Wednesday, April 30, 2008

Another game commented - Game #11

[rant] I think I may need to come up with some other way to label the games. It is pretty tough to know how many games you have played on KGS if you use any level of qualification around it. Do you count a game where your opponent escapes after 10 moves? What about a free 9x9 against a player that just started (a true 30k)? A teaching game? All these show up as games played on KGS, but not all games provide a player with learning opportunities. [/rant]

I think the game I labeled Game #11 is a close enough match that it is interesting. I put a comment on every move and I hope I captured the thinking process and vision I had of what was going on. With any luck I will get stronger. Then, I hope when I look back on this game (and other commented games) I will understand what important things I am currently blind to. If that understanding can be put into words I will have some insights to share.

Seems a reasonably modest goal...until I realize that the game has been played for 3000 years! Why do I think I can explain the steps needed to advance better than all the other players and teachers before me? I'm not sure why I think I can, or why that seems motivating. But I am motivated to try and I do think - in some small way - I can be successful.

Sunday, April 27, 2008

Urgent before Big

The proverb is "Play urgent moves before big moves". I think it is a bit hard for the beginner to know how to follow this, but I sure found a great example of NOT following this in my own game. Maybe this sequence will help you follow the proverb.


I was playing white. Above is the result of my 3-3 invasion. The black group at the bottom has only 2 liberties, and my 7 stone group has 4. The only cutting point is at S3, and even if black plays there, I can kill black's 5 stone group faster and live. I make a mental note to remember I need to be careful about the cut at S3, but I don't move there. I decide all is well, and my next move is a "big" move in the center of the board.


About 50 moves later, we return to this corner group. Not only is the set of stones that I had before still there, I have expanded a bit on the bottom. Black seals the border with J1 - he also puts my 3 stones on the left in atari. I continue to assume that my 8 stone group is alive and well. I don't pay attention to the fact that I had some worry down here, and I play without thinking.


I quickly save the 3 stones that were in atari by connecting. Doh! Bad move. It was a reaction rather than a thinking move. If I had looked at things a bit longer I would have seen the play at L3. L3 takes the black stone at L2. If black plays at L4 (threatening to retake by playing L2 next, I could have played L2 myself and gained an eye at M1. Not only did I lose that eye, I filled in my own liberty! Now my large group has only 2 liberties.


Black does the right thing and punishes me for my mistake. He cuts at S3 and leaves my group in atari. My 12 stone group is dead and Black claims the entire corner.

The toughest thing about this was that I lost the game by 19.5 points. My mistake in this corner cost more than enough to switch the win from white to black.

Looking back I want to call a play at S3 "urgent for white". Playing that one stone would have won the game. I could even have played it as late as when black put my 3 stones in atari. I would have traded 3 stones for the corner - definitely a better deal.

Next time I hear myself making mental notes about an area, I will try to recognize that I am seeing an urgent move. The other half of that is, when you see yourself wanting to tenuki (play someting "big" on another party of the board) remember this proverb and recheck the current area - can you really leave it safely? Last thing, I am going to try to force myself to count the liberties of the group I plan to play a stone in, before playing it. Hopefully that will help me see that a group is in danger earlier.

KGS Ranking System

KGS seems determined to keep me from playing even strength players with no handicap. It bugged me at first, but now I am coming to terms with it. I guess the insight is: KGS has a ranking system, but it is not stable at higher kyus. There is a lot of inconsistency in the 20k - 30k skill band. There are multiple reasons for this as far as I can tell:
  • Churn - There are a lot of new players. Many play a few games and give up. So there is a constant turn over of temporary players.
  • Sandbagging - There are strong players that create a second account so they can pummel noobs.
  • Good Students - These are the new folks that stick to it and start to get better. The kyu rank KGS gives them is off because they are rapidly gaining strength.
  • Small Population - After removing all of the above you may have a few players that are accurately ranked, are sticking with it, but are truly 20k+ players.
  • Non-interacting Pockets - Of these real 20k+ players, there are groups that may never play against each other, so the ratings aren't equal. Imagine a small group of friends that only play each other (after establishing ranks around 25k). Lets say they all improve 10 stones in skill, but just by playing each other. Their ranks will all still be around 25k, when they are really 15k if they played against the rest of the KGS population.
So, what should a new player do about this? It sounds chaotic and intimidating! Well, I guess I would recommend the following:
  • Don't stress out - Have low expectations around exactly what will happen in your initial games online.
  • Play all challengers - Don't just lurk and hope the perfect game for you gets put up by someone else. Create your own custom game, make sure it is ranked, and then play whoever is available.
  • Get 25 games done fairly quickly - This sort of goes along with the original proverb "Lose 100 games quickly", but for KGS I would change it to "Play 25 Ranked KGS games quickly". Once you have done that you should have a pretty decent understanding of how KGS works and your rating should be fairly accurate. (By the way, 25 isn't a special number, really just a guess...)

