Game Analysis 3, ~19kyu vs. 16kyu

I’m skipping ahead to more recent games. Also it turned out there was a problem with how online-go was calculating ratings, so I bumped up to 19kyu. Here’s a link to this full game: http://online-go.com/game/758387

Here are the first six moves. I’m black. The first three were to the corners, then on 4 he played at D7. I applied pressure with E7, but that was a mistake. Look at the influence he now (roughly) has from that one move counter-threatening my piece. I lost influence towards the bottom, as well as some influence on the top. It was a very good move on his part.

I applied some counter pressure on top, then blocked his expansion downwards. This secured me a bit more space on the bottom, but doesn’t solve my problem on top. He still has more territory influence than I do, which is bad when you add in Komi.

We finish building towards the top, and play a few moves in the lower middle. As you can see (sorry for swapping the colors) he clearly has the dominant position.

At this point I actually have a very slight territory advantage (30-26) but that would have still been a loss with komi. I placed an aggressive piece at C2 which he took.

My move at E1 was a mistake, I didn’t have to do it right away and I could have blocked the attack at F5. This is kinda where things go to shit.

Another mistake, this stone at F3 let him continue the snake into my base and eventually lets him put some pieces in Atari.

Because of that mistake he was able to kill off my group on the bottom and most of my territory. I tried expanding into his area, but it didn’t accomplish anything.

So I lost this one pretty badly.

Key Takeaways

  1. Don’t be aggressive early if it’s going to let them secure extra area
  2. Let a piece go early to go secure other areas
  3. Don’t make a move if there’s another move available that you won’t be able to make later

Game Analysis 2, ~30 Kyu

Still hanging out at 30 Kyu according to online-go.com, here’s another game analysis.

There were the first six moves. I’m black. I don’t like the spread out triangle I made in the middle. It feels very weak and doesn’t secure position for me along any side or corner.

White does a good job blocking off the invasion of that one piece in the upper right. I don’t know why I tried to defend it by adding a piece, I should have let it go and secured territory elsewhere.

I tried threatening his group in the upper right when he had the clear liberty advantage. Again I should have gone for control elsewhere.

He expands towards the bottom, and starts threatening my group of three while I back up to defend what space I have.

I’m trying to secure some territory on top, but he already has my pieces mostly surrounded. There’s not that much hope for securing a base here.

He easily surrounds my pieces in the upper left and captures them as well, cementing his control over the top left corner.

I close off what I can of the bottom, and he keeps control of the top. He wins by 16.5 points since he made some captures and has the komi advantage.

I think that if I had spent more moves in the beginning gaining influence on the top left instead of trying to protect my stones on the top right, I could have won. I also shouldn’t have made moves that gave him free atari plays, or “thank you” moves.

Key takeaways

  1. Don’t play “Thank You” moves
  2. Let a piece go so you can secure territory elsewhere
  3. Don’t focus on the middle in the early game

Game Analysis 1, ~30 kyu

This is my first game analysis. I’m going to go through one of my games and try to figure out what I did wrong. It started playing about a month before starting this blog, so the date will be a little off. online-go.com says I’m at ~30kyu which is as low as it goes, so a total beginner.

Here’s a link to the game if you want to see every move. (Feel free to send a friend request too!) I’m going to focus on the parts I think are most important.

I was black in this game. This was not a great starting move. The 4-4 point is fine on a 19×19 board, but on a 9×9 board it does very little. White can easily infiltrate the top, left, and corner, and I haven’t given myself much security.

I did a good job expanding around the board while white focused on blocking off that one stone. I think I should have gone closer to the right  wall though, since it’s still very invadable by white.

By playing like this I gave away the top of the board to white, and I put my stones in the top right in jeopardy by not connecting any of them. Meanwhile White got a strong footing in the top and started working into the right.

I protect some of my pieces on the top right, and prevent white from extending further downwards. His group of three on the right should be dead, but we’ll see if I kill it off.

