Useful Go (Weiqi) Concepts for Bughouse Players
In Go, every move is a drop, and the stones don't move. They stay in place unless they are captured and removed. This is the basic reason why some Go concepts are useful in bughouse, where pieces are dropped. I'll explain the rules, explain a little strategy, and then show some of the concepts useful to bughouse. It may be possible to skip to the third section if you don't care for a long post. This post is necessarily long, but there may be surprising payoffs. I realized a post like this one might be worth writing while teaching bughouse. I often caught myself wanting to use the distinctive terms and explanations from Go. It's fun to look for insights is unusual places. improvement is so elusive that it's worth expanding our search for it.
Some of the concepts are so close, or they already have a close enough chess equivalent. Some have no equivalent but it would be stretching to find examples where they genuinely applied.
But there are some which are distinctive and applicable. They don't describe new strategy, but they do describe things strong players do intuitively, which have no names. Giving things names makes them easier to see.
One: The Rules of Go
The Go board is a 19x19 grid, and the two players alternate placing stones on the points. In Go, Black moves first.
The object of the game is to end with the most points. Points are gained by surrounding empty spaces, or territory. Points are lost by losing stones. The game ends either by resignation, or by mutual passing. If both players agree that the game is over, the players will add up their territories, and subtract their lost stones. The board will look very full at the end. For instance, this is how one of my games ended.
Every stone has liberties. The liberties are those points directly adjacent, but not diagonal to the stone. If a stone has no liberties, it is captured.
The White stone is in Atari. If Black plays at A, the White stone is captured and removed from the board. If White plays at A, the stone is saved.
Groups of stones share liberties. White has gained three liberties, and Black cannot capture the group in one move. Every stone has four liberties, except stones on the edge of the board. Stones on the side have three liberties, and stones on the corner points have two. It's illegal to remove one's own final liberty, so self-capturing is not allowed.
Some groups can never be captured. If they make living shape.
The Black stones are surrounded, but White cannot capture them. It is illegal for White to play at A, and it is illegal for White to play at B. So long as Black does not inexplicably fill in A or B, Black is alive.
Some eyes are false.
White can capture at the point marked X. The single stone is not properly connected to the rest of the group. This entire Black group is dead. In a real game, this would give White 13 points of territory, and Black would lose 11 points. A big swing.
The last point is the Ko rule. The board position may not be repeated. Players are forced to play away before retaking.
Black has just played the marked stone. When White captures the stone, Black must play somewhere else. Black cannot take again.
It's illegal for Black to play at A. Black must play somewhere else first. This position is a Ko. Some Kos are unimportant, but sometimes Ko decides the life or death of a group. Both players want to finish the Ko, perhaps by connecting. Each side makes a Ko threat, often a throwaway move which has no other value than as a one move threat. Both sides would like to be able to ignore the Ko threat, and win the Ko.
In the endgame, a Ko may decide which player wins by a half point.
The Ko rule is a way of turning a bug into a feature. Rather than allowing draws to become ubiquitous, Ko adds a level of interest.
Since the mid-20th century, White received compensatory points for moving second. Usually this is 6.5 points. The half point means there are no draws.
This is more or less the rules of Go.
Two: Go Strategy in Broad Outline
Go is a balance of safety and efficiency. Every move is an asset.
Black can say that the bottom right is Black territory. But 1, 2, and 3 are so close together that they are not very efficient moves. On the other hand, White can say that the top is virtually White territory. White's moves are more efficient, though less safe.
Staying Connected is Key. Good players make their stones work together, and reduce the harmony of their opponent's stones. There are different kinds of connections. Some are solid, others can be cut. The following diagram has various strong connections.
A: Extension or Iron Pillar
B: One Space Jump
C: Diagonal Connection
D: Two Space Extension
E: Small Knight
F: Large Knight
The shapes B-F are such that while they can be cut, they are close enough that any fight will favor the one who was cut.
In this diagram, White is not really worried about Black playing at the point marked X. It's technically possible, but White can stay connect by responding with an atari on the cutting stone.
