Hi,

I have by no means any comprehensive knowledge about trends in pro games, but based on the limited knowledge I do have, skimming
through most recent pro games and several dozen Go World issues from years 1999 - 2010, I can't help but notice the absence of
3-3 and 3-5 plays in the opening. Correct me if I'm wrong - you can see an occasional 5-4, but 3-5 are almost non-existent, as
are 3-3 point plays.

Are these points inferior by today's top players' standards? Why is 3-4 and 4-4 favored so much over the other two points nowadays?

I'd love to hear both speculations and well-grounded arguments about this fact (if there are such). I, for one, heard an interesting
argument against 3-5 in contemporary pros' play a while ago, but I'd like to hear others' thoughts first (maybe
there's a well-known explanation I've been oblivious to?)

_________________

Do you mean for move 1/2? Or anytime during the fuseki. White's move 2 at san-san is not that uncommon, from what I can tell. Interestingly, it seems to invariably be Japanese players. Kono Rin, for instance, seems quite partial to it. If you allow it to be move 4 or later then you get some Chinese players and Korean players playing it.

I can't find any instances of it played as move 1 (but smartgo seemed to get confused by this, so maybe it's a software fault).

Move 1 at 5-3 is pretty rare, but it's been played at least a few times in recent years, e.g. by Choi Cheolhan in 2009.

btw, while looking, I found this remarkable opening from Yoda Norimoto last year:

Click Here To Show Diagram Code

[go]$$cm1 Prisoner Count: B-0 W-0.5
$$ ---------------------------------------
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . 4 . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . 3 . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . , . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . 1 . . . . . . 5 . . |
$$ | . . . 2 . . . . . , . . . . . , . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . . . . . . . |
$$ ---------------------------------------[/go]

Both happen plenty, the 3-5 rather more commonly than 3-3 (as I remember). It is true, though, that other points are much more favoured.

I think it is worth noting that the 5-3 and 4-3 are very much entangled. It is quite common for professionals to tenuki an approach to 4-3 (not that I know exactly how common), which leaves 5-3 patterns if the opponent gets back first. Also, long pincers to 4-3 low approaches leave essentially a 5-3 for the opponent to develop in the corner, but with a checking extension already in place. Due to these two effects, I think the knowledge base of the pure 5-3 point is used far more than might be obvious from 'first corner play' alone.

I did a quick search for you in my database (around 50.000 games)

there are almost 500 games with 3-3 as first move.
almost 6000 with a 3-3 in an otherwise empty quarter of the board
basically it means 3-3 was an opening move from at least one of the players in those games.
This in turn means that close to 12% of the games in the database has a 3-3 move in an empty corner.

there are almost 800 games with 3-5 as first move.
almost 7000 with a 3-5 in an otherwise empty quarter of the board.

_________________
Laurens
Go school: freegame's Teaching School
Author of: The Next Move a book for 15-3kyu players.

freegame, I think the poster was referring to recent trends, perhaps such as the last 5 or 10 years.

It would also be interesting to get an idea of what proportion of empty corner moves 3-3 or 5-3 was. Would it be as easy to taking the number of games and multipling by 4? I'm not sure it's that simple. But if it is, using your full dataset, 3-3 was played 6000/200000 times (0.3%) and 5-3 was played 7000/200000 (0.4%)

Nitpicking as always: 6000 / 200000 = 3%, which is rare but still worth mentioning. But if you can look at the historical trend, i guess it was popular in the shinfuseki time and maybe in 70ies 80ies but not nowadays.

PS 33PointJosekis

"(...) komi is a bigger advantage when neither player has much territory."

Damn, it never occurred to me to think of it that way.

_________________
So you've got an eye?
That don't impress me much

A Korean professional gave a lecture on new openings at the 2010 U.S. Go Congress and this question did come up.

Essentially, it relates to the high value of komi in modern games. Black essentially has to be more active and efficient than before and this drives black's strategy in a game with high komi. Openings are constantly evolving and he most of the innovations in the past few years have been coming from China.

But he didn't say much more than that, other than expressing strong opinions that American go players were still playing openings that were very old because they learned from old Japanese books. I showed him a copy of "Jungsuk in our time" (published in 2000) and he even said that was very old. He said he had written many go books, but they are out of date by the time they are finished, so professional players in Korea and China rely on magazines to keep up with current trends. He said he would try to help get English translations of some articles for the AGA E-journal, but I have not seen this yet.

I'll say a little bit more later, but I'd like to hear what other people think about why komi relates to this. You may know, for example, that Shusaku's kosumi is now being played by white...