Nope. I get distracted by his plays on the left side and he’s able to come back and save his pieces in the top right. Lesson: If you can make sure a group will die, secure that advantage before moving on.

After some more play on the right he does a Hane to J4 and I respond with… H3. I don’t know what I was thinking here. I could have played at J3 and put him in atari, then connected at H3, but I did this for some reason.

I realize the mistake and play at J2 after he expands to J3, but I’ve already given away more territory than I had to. You’ll notice he’s slowly chipping away at my territory in the bottom while his territory on top goes untouched.

After some more play on the left, the game is effectively over. Neither of us should be able to capture the other’s territory at this point if we play intelligently, and it looks like he’s won by 11.5, at 32.5-21 with 6.5 komi included. We keep playing though.

This wasn’t a smart move. I don’t know if I didn’t notice the atari at D6, but this move did nothing. If I had played at C6 I might have been able to capture something, but I didn’t. Oh well.

He tries to infiltrate the bottom but I put an end to it by making both of his eyes fake.

If I had extended the four group he would have captured my pieces later anyway, so I suppose this was the best move

He takes my group, I do a return capture, and the game is over. White wins by 16.5 points.

Key Takeaways

  1. Start closer to the walls in 9×9
  2. Don’t make careless errors, look carefully before moving
  3. Make sure a killable group is dead before moving on
  4. Stop territory advances as soon as possible

 

9×9, 13×13, 19×19

Go is traditionally played on a 19×19 board. On a 19×19 board, there are 361 spots you can play a piece at, though most games last 200-250 moves.

Games on a board that size will frequently be a minimum of 45 minutes, with many going into the hours. It would be hard to learn at that slow a pace, with each iteration requiring 45 minutes or more.

Luckily, because of Go’s rules, you can play it on smaller boards as well. 9×9 boards, with only 81 spaces, are popular among beginners. They’re also excellent for practicing the endgame strategy I’m focusing on. Since they only have ~1/4 the spaces, they tend to last ~1/4 as long, and can be comfortably played in 15 minutes.

For my Go game analyses in the beginning, I’ll likely play a lot of 9×9 games.

Go: A Complete Introduction to the Game

Before I jump into some real games and the more detailed strategy, I want to read through “Go: A Complete Introduction to the Game.” In browsing through Reddit’s “Go” section (baduk is the Korean name for Go) this book seemed the most frequently recommended. I’ll pull out the pieces most interesting and noteworthy to me as I go, but I recommend you pick it up for yourself.

Each heading will be in reference to a chapter of the book. Again, this content is from the book, not my original work. I recommend you pick it up for more in-depth insights and to support the author.

Capturing Stones

A stone or a group of stones becomes captured when all of its liberties are gone.

A liberty is a space next to a stone that it could theoretically place another stone and expand to. In this picture, (source for all images)

[Diagram]

Black 1 has two liberties, white 2 has three liberties, and black 3 has four liberties. If white were to fill the liberties for 1 and 3, those black pieces would become captured. If black were to fill the liberties of White 2, that white piece would be captured. As you can see, pieces in the middle are safer from capture and have more escape paths.

Atari

A piece is in Atari if it only has one remaining liberty. Here are some examples:

[Diagram]

For each of these pieces, if they don’t “escape” by adding another piece to their one remaining liberty, they’ll be captured on the next turn. So putting your opponent into a state of atari is a way to force them to play a certain move, or else surrender some of their pieces.

When a piece is captured, it’s removed from the board. At the end of the game each of your pieces captured by your opponent is subtracted from your score (total territory controlled).

Eyes and Living Groups

One of the additional rules to Go is that suicide is illegal. In this image below, white plays at 1, and in doing so removes all of the liberties for that group which means it gets captured by black:

[Diagram]

These types of moves are not legal in Go–you can never make your pieces commit suicide.