After the sequence 1, 2, and 3, White may defend A or B in some way. But White could also decide that those stones are fine enough, and that there are bigger things on the board.
Corner, sides, center. That's what all Go players learn. The corner can be surrounded in the fewest moves. The sides are the most easily surrounded after that. Surrounding the center is the hardest, because the center can be invaded or reduced from all sides.
Go did not have its "Deep Blue" moment until 2016, 19 years after the Kasparov-Deep Blue match. Since 2016, Go professionals have studied the AI's recommendations. The AI really likes the corners, valuing them highly.
Black has easy territory, or solid cash. What White has is more nebulous. It's influence. White's wall of stones projects power along the left side. White's has thickness on the outside. Any fighting which occurs on that side will favor White, because there is one direction which is blocked for Black. A strong player can use influence to control the game. If Black lives inside White's area of influence, White can use the fighting to get influence in a different direction. Some players call this the ripple effect.
Before the AI, Go players were taught not to take too much cash too soon. Now influence is valued less than it was before the rise of AI. But the AI can also use influence powerfully. As in chess, the engine recommendations are not necessarily the easiest for human players to implement.
Invasions and Reductions
Many players over-estimate the dangers of their opponent's frameworks. They take unnecessarily extreme measures and invade where their opponent has the most strength. Often, it's better to play a reducing move, just outside of the opponent's territory, not to steal it, but to contain it. It's important not to be too jealous of the opponent's potential.
This is just a toy example, to illustrate the point. Here, White could worry that Black has too much cash, and jump right in.
White can probably live here. But it might have been better to play more lightly.
This is a shoulder hit. This White stone can run more easily, but it doesn't stand to make much in territory. White could also try to reduce in different ways.
Three: Go Concepts for Bughouse Players
1. Aji, also known as Trouble or Potential
In Go, stones have aji if they have the potential to be annoying. Many stones in Go are simply assumed to be dead, and they are not captured because it is wasteful to capture dead stones. If a player's shape is bad, their position can be said to have bad aji. Sometimes a single enemy stone inside a position can create significant aji.
The marked White stone was naturally surrounded in a normal sequence. But it has aji. It is too soon for White to make good use of it. But as the game changes, White will try to use this stone. Black cannot take the time to capture it, so Black can only be careful in the middlegame.
The following diagram is from one of my online games.
White's position on the bottom is dodgy. White has tried to do too much, and Black has gathered power on all three sides. There is no stone inside yet, but Black's outside strength makes trouble likely. White looks close to alive on the bottom, but Black can put a stone in, and cause White problems.
White is able to contain Black's 1 stone. But the resulting shape of White's allows Black to get tricky, and in this game, White's group died. There is a false eye.
Sometimes Go players take the time to clean up the aji in their position. Sometimes the aji cannot be removed--but it can be controlled.
Aji also exists in bughouse.
Black is reasonably well developed, not particularly active. But after placing the pawn at g3, I would say there is aji. White's knight defends against Queen at h2 checkmate. But suppose this happens...
Black dropped a knight on f2. White could respond in many ways.
1) Tell the partner not to give Rook or Queen, and thereby constrain the partner's play.
2) Give up the Rook for the knight, potentially bothering the partner.
3) Prevent a mate on h1 by playing there first.
Suppose that White was up time, and wanted to thoroughly clean up. Instead of dropping the Knight on h1, White could play Rfc1, and Nf1, aiming to take g3 for free, and forcing Black to give up the f2 Knight for just a pawn.
When defending, how many steps will you defend in? A one-step defense evades a mating threat. Defenses with many steps may improve King safety, improve pieces, and win material. It's also important to know if your partner is aiming to defend in one step, or many. Sometimes games are lost because a player is defending in many steps and their partner thinks they are going to defend in few steps. Sometimes counterattacking chances are squandered, leading to this sort of exchange.
"Let me fix."
"We don't have time. Counterattack, I feed."
"I don't see the mate." or "I am winning material".
Defenses with many steps are often too ambitious. And defenses with many steps may fail to take into account the opponent's aji.
