, undo bad short-term moves, and ponder when I need to. Go is taken seriously, but as handicapping recognises a pre-existing skill difference, there seems to be less of a competitive air.
However, I think the real difference is that I have less of an ego-attachment to winning the game. Most modern boardgames make losing less painful:
* you have excuses - luck was against you, you were ganged up on, the winner had help, or you were still learning the system
* you tend not to lose things or be attacked - most Eurogames don't allow you to be eliminated or wiped out, and limit direct attacks
* you have a score rather than just a win or loss
* there are lots of colorful things that happened even if you lost - you built up a farm or an economic engine, or led many fantasy races to their fate
* they are shallower - with a few plays, I'm usually competitive against veterans, or know what I need to do to get there
I usually feel like I'm learning rather than just trying to win, which is a much healthier mindset to approach the game.
The standard criticisms of Go boil down to variations of being dry, or requiring study to be competitive. The first is pretty subjective, but tends towards lacking a narrative and requiring brain-work. I'd argue that once you know the game these actually lessen - there are peaceful and combative styles, traps, attacks, invasions, evasion and defence, and Go is more intuitive and less calculating than many other games. The second is not really true - like any other game you can improve just by playing, though given the depth of the game the possibility of walkover matchups is inevitable.