The exception to this is if you can capture an enemy group by placing a stone in a situation that would normally be suicide, except that it removes the last remaining liberty for a group.

[Diagram]

In this example, black only has the liberty in the center since white has him surrounded. White wouldn’t be able to play in the middle normally, but since the otherwise suicidal move results in capturing black’s group it is allowed in this case.

Safe Groups

So, the only way to make a group unkillable is to make sure it has two “eyes” (internal liberties). That way it can never get down to just one liberty without the opponent making a suicidal move.

[Diagram]

In this image, black has eyes at the two red dots. White can’t play at either, since they’re both suicide moves, so he cannot reduce black to one liberty and thus cannot capture their pieces. Black’s stones are safe.

False Eyes

But you have to be careful that you don’t have “false eyes.” A false eye is a liberty that at first glance appears to be protected, but that is actually able to be infiltrated by the other player.

[Diagram]

The black group here has one real eye, but the eye at “a” is a false one. If white plays at “b” and “c,” they’ll then be able to play at “a” and capture the two black pieces on top, followed by the larger group below it on the next turn. Don’t make false eyes. If black had played at either of the white stones marked with a red circle, then he would be safe with two real eyes.

Opening Strategy

As I mentioned in my previous post, I’m going to focus only on high-level advice about opening moves in the beginning. Here are the most important pieces I gathered:

  1. Spread across the board gaining influence on different areas until you absolutely must respond to an attack or begin securing territory
  2. Securing corners is easier than securing along the wall which is easier than securing the middle
  3. You don’t need the whole board, just influence over a majority
  4. Play your pieces in a way where they can be connected later, they don’t need to connect immediately
  5. “Knight moves” (2 or 3 spaces away and one space over, like moving a knight in chess) are pretty safe bets
  6. Don’t start below the 3rd line in on any side

Capturing Techniques

Some strategies for capturing pieces:

Double Atari

You will sometimes find, and will eventually be able to build into, situations where one move creates two ataris for the opponent. Here’s a very simple example:

[Diagram]

The move at white 1 puts both black pieces in atari. Black can save one, but the other will be captured. A more complex example:

[Diagram]

The white move at 1 puts both of the red-marked black groups in atari. Black can only save one of them. If black had played at 1 first, the two groups could have been saved.

Ladders

Ladders will occasionally come up in Go games. They look like this:

[Diagram]

White is stuck inside a ladder being built by black, and now that it has reached the wall black will capture white’s pieces. The only way to survive a ladder is to already have a piece in its future path, like this:

[Diagram]

If the ladder were to extend, it would eventually reach the piece, and then white would break out of the ladder and put black in a bad position:

[Diagram]

Now 5 is in atari, and even if black connects it to 1 white can break out through the top and start surrounding blacks stones.

Nets

A net is a move that, while not putting a stone into atari, makes any extension it could make detrimental or unhelpful in keeping itself alive.

[Diagram]

The white stone marked with red is “netted” by black’s play at 1. If white extends to “a” black can block at “b,” and if white extends to “c” black blocks at “d.” So no matter what it does, the marked white stone is dead.

Life and Death

I think this will be the most important skill to master. No matter how the game starts out, you need to be able to defend your spaces and kill off those of your opponent.

Placement Moves

One important thing to understand is where to play a piece in order to prevent the opponent from securing two real eyes.

For example:

[Diagram]

When black plays at 1, white’s group is dead. White can’t capture black without removing the possibility of getting two eyes, and black could fill out the inside to capture white. If white had played at 1, however, white would live. It would have two eyes and be uncapturable.

Here are some more examples.

Hane

Hane is a method of infiltrating a group that does not yet have two eyes, by “reaching around” the pieces. It can also just be in reference to reaching around a piece, reducing its liberties, but without having a direct link to your other pieces, like here:

[Diagram]

 Linking Up Stones

More text notes here:

  1. Connections are good. They keep stones alive
  2. Diagonal connections don’t really count, though they’re useful
  3. A “one space jump” (two stones with one liberty between them) on the third line cannot be separated.
  4. Bamboo joints are strong, they can always be connected:
    [Diagram]

That’s all for now. I think these are the most important parts of the book to my learning, but I may revisit parts of it later.