Another example:
Should White be concerned about Black's b2 pawn? Perhaps White should play Qd1-b1 or Bishop at b1. It takes Black time and piece flow from the partner to do much with this. But another pawn on a2 would begin to look bad for White.
White's King is fine, but if White is forced to give up heavy pieces in return for only pawns, this could cause problems for White's team to snowball.
How White best deals with the pawn on b2 depends on the partner's prospects, as well as the time situation.
The two examples above are of reasonable aji. A strong opponent will always have some potential. But once you have the idea of aji, you can wonder if you are giving your opponent too much of it, and if you are getting enough out of the deal. You can also ask about the aji in your opponent's position.
For a bughouse game, Black's opening is provocative. Would you, as White, put something on f7? Many players would. I would just go Bc4. What if we changed the position slightly?
Black played c6 instead of Nc6. This might keep Black safer in some variations, if the Black King can reach c7. I would be even less likely to throw in a check here. Instead, I'd prefer to develop, threaten to open the center, and challenge Black to find the time to fix the f7 weakness. As the game goes on, the f7 weakness is likely to become more of a problem for Black.
It's common in Go to say "count to two". A strong player can handle one weakness, but two weaknesses is too many. And in this position, rather than pounce on the f7 weakness, which may not be exploitable soon in the opening, White might play like a Go player, and simply make a second weakness. When the trades do come for White, Black will be unable to deal with the f7 problem and with the opening of the center.
Why say "aji" and not simply say "activity"? I think the difference is noticeable enough. Aji is a more specific form of activity. It refers specifically to almost-captured stones doing just enough to bother the opponent.
2. Overconcentration
Overconcentration is too much stuff in one area. In the following diagram, Black has two strong walls, but they are much too close together, and Black has fallen behind in points.
Black would be doing much better, if their stones were further away.
Does this really exist in bughouse? Sub-2000 rated players often overconcentrate their pieces. Around 1200, a defensive mindset creates moves like this...
And sometimes unnecessary g6 or h6 drops.
The reason we can know these moves are bad is that they add nothing. They add extra protection to squares that don't need it, which not only hinders easy piece development, but it also wastes droppers. Droppers are costly in bughouse.
Another way in which pieces get overconcentrated...
Black should just play Be7 here, and retake with the Queen. But some players want more than this. They want a trade ending with a Bishop on e7. So they play...
If White takes on e7, Black got what they wanted, a small plus. But White prefers to keep Black overconcentrated.
Black has some work to do to untangle the back rank.
There are many positions where it's better not to trade, and leave the opponent with a confused mass of pieces. A decision to not trade has to be balanced against the partner's requests, and most partners want trades. But an overconcentrated opponent can be a huge asset.
3. Throw-in.
Sometimes throwing a stone into the opponent's position is the only move.
If Black plays at A, White captures it. But then Black plays at A again, and White loses all of the stones.
There are throw-ins in bughouse too. Sometimes the only way to checkmate is by throwing in a piece, forcing the opponent to capture it, and blocking the enemy King's escape to safety with his own pieces. This is usually done with Bishops and Rooks.
Suppose White wants to make sure that Black's King never reaches g8.
This looks like a pure flagging move. But there are positions when this motif is the only one that works.
White can't just play Knight at c7, because the King can run to b8. First, White needs to prevent that.
Now Rook at b8 is the only way White can mate with two Rooks plus Knight.
If Kc8 Rc7. If Ke8 Rook at d8, Nc7 mates.
It's also possible to do this with a Bishop...
However the Bishop is taken, the King cannot find safety on c1 or b1.
It's also possible to throw in a knight.
This was actually played, and it led to a Black win. The flaw in the idea might be that White can play Kc1. In the actual game, White captured it and was mated.
In this position, I wanted to prevent the opponent's King from getting to e7. The intuitive Rh8 doesn't do that, and White can't give rook anyway.
Knight at g8. We can call it a placement. It's quiet but hard to deal with. It looks pointless but the threat is Rh8. White mated shortly after.
4. Probe
A probe is a kind of asking move. "What do you want?" In Go, there is the concept of miai. If you play at A, I play at B. If you play at B, I play at A.