Monday, April 21, 2008

Some "noise" in my planned games

I looked to play another rated game recently and try to get rid of my KGS question mark. When I logged on there were no players I could find in the 20 kyu - 30 kyu skill range. I put up a game and was challenged by a 4 kyu. Even though he gave me 4 stones it wasn't much of a battle.

Then I played another rated game against a '?' player and was losing. My opponent chose to escape rather than finish the game. I have looked him up since and see that his rating has started to stabilize at around 13k. So, I have yet to win a rated game, and my KGS graph looks like this:


(How do you get a graph with 8 points when you have only played 3 rated games - one which was escaped by a ? player...riddle me that!)

My ego is having a hard time believing I am a 27k, but I need to win a game before I can say anything. Hopefully there will be some 20k-ish folks on next time I go to play.

My attempt at a perfectly controlled science experiment has been foiled! Hehe. I won't get to choose exactly what skill level my opponents will be and my attempt to play 100 more-or-less even games in a row isn't going to happen as I envisioned it. I still feel that I can proceed with the general goal, so I will. So, bottom line is that I am not counting either my loss to the 4k or the unfinished game as "Game #4". I will try to get in games 4 and 5 soon.

Tuesday, April 15, 2008

Destroying Eyes

Groups in GO need 2 eyes to be alive. So, if you could block your opponent from making any eyes in their groups, you would have a strong chance to win. To effectively block the creation of eyes we should look at what it takes to make an eye.


Above is the basic scenario for making an eye. If we want an eye on the star-point above, we need to have a stone of our color on 7 of the 8 spots around it - we can skip one diagonal. Anything less means the star-point will likely not become an eye for us. Let's look at some situations where we don't have the right 7 spots covered.


Above here we have a classic "false eye". When 2 of the diagonal spots around your eye have been captured by the opponent, the eye is in danger. You can see, white just played "1" and has put a black stone in atari. We could save that stone by extending...but that would fill in the eye. If we don't save the stone, white will capture it and take over our eyespace.


Here is another situation, but with even worse problems! Here white has captured 2 opposite diagonals around our eyespace. Now there are 3 black stones in atari. This is another example of a false eye. We either need to fill it to save our stones, or let our opponent capture them and take over the space. So, if an opponent gets any 2 of the diagonals around a spot you want as an eye, it is a false eye.


What about if you get all the diagonals, but give up 1 horizontal spot? You can see that in the picture above - does black have an eye? No. The star-point that black wanted as an eye is now a neutral point on the board. It is in "no man's land", it is worth no points to either player, and could be filled in by either, but it won't become an eye. Losing 1 horizontal spot around an eye is all it takes to lose that eye.


So, now you can add "destroying eyes" to your strategy. Imagine you are white, playing the game above. You can see that black may try to build an eye at the star-point. If you play at the X, you will take one of the horizontal spots around the eyespace, and that eye will be lost.

19x19 - Center vs Edge of the Board

Take a look at the following position:


Who has surrounded more territory - black on the outside, or white on the inside? This was an "aha!" moment for me when I first saw this question and heard the answer. Black has claimed 136 points on the outside and believe it or not, white can only claim 121 points on the inside. The point is: the outside of the board is very valuable!

Not only are there a lot of points available at the corners and the edges, they are actually "cheaper" to surround.


Above we have 3 groups of stones, each group surrounds 9 points of territory. Look how many moves it takes to get 9 points in the corner, on the side, and in the middle. You have to work twice as hard to get the points in the middle (12 stones) as compared to the corner (6 stones). Capturing territory on the side is right in the middle in terms of cost (9 stones).

These 2 "facts of geometry" on the GO board should help you see why it makes sense to try to claim the corners first, then the sides, and finally wrestle for the middle.

Interesting sidenote: 19x19 is the largest board one can have where the outer 2 rows are more valuable than the points inside the 4th row. I guess the game would play differently on a 21x21 board!

Ladders and a Joke

A ladder is a basic form that happens often on the GO board. In its most basic form, it happens when one group of 5 stones puts a pair of connected stones in atari. It is something you need to know about so you can avoid getting caught in one. It is fundamental enough that the proverb is "If you don't know ladders, don't play GO".


Here we have a bunch of stones that could become a ladder. If black gets one more stone on C17 or D17 he will put white in atari.


Here we see the ladder. Black puts white in atari with C17, white extends to get 2 liberties, black takes one and it is atari again. Repeat until you reach the wall. Notice the zig-zag, stair step pattern made by the white stones - this is the signature of a ladder. By choosing to start the ladder with C17, black pushed white up and to the right. If he had chosen D17 instead, he could push up and to the left.


