Using my Winter 2010 copy of GoGoD, I analyzed the first four moves of games played in 2010. (GoGoD Winter 2010 goes up to 4 December 2010). This is 2244 games.

In selecting popular sequences of the first 4 moves, I eliminated any that appear in less than 1% of the games. The 18 openings represent 74% of all games in that set. The top 6 openings account for 48% of all games, and they all start with black's first move on the 4-4 point. The top 3 account for 1/3 of all games.

Some properties of popular modern openings:

1. With one exception (R16 D16 Q3 P17, which has an early approach like in Shusaku style) every opening has both players playing in empty corners for the first 4 moves.

2. With one exception (the same R16 D16 Q3 P17), all of the first 4 moves are on either the 4-4 or the 3-4.

3. Some popular variants of the past, such as nirensei (parallel 4-4) or mukai komuku (opposing parallel 3-4) for black, simply didn't make the cut.

The following is the breakdown.

Q16 D16 R4 D3 285
Q16 D16 Q3 D4 242
Q16 D16 R4 D4 211
Q16 D4 R4 D17 132
Q16 D16 Q3 D3 100
Q16 D4 Q3 D16 99
R16 D16 Q3 D4 99
Q16 D4 R4 D16 80
Q16 D4 Q3 D17 70
R16 D16 Q4 D4 62
R16 Q4 D17 C4 45
R16 D16 Q4 D3 42
R16 D17 Q3 D4 41
R16 D17 Q4 D4 38
Q16 D16 Q3 C4 36
R16 D16 Q3 P17 28
R16 D17 D3 Q4 24
R16 Q4 D17 D4 24

I will post more thoughts later.

_________________
Help make L19 more organized. Make an index: https://lifein19x19.com/viewtopic.php?f=14&t=5207

Reviewing the patterns I listed in my first post can be an exercise. Let's start with the top 3.

Click Here To Show Diagram Code

[go]$$B
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , 3 . . |
$$ | . . . 4 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

White prevents mini-chinese with .

Click Here To Show Diagram Code

[go]$$B 19x19 diagram
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . a . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 4 . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . 3 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Black may next aim for a chinese or variant near 'a'.

Click Here To Show Diagram Code

[go]$$B 19x19 diagram
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 4 . . . . . , . . . . . , 3 . . |
$$ | . . . . . a . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

will allow an approach at 'a'.

In each of these cases, we see white which also allows black to play a diagonal opening. Interestingly, in none of the top 18 does black actually take white up on this offer if is on the 4-4 point. In fact there is only one true diagonal opening left, and with only 24 hits, it was close to missing the 1% cut:

Click Here To Show Diagram Code

[go]$$B 19x19 diagram
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , 1 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . 4 . . . |
$$ | . . . 3 . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

So are we witnessing the death of diagonal openings? It is looking like a trend.

You're saying the top 18 account for 74% of the reduced set, while the top 6 account for 48% of all games? Something about the phrasing is makling me second guess myself.

_________________
Occupy Babel!

Sorry, I should be more clear. All percentages are relative to the 2244 games from 2010 an that CD.

My favorite opening for both sides these days is this one:

Click Here To Show Diagram Code

[go]$$B 19x19 diagram
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 4 . . . . . , . . . . . , 3 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

As a side note, I have mixed feelings about a set of 2244 games. On one hand, it seems like a lot of games. On the other hand, there's a lot of variation in go, so 2244 doesn't seem like that much data.

However, we are only looking at the first few moves, so observing these tendencies is probably valuable.

_________________
be immersed

This goes past the first 4 moves, but I thought that it might be interesting to some people. In case anyone didn't get the memo, the Kobayashi is dead.

Click Here To Show Diagram Code

[go]$$B
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 2 . . . . . , . . . . . 1 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . 7 . . |
$$ | . . . . . . . . . . . . . . . . a . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . 6 . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 4 . . . . d c . . . . . , . . . |
$$ | . . . . . 5 . . . . . . . . . 3 . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

or 'a' are the surviving openings... 'c' and 'd' are down to curiousity level now.

Although I like looking at this kind of thing, actually I'd be one of the first the claim that it is not immensely valuable. The differences that show up in pro games in terms of win/loss stats don't seem to translate into, for example, high dan KGS games. Go is not chess. It is rare to see an opening get a clear refutation that's comprehensible by amateurs.