White plays the bizarre-looking attachment. Black is forced to reveal their intentions. Either they take the corner, giving White free moves on the outside, or they take the outside, and White can live inside the corner. It's never used so early in the game, but there are situations in which a probe is disruptive.
When White takes the outside, they usually sacrifice some stones, treating them lightly. White doesn't need to keep all of their stones, only some of them. Treating stones lightly means being ready to give them up.
White gave up a stone in this sequence, and is even ready to give more. But it's a very cheap investment for White, who can get stones flowing towards the side while Black continues to solidify what is already Black territory. The three White stones are like banana peels--slippery, but not something wants to take the time to pick up. In the endgame, White may finally save some of those stones.
Chess players already know about asking moves. But I don't really think that the concept of treating pieces lightly has caught on. Maybe it should. Whenever your opponent sac-sits against you, it's often good to treat some pieces lightly.
Black sacrificed. To clear the back ranks quickly, White launched his Bishop into the ether, giving back the piece, and giving Black a pawn in the end. But the Bishop check should be regarded as a probe. If Black blocks with a pieces, White is happy to trade, and keep a material advantage. If Black blocks with the pawn, White encourages Black to take it. Black takes two moves to weaken themselves on the dark squares.
If Black doesn't take the Bishop, White may eventually save it. But if White has the right outlook, White won't be worried about a single Bishop. In the above diagram, I think White's position is much better. White is a pawn down, but with no weaknesses, and more activity.
Sometimes probing moves should be used before going all-in with drops. Non-forcing moves give the opponent more chances to make mistakes, and they give the attacking player more ideas about what the right way in is going to be.
In this position, White played pawn at e7+. Black doesn't have a good response. The best is to take with the King, but then Queen at h4 for White trades off the Queens, and White lives easily. If Black played Ke8, White could play Knight at f6. In the game, Black played Nxe7 and Queen at d8 was a chess-mate. Black had no blocker.
In this position, White has a nice resource. Pawn at c6 forces Black to decide between breaking the queenside pawns, or bringing out the King.
Once you know about hitting c6, you should never miss it.
White can play Bxc6+ without thinking. To place a Knight on c5 would be harmless. Bxc6+ makes the Knight drop much stronger, because with broken pawns, White can drop a pawn at b7. Black does not have to play bxc6. Ke7 is possible, but even here White has prevented Black's King from finding safety on the Queenside.
5. Capping Moves
Capping moves keep the opponent contained.
If Black had a stone where the marked stone is, Black would be crushing. The right side would be too big for White to ever catch up on points. White observes the Go proverb, my opponent's vital point is my vital point, and plays in the same spot Black would play in.
Does this really exist in bughouse? I think so.
Not only is f7 under threat, Black cannot play e5, and this hinders Black's whole game. Instead of a6, many strong players would just sacrifice the e pawn as Black.
The pawn can be taken in two ways, but Black prefers to lose a pawn than to suffer closed lines. In the highest level of bughouse, players treat center pawns very lightly. For White, Nxe5 or dxe5 are both non-line-opening moves.
Beyond techniques and concepts, the culture of Go is also interesting. Professional Go players not only have ratings which are fluid, but they have more static ranks. The highest rank is 9P, or 9 Professional Dan. The professional dan ranks are tied to rating and to achievements, but professional players don't get demoted even if their rating goes down. In Go, there are Kyu players, with the strongest being 1 Kyu. Then there are the amateur Dans, where 1 Dan is the weakest and 9 Dan is the highest. Then professional Dans. I wonder what this system could do for bughouse or for chess. (Or perhaps what it would do to them.)
Go also does not have a single World Championship Title. There are many localized titles. Some players are able to collect many titles at a time. There is a highest rated player, right now Shin Jinseo from South Korea, who is rated 3863. Historically, there have been more players with Morphy-like dominance in Go--players with no serious rivals. Perhaps there is more scope in Go for strength.
Perhaps bughouse doesn't deserve to be mentioned in the same company as Go. Bughouse is a team game, and there are no full-time professional players. But becoming strong in bughouse can teach you a lot. It can teach you how to become strong at other strategy games.