As the ladder approaches a side, black has to break the pattern a little to avoid an atari of a stone against the side, but as you can see above, one more black stone will kill the entire white group! (White at D19 leaves C19 for black as the capturing move.)


Ladders are deadly, but sometimes they don't work. All white needs is one extra stone to help escape the deadly ladder. Stones that do this are called "ladder breakers". Set this up on a board and see why both A16 and F18 are ladder breakers. One breaks the ladder started with C17 and the other breaks the D17 ladder. As the ladder pattern forms those breakers give white an atari on a single stone in the ladder that helps escape the atari on the large ladder group.


Potential ladders are formed on the board all the time - can you find any in the diagram above? The player who gets to go next can start a ladder on their opponent.

So, now you know about ladders! A few more thoughts to deal with them in game. If you see you are caught in a ladder, look to see if you have any ladder breaker stones to help you escape. If you don't, stop trying to save the group and play elsewhere. If your opponent doesn't kill the group instantly, try to play a ladder breaker sometime later. If your opponent misreads the purpose of that stone and doesn't kill the group in the ladder, you can now return to playing and breaking out of it. Secondly, if you are in a game where you have a ladder or are caught in a ladder that you can break, play it out to the end! Get as much advantage as you can with the ladder - drive it as far as you can before capturing or breaking it. This should only happen if your opponent doesn't know about ladders (although there are known games by professionals where they misread ladders).

Well, that was a long post just to set up the following joke (which I found on another blog somewhere):

Q: How many 30-kyus does it take to change a light bulb?
A: They can't do it, because they don't see the ladder.

Preparing to Cut

When your opponents stones are diagonally connected, he can be cut.


Here is a pretty common position (result of the 3x3 invasion joseki). Here white can be cut by black at D18 and black can be cut by white at E16. Note that white can't cut black at C15 due to the Tiger's Mouth black created.


Cutting helps you split groups apart and limits their ability to connect with other groups and share liberties. In the position above, black has 2 groups with 1 eye each. If he can connect at P3, the groups will live. If white cuts there, it is VERY unlikely the black groups will survive (although I do see a potential cut for black at R6).


Going back to the first diagram - what if white tries to cut here? That first white stone has no other help on this side of the wall. If black plays at F15, he can quickly contain and kill the cutting white group. So how do you successfully use the cut?


This is the point of this post! Here white approaches to prepare to cut black. To have a better chance at success, white set up the cut from a distance. It is easy for a beginner to completely miss the intent of this move. Study this diagram! Burn the image into your mind - this is a strong attack on black! You want to see this every time your opponent does it in a game. Memorize the danger of the cut!


If Black misses the intent of this and tries to surround white with something like G15, white can spring the trap and cut white with E16! Now black can not really stop white - the cut is successful. Black can try to get white in atari with E15 or F16, but the position has miai for white - whichever spot black takes, white can take the other to connect and escape.

So, pay close attention around spots where you have not solidly connected - diagonals are dangerous! A play that looks like a simple territory grab or a low threat surrounding move may be setting up a cut. At the beginning, I didn't see anything scary in moves that didn't touch my stones. Now I am starting to see these threats that are a row away from my groups - hopefully this post will help you see threats like this more easily as well.

Basic Japanese GO Terms

Go originated in China, and then gained popularity in Japan and Korea. For some reason, English speakers learning GO are taught the Japanese terminology for special GO words. (Even the word for the game is the Japanese word! Wei Qi = Chinese, Baduk = Korean, Go = Japanese.)

It is just about impossible to understand GO books, websites and comments on moves in games without understanding some of these Japanese GO terms. I will list some basics, but if you ever encounter one you don't know I suggest doing a search on Sensei's Library. There are some long lists there as well - Like this one and this even longer one with kanji.

Aji - Potential. Stones can have good or bad aji.
Atari - A stone or group of stones with only 1 liberty.
Dame - A liberty or neutral point between two groups of stones, without territory value.
Fuseki - A general term for the opening phase of the game.
Goban - The board you play GO on.
Gote - Losing the initiative, playing last, a defensive play
Hane - A play that reaches around the opponent - the response to a contact play. (Look this one up!)
Honte - the proper move, generally not flashy, just a solid, honest move
Hoshi - a "Star point" where handicap stones are placed, also good spot for an opening move
Ikken tobi - a one point jump
Joseki - a well tested series of moves that is considered to give even results for both players
Keima - a knight's move
Kifu - the game record, list of moves that allow others to reproduce the game
Ko - a spot where both sides can make continual captures, limited by a special rule for ko
Komi - the points given to white in compensation for going second, usually 6 1/2.
Miai - a situation with two equally good spots to play, if your opponent takes one you can take the other
Moyo - framework, a loosely connected set of stones intended to allow the player to build a living group and territory.
Sansan - the 3x3 point
Seki - called "mutual life", a situation where both opponents stones share the same liberties
Semeai - capturing race, mutual attack
Sente - initiative, playing first, being on offense, driving the direction of the game
Shodan - a ranking denoting GO mastery (1-dan)
Tenuki - playing elsewhere, choosing not to continue a local fight
Tesuji - a surprising, skillful, tactical play
Yose - the phase known as the endgame