Selecting What to Learn First

In my last post I deconstructed Go’s strategy to six major groups:

  1. Good and Bad Shapes (Formations that guarantee some territory and are uncapturable vs. ones that can be destroyed)
  2. The Early Game, or Opening
  3. The Josekis, common series of play that happen in many Go games
  4. The Tesuji, or middle game local tactical problems
  5. Life and Death, how to keep groups of stones alive or kill off enemy groups of stones
  6. The Endgame

Now that I have these, the question is which should I focus on first? Joshua Waitzkin is a famous chess prodigy and was the subject of the movie “Searching for Bobby Fischer” which documented his early chess career.  In his book The Art of Learning he teaches you how to master skills effectively, but he also describes how he learned chess.

The key difference in his learning method was that his teacher started with the end game. Instead of pouring over endless opening strategies, he and his teacher started by focusing on how to win when it’s one pawn and a king vs a king, and slowly worked their way back towards the middle game, and eventually the openers. In doing so he learned to be comfortable with wherever the game ended up, and did not need to worry as much about having the perfect opening.

Many successful runners use a similar method. They dedicate parts of their training to being able to pull out a last surge of energy and speed at the end of the race to secure a higher position, instead of focusing on moving as fast as possible right out of the gate. This analogy makes the benefits of the endgame focus clearer. A runner could have an amazing start, but if they don’t know how to end strong they’ll be beaten.

Following this wisdom, here’s how those areas of strategy would be re-ordered based on mastering the endgame first:

  1. Good and Bad Shapes (You thought “endgame” would be first? Nope, without understanding this everything else is impossible)
  2. The Endgame
  3. Life and Death
  4. Tesuji
  5. Josekis
  6. Openings

Now this isn’t to say that I shouldn’t do any study of opening moves until I’ve worked my way up to them, but I’ll just focus on the core concepts of them instead of studying the endless positions that could come up. To do this I’ll start by reading “Go: A Complete Introduction to the Game” which has been widely praised as the best introductory book, and after that I’ll dig into more of the minutia of the theory.

Deconstructing Go

In order to figure out what I should focus on in learning and improving at Go, I need to break it down to its elements. At the highest level, there are two parts to this:

  1. Learning the Rules
  2. Learning the Strategy

The Rules of Go

One of the great things about Go is that the rules are actually extremely simple. In fact, there are only about 6 of them (some people condense/expand the rules, but it’s always the same core ideas).

  1. The board starts out empty
  2. Black moves first, then white and black alternate placing stones on intersections of lines
  3. A stone or group of stones adjacent to each other is captured when all of the spaces adjacent to it are filled by the opposing player’s stones
  4. No stone may be played that returns the board to its previous position (“Ko”) (This prevents infinite loops)
  5. Two consecutive passes (choosing not to play a stone) end the game, as does running out of stones
  6. The player with the largest amount of surrounded territory wins

And that’s it. Unlike a language which has tons of rules, and unlike chess where every piece moves differently, Go’s rules of play are very simple. That’s the easy part of the deconstruction.

The Strategy of Go

Here’s where things will get crazy. There are dozens of good books on Go strategy, scores of websites, and thousands of “Graded Go Problems” which you can think of like those chess puzzles in magazines. All of these resources are focused on strategy. But we can break the strategy down as well. I’m going to think of it like this:

  1. Good and Bad Shapes (Formations that guarantee some territory and are uncapturable vs. ones that can be destroyed)
  2. The Early Game, or Opening
  3. The Josekis, common series of play that happen in many Go games
  4. The Tesuji, or middle game local tactical problems
  5. Life and Death, how to keep groups of stones alive or kill off enemy groups of stones
  6. The Endgame

There will, naturally, be a number of other parts of the strategy that merit discussion, but I think 80% of it at least will fit into these six categories. I’ll try to put each strategy post into one of these categories to make everything nicely searchable as I go.