I think the main value may be for players who would like to be able to follow the lastest pro games and would like an update on what to expect. Maybe for very strong amateurs it may be possible to consider that if black plays an opening that is too slow for modern komi, simply playing solidly as white and seeking a close game is good enough. But if your endgame is not good, it may still be better for white to play to create confusion just like in pre-komi days.

I have no fantasies that following pro fashion makes me stronger. It's just a hobby. My wife watches "America's Next Top Model" and I watch Cho U and Lee Sedol. They are both indulgences.

They're on America's Next Top Model? I have to start watching that show!

_________________
Occupy Babel!

BTW, here is the data for KGS high dan amateurs from 2010. I selected 6966 games with no handicaps and 6.5 komi.

Q16 D4 Q3 D16 404
Q16 D16 Q3 D4 366
Q16 D4 Q3 D17 314
R16 D16 Q3 D4 206
R16 D17 Q3 D4 190
Q16 D17 Q3 D4 189
Q16 D4 R4 D17 181
Q16 D4 Q4 D17 175
Q16 D16 Q3 D3 170
Q16 D16 R4 D3 153
Q16 D4 R4 D16 151
Q16 D16 Q4 D3 138
Q16 D4 Q4 D16 106
Q16 D17 Q4 D4 105
Q16 D17 Q3 D3 104
R16 D16 Q3 D3 99
Q16 D16 Q4 D4 94
Q16 D16 R4 D4 77
Q16 D17 D4 Q4 75


The first thing to note is that there are 19 patterns that made the cut about 1%. More importantly, they only cover 47% of the games, which means the KGS results have a "long tail" with more options. However, the options here aren't very strange. Except for the fact that the KGS high dans are not abandoning the nirensei (parallel 4-4) or diagonal 4-4 patterns as much as pros have, the openings are similar. There are no moves other than 3-4 or 4-4, and they are all in empty corners with no early approaches.

I'm a bit surprised but perhaps shouldn't be. Some of these players were insei or had professional training, for example.

As usual I exaggerate for the purposes of stimulating discussion, but I think there is a bigger trend going on. It's to do with time limits.

In the good old days, the aim of the fuseki was to get to the endgame. Nowadays the aim is to get to the middle game.

When time limits were generous enough to ensure that pros could avoid most reading mistakes, they did not expect to win in the middle game. They hoped instead to emerge into the endgame with a thicker position. The fuseki stones were seen as vital in that process and it was considered very important that when the endgame started (about move 120) each fuseki stone should be seen to be in just the right place to contribute to that thick position.

Accordingly, a large proportion of a player's time was spent in the opening. This in large part stimulated interest in varying fusekis (which, in a vicious circle, meant spending even more time on the fuseki). Players (this was an era when it was all essentially about Japan) were aware of this because there was an explicit comparison with sumo. Because right/left-handedness comes into it, wrestlers tend to grapple in a standard way, and switching hands was done to inject an element of surprise. In go, however, playing the same way repeatedly was the thing that caused surprise. It was considered lazy and the sumo term for this was borrowed (namakura yotsu). In Sakata was an early proponent of this, and this was one of his early nicknames. In sumo, of course, if you were stronger or bigger than the opponent you could afford to be lazy, and maybe the signal Sakata was sending was that he was strong enough not to need to worry about the endgame. But the prevailing view, especially in a match, was that you had to vary your fusekis and rely on stamina for the endgame.

With modern go, especially the Mickey Mouse events in Korea, you just want to be the first to start a successful fight. It's basically a crapshoot, but at least if you can control when and how the fighting starts you may have an edge. Playing openings you are familiar with (not just fusekis but also new josekis and hamete moves) is one way of controlling the tipping point. There are other ways of doing this, of course - the old chestnut of grabbing territory and then relying on the one-weak-group-can-live strategy (significantly, associated most with Sakata) - is not just still around but is much more prevalent, but a nakamura yotsu fuseki seems to be the favoured strategy either way.

One interesting way to see the difference is to look at games at about move 120 and then highlight the fuseki stones. In the older games you can see how they tend to form linchpins in territories (or, if they don't, this is often part of why a player lost), but in modern games it is more chaotic. If anything the key fuseki moves are those that feature in the main fighting, which means a lot of emphasis on high stones.

Behind all of this, obviously, is also the observation that resignation is much more common in modern games with short time limits. More blunders or more accent on fighting? Probably both.

Is Mickey Mouse known for fighting or something? I suppose he does have the gloves for it. I'm also surprised that you feel that modern games are a "crapshoot". I'm sure that today's pros would be happy to hear that.