Basic Sente and Gote on the Edge

Many times in GO, a shape will head toward the edge of the board, but stones will not be placed all the way to the edge of the board. Players feel that early in the game there are more important moves than sealing up the borders on the edge. But as the game moves toward the end, these borders generally get locked down. There is a pretty predictable set of moves to do this, but there is a little more going on than you might originally see.


Take a look at the diagram above. Black has claimed a corner, and assumes that he can make 2 eyes and live there. White has surrounded him. On the top right border (at T4 and T5) you see that both black and white extended their wall straight to the edge to determine the border. Although that can happen in a game, it usually doesn't. If we look at the bottom (at P1 and Q1) you can see that the border remains open.

Imagine white decides to settle this border and moves here first. Instead of playing straight down to P1, he squeezes black by taking Q1. This gives black one less open spot to build eyes with, and one less point when the score is counted at the end. Black can't respond with P1, because that would put P1 in atari right away. (A white stone at O1 would capture P1 immediately.) Black has to stop the white advance with R1, which puts white's Q1 in atari.

White needs to save Q1 from atari. P1 does the trick. Now white is solid and he stole a point from black. Should black be happy with this and play elsewhere? Unfortunately for black, there is still a threat by white.

The stone at R1 isn't strong enough to stop another attack by white! White places a stone at R2 and puts R1 in atari. How many more stones will black have to play to stop this attack? Black is not only losing points, he may lose this whole corner if he can't make 2 eyes!

Black must respond to this series of moves by playing at R2 to seal the border from further invasions, even though it costs him another point of territory. Replay almost any game and you will see this series of moves. Now you know how to squeeze 2 points out of your opponent on the edge, and what to do if they start to reduce your territory. If black had played here first, he could have played at P1 and pushed white the other way.

Is there anything more subtle going on? Yes. These moves act as a series - once one side starts them, the other "must" respond. (Obviously you don't have to, but it is a series of forced moves.) The person that starts the series has the advantage. Not only do they get to steal a few points from their opponent, they get to pick where the next battle will be after this little border series is finished. When you play, if you have the freedom to move where you want, and are causing your opponent to react to your moves, you have sente (Japanese word pronounced SEN-TAY). You are on offense and are forcing your opponent to react. If you are playing moves to counter threats from your opponent, especially moves that don't really help improve your position, you are in gote (GOH-TAY). The move by black at R2 to end the series above is a gote move. The first move of the series by white is a sente move.

Sente and gote are talked about a lot. They are particularly useful in battles over a ko (which I haven't talked about on this blog yet). In general, you should be trying to play moves that attack your opponent, grab more territory, and force them to react defensively. You want to have sente. The less your moves do to force your opponent to make a gote move, the more you open the door to letting them play a strong sente move that forces you back on the defensive.

Makes me think of a cool GO T-shirt:

Monday, April 14, 2008

Game #3 complete & commented

This game was a lot of fun. I started off horribly, and came back to lose by only 8.5 stones. My opponent was a 26k, which seems to be about where I should rate. Some of the most fun about this game is that I am still kicking myself for opportunities missed. Here is the corner that is haunting me the worst - I invaded a bit, but not as well as I could have. Black to play:

I learned a lot in the game. I was expecting a timed game, and so it was no real issue. I really played awful at the start of the game. I need to go back and see how I got such bad position. But, I think I did a much better job at "tenuki" in the opening. (Tenuki means to play somewhere else - last game I followed my opponent around like a puppy.) This time I feel I did a decent job of staking out unclaimed territory in the opening. The stones I had sitting around the board on the 3rd and 4th line ended up helping turn the tide of the game - or at least make it closer!

Other things: Wasting moves to kill an already dead group. It is difficult emotionally to leave a big group that you can capture alone, even if it is dead. It feels good to surround and capture - it makes you feel like you are winning! It is an emotional pump, and a momentum shift to remove a bunch of your opponents stones. Unfortunately, it may not be an efficient move. Leaving some big, dead opponent group in atari and using your stone to improve things elsewhere could be a much better move.

Being satisfied at achieving a minor goal and relaxing. Maybe you have a group that is squeezed between opponents stones. You think to yourself, "If I can just make 2 eyes I will be happy over here". You fight and keep the group alive by making eyes and then celebrate. "Yeah! I did it!" What if you can do a lot more in that area now that the group is safe? Maybe you can squeeze your opponent - maybe you can kill one of their groups! Relaxing after a minor achievement and thinking you are done in an area is a mistake. Look for further opportunities now that the first goal is met.