The Method to the Madness

Before I get into my progress learning Go, I want to explain the strategy I’m going to take.

I’ve been a fan of Tim Ferriss for four years now. He recently (2012) published a book called “The Four-Hour Chef,” which despite its title, is more about learning than cooking. He uses cooking as a vessel for teaching you how to learn. I highly recommend the book to anyone who wants to learn new skills quickly, or actually learn to cook.

In the beginning, he outlines a method for “meta-learning” that you can apply to adopting any skill. He provides examples from language learning, fire building, tango dancing, cooking, and more. This method is what I’ll start with to make my learning as efficient and effective as possible, and I’ll draw psychology studies and on other learning methods as I go.

You should pick up his book to get the full methods, but here’s a quick breakdown as is relevant to learning to play Go. He provides four major principles to skill development: (Deconstruction, Selection, Sequencing, and Stakes); and three supplemental ones: (Compression, Frequency, and Encoding).

Deconstruction

First you need to break the skill down to its elements. Find all of the necessary sub-skills, and figure out what you do and do not need to know. You can also find existing things to peg it against, and interview experts to figure out what they believe to be the most important aspects.

Selection

Instead of trying to learn everything at once, it’s important to focus on the most important things to learn first. This means taking all of the available information on Go and figuring out what would be the most beneficial skills to master for rapid skill development. It doesn’t make sense to focus on minute skills of the grandmasters when I don’t know basic introductory skills.

Sequencing

Once you know the most important pieces you need to learn, it’s important to learn them in the right steps. Learning how to stop a car would not be much use if you didn’t know how to turn one on.

Stakes

You’re more likely to succeed if you put some stakes on your goal. It could be a fee you have to pay a friend if you fail, or some presentation you’ll have to put on of your work, whatever will motivate you to make sure you get it done.

Then for the three supplemental pieces:

Compression

Can you put the most important pieces of the skill into a one-page document to help you remember them? This is also a good exercise in determining the most essential pieces.

frequency

Figure out how often you need to practice, and how you can make the best use of that practice time.

Encoding

Find ways to combine the new knowledge with things that you already know.

I recognize that these definitions are sparse. I don’t want to take too much from Tim’s book, and I’ll elaborate on them as they become relevant.

I’ll be playing most of my games though http://online-go.com, and will use their rating system as my barometer. The goal is to reach the rank of 1d, or 1st dan. I’d like to hit that number by the end of September… but I have no reference to assess whether that goal is ludicrous or not. I’ll adjust as need-be.

 

 

 

Hi! What’s Going on Here?

Hey there,

Glad you found the blog. My name is Nat, and I decided about a month ago that I wanted to become very strong at the ancient Chinese board-game “Go.”

If you’re not familiar with it, this is what a game of Go looks like:

I’ll explain the rules and such in a future post, but think of it as a game of territory. The board starts empty, and the objective for both players is to claim as much territory as possible. That means the game is “additive” (pieces get added) as opposed to the more common “subtractive” (pieces get removed, like chess and checkers).

I picked it because the strategic depth of the game is astounding. In my real life I’m working on a startup called Tailored Fit, and this provides a nice respite and way to stretch my mind in different directions. Also, strategic games such as Go have endless parallels in real life.

I started this blog more for me than for you. About a year ago I worked on a blog called “52 Weeks of Habits” about learning new habits to improve your life. In doing so, I learned an incredible amount about productivity, habits, and human psychology, and realized that learning through teaching was an extremely effective method.

That said, I’ll be very methodical in my learning, and I hope that what I’m doing here can benefit you as well. Feel free to leave me a note anytime at nateliason [at] gmail [dot] com if you find this blog helpful.

Nat