_________________
be immersed

i don't know much about Mickey Mouse, but from John Fairbairn's posts i understand he means short time limits. i suppose he uses the term 'Mickey Mouse' for reason that such games are meant only to fast/shortly entertain spectators, and can't be really taken serious

it is a recurring theme in JF's posts and i mostly agree with his opinion

_________________
Spilling gasoline feels good.

I might be wrong, but probably not.

So combined with:


the implication is that events in Korea "can't be really taken serious"?

_________________
be immersed

Depends on your definition of serious. With events made for TV, where the audience demands to be amused constantly, they enter byoyomi almost immediately. The probability of errors increases greatly, so the playing level simply drops.

So if your definition of serious games includes "high level of play", then those events cannot be taken as seriously. If your definition is only "the players are strong and do their utmost to win", then any pro event is serious.

There are many Korean events with longer time limits, so this discussion is kind of moot, anyway. It can be seen in other posts on this forum that some members here have "exaggerated" about the time limits used in Korean tournaments.

Ignoring that, though, it seems like I am in a minority in thinking that "high level of play" should also include time limits as a factor of the game state.

Being able to calculate quickly and accurately is an admirable skill, in my opinion. I think that such play can be taken seriously, and can in some cases provide more insight into "go skill" than a slower game could.

If we assume that longer time limits leads to more optimal play (which is still simply an assumption), you can still gain insight toward a player's thought process behind making a move, even if a particular sequence was not optimal.

I do not think that "strength" equates to "brute force". A brute force algorithm on a computer could work on solving a go position. It might take years to finish.

There's nothing special about that type of an algorithm, though. You could probably make that type of an algorithm with a couple of weeks of programming experience. When faced by the additional problem of having time as a constraint, you start to see some real magic happen.

And I'm still sure that modern pros would be happy to hear that their games are "crapshoots"... :-p

_________________
be immersed

http://senseis.xmp.net/?ProfessionalTournamentTimeLimits

(can't resist )

So I guess the next thing to do would be try this same analysis for 1865, and see how much more or less variation there is.

I do think that being able to handle sort time limits is an important skill on the part of the player. Where I am talking about "high level of play", I am talking about the resulting game, not about the skill of the players. IMO, the more errors a game contains, the lower the level of play of the game.

For a given player, it is possible that more time will allow you to weed out more errors (though, this is also not proven; The Art of Learning (http://www.amazon.com/Art-Learning-Jour ... 0743277457), for example, suggests that the author's best plays as a chess player were those that he spent a medium amount of time on. When he spent too much time on a given move, he often played poorly.).

But it seems a bit biased to claim that go events in Korea "can't be taken seriously". It's feasible that a game played at a go event in Korea is of a "higher level of play" - having fewer errors - than a go event played under slower time settings. With go events in Korea still having multiple hours for time limits, it is irrational to throw out the quality of the game simply because of the time limit. There's still a lot of time that was put into the game, and it is quite feasible that certain games played under such time limits can have fewer errors than those with greater time limits.

I could also say that games played by Japanese players "can't be taken seriously" because some of the said players have lost to Korean players in international tournaments. I won't say this, though, because I think that it's insulting to the players to say something like that.

But thinking only of the time controls:
1.) There isn't a huge difference in time between Korean pro events and Japanese pro events. Have you watched baduk TV? On a number of programs, the players still spend a considerable amount of time making moves. Sure, there are fast events as well, but this doesn't mean that games from Korean pros "can't be taken seriously".

2.) It has not been shown that, overall, games played under faster time settings in Korea contain fewer errors than games played under slower time settings in Japan.

_________________
be immersed

Hehe, I wrote about 500 words to try to argue against Herman and John, but ended up having to trash it because there were so many arguments (how level of play cannot be determined by number of errors because other factors such as length of game and complexity will tend to overshadow the effect of half a stone or so at pro levels, number of errors is not measurable, games with less detectable errors are not necessarily more fun to watch, reviews of Japanese top games such as title matches demonstrate a comparable number of errors, performance of Japanese players against Koreans and Chinese do not suggest that they will be better players even given long time limits, etc...) that I can't end up choosing one over the others (I guess in this case, the best defense is no defense?) :p And that's not even to get started on amateurs talking about how pro games cannot be taken seriously?

In the end, I make this post just so that my previous attempt is not entirely wasted (it's clear that I'm the type who values consistency over rationality) :p