Don't waste stones. I am sure I wasted 10 or 15 stones at a minimum this game and the final score had me losing by 8.5 points. If I had not thrown those stones away I would have won! (To be fair, my opponent made the same mistake, and had some easy opportunities to waste less as well.) Don't play "hope GO" and try to save a group that is dead unless your opponent gives you 3 or 4 unanswered moves - it's dead! Leave it alone! Don't over fortify behind your own lines - resist answering a foolish threat that is in your area but will never make 2 eyes.

More once I have this game reviewed!

Friday, April 11, 2008

Ranks and Rules

Two areas you want to be able to understand and talk about are GO rankings and GO rules. My opinion is that you want to know enough to talk about these, but don't spend too much energy on them. There are complexities with both that cause a lot of discussion about subtleties but the details don't matter so much for the beginner.

RANKING: Ranks in GO are like belts in martial arts. GO is complex enough that one can differentiate between 40 or 50 different levels of skill. A true beginner starts at about 30 kyu. This is often abbreviated to 30k. As they start improving the number decreases. Theoretically, for each kyu you move up, you have improved 1 stone. So, players can have competitive handicap games even though they are at different skill levels. So if you work your way up to 20 kyu and you want to play against a 17 kyu, you would receive 3 stones at the start of the game to make the game even.

(SIDENOTE: A shorthand I have seen that isn't obvious is DDK and SDK. DDK = Double Digit Kyu - the weakest group of GO players - from 30k up to 10k. SDK = Single Digit Kyu - 9k to 1k.)

Players move up through the kyu ranks to 1 kyu, and if they improve 1 more stone in strength they go to 1 dan. You can think of 1 dan as equivalent to "black belt" in a martial art - it shows that the person is a master of GO. Many players originally aspire to reach 1d (or "shodan" in Japanese), just as martial artists wish to receive their black belt. As a dan level player increases in skill their rank now increases in value, so a 4 dan (or 4d) player would give a 3 stone handicap to a 1d player. The dan ranks seem to go up to 9. I have never seen a 10 dan rated player. But there are some different rankings up at the top end. Some ranking systems differentiate amateurs and professionals. The American GO Association (AGA) uses the dan label for amateurs up to 6d or 7d and then continues up to professional at 1p (p = professional) and beyond. (2p, 3p, etc.)

So here is where ranking starts to get unclear. First you have to realize there are lots of different groups that rank players. These different groups rank in a way that is internally consistent, but that may not match exactly with another rating system. The AGA, the European GO Association and the online GO server KGA all use the kyu and dan ranking system, but a 1d on KGA may not be as strong as a 1d somewhere else. Here is a table of ranks compared across some of the well known ranking groups.

So, now you know something about ranks! Does it really matter that an AGA rating of 20k doesn't equal a Japanese 20k? No - the important thing is that once you pick a system to be rated on, and play enough games to be rated, you can find opponents of equal strength in that system. It will also let you play challenging games with people that are not at your kyu level by giving or receiving some handicap stones.

RULES: Unfortunately, there is no single set of "official" rules that all players play by. The differences in the rulesets are small and for the most part shouldn't really change the outcome of the game. On Sensei's Library I found 10 sets of official rules listed: AGA Rules, British Go Association Rules, Chinese Rules, EGF General Tournament Rules, EGF: Simplified Ing Rules, French Rules, Ing Rules, Japanese Rules, Korean Rules, New Zealand Rules.

The most popular rules are Chinese rules and Japanese rules. Many of the other rulesets are just some form or deviation of these rules. Even this has complications - Japan seems to have published the "official" Japanese rules more than once, and each published set of rules is different. So, you can spend time worrying about little details, or just get the basics and start playing. Once you get to the point that you are playing in tournaments, then you probably need to read the fine print.

Basics of Chinese Rules: Points are scored using area counting. (Both the spaces you have surrounded and the stones you have on the board all count as points.) Prisoners that have been taken off the board (both those captured during the game and groups recognized as dead at the end of the game) do not factor in to the points.

Basics of Japanese Rules: Points are scored using territory counting. (Only the empty spaces you have claimed count as points.) Then, subtract the number of prisoners your opponent captured. (We always put the prisoners back onto the board, filling up the spaces that were our points. That does the same thing as subtracting them out. It feels like the POWs are returned after the war and you have to find a place for them to live! :) )

Online GO servers will help score the game once it is complete, computer opponents will do the same. If you are playing over the board, hopefully your opponent can fill you in on what rules they want to follow. It won't take long until you are able to score a game.

Last thought - sometimes Japanese rules and Chinese rules score the game exactly the same, other times it varies by a few stones. If you are super concerned about it, make sure you understand the rules you are playing by. I think the truth is that those first 100 games or so that are really learning games will likely be losses. By the time you are getting to a competitive level of skill, you will have a good feel for the rules. Don't worry about the details until then!

Thursday, April 10, 2008

Efficiency

GO is all about being efficient. That sort of translates to capturing the most area with the fewest stones. There are opportunities to do this all the time on the GO board and I only have a few examples to show.


Here is a situation that might exist in a game. Both ends of the black wall are single stones that are not solidly connected to the rest of the black stones. White could cut across and separate them from the rest of the wall. If black gets the chance he could simply connect each of those stones to the rest of the wall and block white - but would that be efficient? There is a more efficient way to protect from having white invade at those weak point while grabbing more territory.


Below you will see 2 stones added to the black wall. These stones make a shape called a "Tiger's Mouth" at each end of the wall. (A Tiger's Mouth is 3 stones in a shape that protects the point in the center - if your opponent moves there, they put themselves into atari instantly!) You can see that white can't cut across the ends of the black wall without putting stones in atari. You can also see that black is efficiently extending out toward the rest of the board with these moves.


Below is another efficiency series. If you start with the position on the left, it seems like the best move is to capture. (The middle is the result of that choice.) But it may be more efficient to play the move on the right. White should know that black can capture that stone with one move, even if white extends down. That white stone is effectively dead. The black move on the right grabs a little more territory on the right and still effectively "kills" the white stone.


Of course if white is allowed to add a few stones in this area while black is focused on another part of the board, that dead stone could be rescued. Black will have to keep an eye on the area and truly kill the stone before white can save it.

Extending and other basic moves

I gained some confidence on the GO board after reading what moves are considered good. The opening is a bit of a challenge, and not covered by these moves, but once you have a few stones on the board (probably near star points) these are the moves you can use to grow your groups and build territory.

1. EXTEND - this is a fancy way of saying "put a stone next to another one". Look at the situation below - you are playing black. If you play somewhere else on the board, white could add another stone around this one and put it in atari. If that happened you could still keep this stone alive by EXTENDING. But why wait? This is a perfect time to foil white's plans and extend before you are in atari.


If you decide to extend before you are in atari you have 2 options of where to play. If you wait, the third stone white plays will leave you only 1 way to extend, and a good opponent will make sure that you are stuck extending in the direction that is worst for you. So don't wait! Look at the rest of the board and extend in a helpful direction - towards some of your other stones maybe. Here black extends to the right:


2. DIAGONAL - this move appears less solid than extending. But the truth is it is a great move. The reason is it has a quality called "miai" (MEE-EYE). Miai means 2 equally good options exist - when your opponent takes one you take the other. If you look at the diagram below you can see that white could move in at the upper right or the bottom left. You can still connect these stones by taking whichever spot white leaves open!


3. ONE POINT JUMP - This is a great move for getting around the board quickly. If your opponent moves right between these stones to try to break them apart, you can instantly put them in atari!


4. KNIGHT'S MOVE - Here is an even faster way to move around the board. If you look at this and imagine adding a third stone between these two to start connecting them, you will see you have an extension and a diagonal move together.


The above moves are all strong and will serve you well throughout you GO career. Fire up a computer GO 9x9 game and try them out!

Rethinking "a review every game"

I think I have learned a lot from reviewing my own game and having it reviewed. I have some specific basics that I will share in a future post. Unfortunately there is a downside to this process - it takes a lot of time. Play, self-review, comment, share for review, get reviewers comments, post, and add learning details to the blog is pretty time consuming. It has been 8 days since Game 1 and I still don't feel I have studied and absorbed my reviewer's comments fully.

Maybe I should stay the course and stick with having every game reviewed. What is pushing me to change is that I need to play at least 4 rated games a month to gain a rating on KGS. If it takes 10 days per game I will only get in 3 games a month, so I won't get a rating. If I do that for all 100 games, that means 10 days per game x 100 games or 1000 days with no rating. That's almost 3 years! That doesn't seem like the right approach. Just for contrast, my opponent in Game 1 has already played 108 games since I played against him!

The proverb is "Lose your first 100 games as quickly as possible". This approach isn't quick. So, the bottom line is that I need a way to speed up my games played. I do want to look at them and see where I could have improved, but time between games needs to shrink to 2 or 3 days I think.

I could look to have occasional games reviewed. Maybe every 5th game or 10th game? Or another geeky thought - I could have my prime number games reviewed (2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97, 101). By picking some reduced set of games I could balance speed with study better. I think there is value in being able to share my rating as I play games just to help set expectation, so I want to get a KGS rating at least.

Hopefully I will still be able to share any insights I have along the way. I am trying to remain conscious of all the thoughts I have around GO, and anything that seems useful should end up on these pages.

Typing this out makes me feel like it is a better approach, even if in truth it is just a justification for changing. So, I am going to switch to the prime number games get reviewed. Heh. This also gets a tiny skeleton out of the closet - I DID actually play 1 game of 19x19 on KGS before this first reviewed game. It was before I had decided to try to blog about GO. So, the previously noted "Game #1" will get retagged as Game #2 and be the first entry in my "Prime Games" on the sidebar.

Alrighty, back to the slightly modified agenda...

Tuesday, April 8, 2008

First Sequence to Understand

Your opponent sets down a stone. You want to attack! But where is the best place to go? You want to smash right up next to his stone, right? Wall him in and stop any expansion, right? WRONG! I think what intuitively makes the most sense is a bad move.

GO has patterns of play that are called "joseki". There are tons! Entire books are written about them. They are a series of moves that have been tested over time and are considered to be a somewhat "even exchange" between the two players. (Similar to chess opening theory if you are familiar with that - although as I understand it, the GO joseki can happen at any time during the game and in various locations.)

I wanted to call this post "Beginner's Joseki", but I realized it isn't an even exchange. It is bad for one side! Which is what I want to warn you about. The proverb is "Don't Attach When Attacking". But what does that mean? Where does the strongest attack appear to be for black in the diagram below? (Assume the first play of the game, your black stone with the "1" on it is on the other side of the board. White "2" is his first play of the game.)



Let's get in his face! That will force him to play defensively, right?


Actually, white has a great offensive response to our attack.


Basically, white attacks us right back. and the 2 stone is helping the 4 stone on offense. We are in what is called a "liberties race". Who can surround the other sides stones first? White already has two stones out there and our black stone has only 2 liberties left. If black continues to try and surround the white "2" stone, he will lose the "3" stone first and fail to capture white "2".



After a few more stones it is white's turn. Both black "3" and white "2" are in atari, but since white was there first (or turned around, because black attacked) white will win the liberties race:


The black stone at 3 was captured. So, in this simple scenario you can see why attacking by connecting (touching the opponent's stone) is to the advantage of the player who had a stone there first. It also gives you your first strategy! If you are attacked by someone who places their stone against yours, you know how to respond - attack right back!

Sunday, April 6, 2008

Insights into the Beginner's Mind

I played a 13x13 against my fiancée (who has not been bitten by the GO bug). It is interesting to see the concepts a completely new-to-GO person struggles with.
  • Lack of clarity between the definition of a liberty and an eye.
  • Eye/liberty confusion leads to the inability to determine if a group is alive or dead.
  • Without confidence to determine which groups need added defense or eyes, it leads to killable groups that are a shock for the new player to lose.
  • This leads to complete frustration with the game. It seems almost arbitrary which groups live and die.
If this is you, or someone you hope to help start up in GO, see if they understand better after doing the following:

  1. Surf over to The Interactive Way To Go. This is a great interactive starting place for learning GO.
  2. If that format is comfortable, set up a beginners problem set at GoProblems. Click the black go stone at the top of the page that says "Problems" and create a set of puzzles rated from 30 Kyu to 25 kyu. (Kyu is a rating of GO skill. 30 kyu is a beginner rating and 25 kyu means you are a bit stronger. So, this creates a set of the easiest problems on the site.)
  3. Once the basics start to make sense, try some 9x9 games by downloading Igowin (for Windows) or GnuGo (for Mac).
  4. Next step is to take off the training wheels - start playing against people! Either live face-to-face games (on smaller boards) or by using an online GO server (where you can still start on small boards).

Saturday, April 5, 2008

Beginner's self-review

Game #1 has already revealed some points to improve on, things I can see that I did wrong. For the most part this felt like an orientation into playing on a 19x19 board, playing with time pressure, and playing via the KGS system. What were my impressions?

Time pressure - no real big revelations here. I guess it is part of the game a player has to get used to.

KGS - this is an online GO server where you can play against other GO players from around the world. Just like any other software it takes time to become a strong user. I was surprised by the pop up windows at the end of the game, and I didn't expect to have to count dead stones. Ultimately it is a pretty straight forward application. KGS has some nice features you can access from the web like a saved list of all the games you have played and a graph of your current strength. There are other GO servers out there - IGS, CyberORO, and others I have never visited (MSN Zone and Yahoo both have realtime play and the Dragon Go server has a turn based "play by email" system). If you want a place to play against other people, and you don't have a go school next door, getting comfortable with an online GO server is a must.
One little hint - when I decided I wanted to pursue GO fairly seriously, I picked a screenname and registered at every GO website I could to sort of "reserve" that name across the many sites in the online GO world. Not required, but I think it may make things easier.

The 19x19 board - this definitely plays differently than a small board. On a 9x9 board you use the first few moves to stake out territory, then you jump into a completely tactical situation. Life on the big board feels like it happens at 3 levels, and they are all relevant throughout most of the game.
  1. The "10,000 foot" view - This is looking at the whole board like an airplane fly over. It is easy to get caught up in the close up "level 3" view of the game and worry about a fight for a corner or some tactical issues, but the goal of the game is to win more than half the board. Forgetting to step back and look at all 361 points as a whole could leave you "winning the battles but losing the war". I almost completely ignored this view in Game #1. I did just what I had done in 9x9 games - play a few moves around star points and then dive into tactical battles.
  2. The middle view - (this view belongs between 1 & 3 but if you are new to GO, read this after reading 3). This is a little harder to explain, but it is about watching how the groups are interacting on the board. What are the local "weather patterns" that are happening in the game? Is your opponent getting a strong presence in the center as he surrounds the group you are building? Is that group you made on the side with only one eye going to be able to connect with other groups so it can live? Is your opponent using your moves against you? Are you being out maneuvered such that you may lose the game? Paying attention to this was almost a completely new need that I didn't come prepared to look at. I would battle, battle, battle and then suddenly see I was surrounded with no eyes. Here seems to be where the heart of GO may be. The winner will use the strength and force of both their own moves and the opponent's moves to adjust strategy and tactics so that they outmaneuver the opponent and win.
  3. The tactical close up - This is watching for atari, counting liberties, doing life or death studies of groups (tsumego), trying to create eyes, and fighting against your opponent. This is what I did 99% of the time in Game #1. It is what is needed on a 9x9 board, but will not be enough to win the war on the big board.
A player seems to need to cycle between all these views every turn to make sure they keep the big picture in mind, know the specific details of the area they plan to play in and then make the move that best tilts the overall game in their favor.

One other thing - the 19x19 game feels like a marathon in comparison to a 9x9. If you have gotten used to the pace and length of a 9x9 game, you will likely run out of steam in the middle of the game. There are over 4 times as many points to play on in a 19x19! In my Game #1 I think I tried a few local skirmishes, got outplayed on them and had the wind knocked out of me. (All in under 81 moves a 9x9 would allow!) Then I turtled up and secured a corner (while giving the rest of the board away) and dabbled in a few other experiments before allowing my (probably very bored) opponent to escape. I can and will do better.

Last point today - don't play "hope GO". That is playing and hoping that your opponent doesn't see what you are trying to do. Unless you are doing something super tricky you have to assume your opponent sees and understands every move you make. You can't hope that they will be focused on the bottom while you are quietly surrounding a group at the top. None of my "hope he doesn't see that" strategies worked. I don't think you can become a good GO player relying on hopeful trickery. You have to win by honestly outplaying your opponent. Why does that sound scary? :)

Friday, April 4, 2008

Game #2 is complete - working on comments

KGS shows game 2 as being played 4/3/08 3:18 AM. I was interested in playing a Rated game, so I could start determining my current strength. I challenged a 19k player and although they hesitated, they eventually accepted. (The game did end up being a Free game though, and I wonder if my opponent changed the setting before we started? Perhaps I just misread that this was a Rated game to begin with...)

Having not played much 19x19 I was fairly intimidated. Adding to the newness and strangeness was the time limit on the game! I had 30 minutes to play all my moves. We began play and before long I could tell I had made some errors. But I wanted to keep fighting and just get more accustomed to the big board. I managed to secure a corner, but basically by giving away the whole right side. I was getting further and further behind, but I wanted to play things out to the end.

Since I am used to playing against a computer, I forgot I was playing against another human. I wasn't taking any of their needs or concerns into account. I learned later that it is good form to resign a game you have clearly lost. (If you want to keep playing things out, ask your opponent if they mind.) Your opponent may not enjoy adding stones to a game whose outcome is already determined.

So, resign once you are confident that you have lost. You can ask your opponent to review the game with you and perhaps learn a few things and even make a friend. (Many of the thoughts here are lifted directly from responses I received on a thread at GoDiscussions.)

Thursday, April 3, 2008

Introduction

One GO proverb for beginners is "Lose Your First 100 Games As Quickly As Possible". My goal with this blog is to play and comment on my first 100 games on KGS. I will then look to have some of them reviewed by a stronger player. (The "prime" games...2, 3, 5, 7, 11, etc.)

My hope is that I can capture many of the early insights into playing the game better in posts on this blog. With any luck, the insights that help me get better will help other players new to GO. In the long term I hope my skill increases enough that I can act as the teacher for others.

May my losses be your gains,

Vultur


Disclaimer: I started playing GO in early 2008. The 100 games I intend to capture are my first one hundred 19x19 games against real opponents. I have played a bunch of 9x9 against computer opponents Igowin and GnuGo. I have worked through a few hundred problems at goproblems.com so I am not a complete beginner.