Thanks snorri, that's an interesting insight. I guess the explanation concerning komi is one that could be expected, as komi is the only major factor in go that has been changing since its introduction in the 20s-30s. How exactly this relates to 3-3 and 3-5 being no longer optimal is too subtle for me to grasp.

Incidentally, a Korean pro writes in a book about joseki that the increasing komi is also the factor responsible for joseki development - often some sequences previously regarded as even are later reevaluated as better for one side or the other, and new joseki moves are researched.

Yes, I mean 3-3 and 3-5 played as one of the first 4 moves in an empty corner being vastly outnumbered by 4-4 and 3-4 plays. They do come up, but rarely, I don't have the exact proportion estimate, but whichever random sample of pro games from the last, let's say, 5 years I've seen - I can't recall a single case of them being played (which may just prove I replay too few games )


That's true, but still, browsing through Shusaku-era games and contemporary kifu one can see how seldom the 3-5 - low approach - pincer variations are played nowadays compared to the Edo period. For one, the komoku can and is very often met by a high approach move which excludes any 3-5 variations.

The argument I heard against 3-5 is simply this: variations that follow 3-5 (after a low approach and a pincer) belong to the richest and most complicated in go, and pros nowadays tend to avoid overcomplicated patterns in the fuseki, which is caused by game times being shorter than before (with the exception maybe of the Japanese 2-day matches).

_________________

As has been pointed out these moves are played in some modern pro games, just not very many. I can't say much about the mokuhazushi but I think the san-san move is too slow. That's because it is hard to develop from the san-san stone compared to the hoshi and komoku. The point about komi is relevant here, too.

So actually, the pro I mentioned was Kim Seong-ryong http://gosensations.com/index.php?id=2&server_id=1&new_id=664.

The impression he gave was pretty didactic. Pretty much, he suggested there are only two moves for black in the first empty corner: 4-4 and 3-4. Others are really played less than 1% of the time. I already knew that to be professional practice, but it's good to check.

If I search GoGOD Kombilo just for the year 2009, I get:

Total Games: 2105
4-4: 1245
3-4: 849
5-3: 6
5-4: 3
3-3: 1
8-8: 1 (That was Gan Siyang 3p, who has been mentioned before in other threads.)

So we have 11 first moves out of 2105 and if you don't count amateur games there are even fewer: it drops down to 8. And most of those 8 are in short-play tournaments like the KBS cup or the Korean Baduk League.

Openings such as the mini-chinese and low chinese are still popular, but some openings that used to popular in the past are getting kind of scarce.

Of that 2009 set, many of the popular openings of the past are now close to or below the 1% category:

San-ren-sei: 20
Kobayashi: 23 (of which 20 have move 7 on the side star point.)
High chinese: 10

Nice stats, snorri.

One caveat: it really makes sense to look at Black's first two moves. I read that if you're going to play a 4-4 and a 3-5, or a 3-4, it generally makes sense to play the 44 first, because it isn't biased towards one side or another. That prevents white from making an advantageous play based on the directional bias of the 3-4 or 3-5. Then Black can play the biased move with impunity at .

If that's accurate, then you'd want to look at the percentage of games where there is a 3-5 on the board by move 3.

_________________
Occupy Babel!

OK, so my gut feeling based on the limited number of games I saw was spot on

Now, the question is: why is it so exactly? Because of the komi, but can we dig deeper, get some fuseki examples maybe, which show the inferiority of the 3-3 and 3-5?

Yes, 3-3 is slow, but it has been habitually played by many great players back in the day (like the aforementioned Cho Chikun, for instance) and it was played extensively in the New Fuseki period, wasn't it? So, after all this experimenting could the final consensus be that it's inferior after all?

A side note about komi: it's been increasing gradually but it won't be increasing forever. Based on the info at Sensei's Library, the optimal komi is thought to be something between 6 and 7 (assuming Japanese scoring rules), and for example, 9 is considered to be way too much.

_________________

I don't think 'it's not played' necessarily implies 'it's inferior' in Go.

I agree with amnal's post above. Another point worth making is that there are fads and fashions in go, just like there are in clothing, automobile design, etc. In the 1970's and 80's there were fads for the Chinese style opening and san-ren-sei. Then came Kobayashi style, mini-Chinese, etc. Since then some of these openings haven't been played so much. It isn't because the openings are intrinsically bad. Maybe people just like to play something different once a particular opening is well understood. It might be as mundane as people all wanting to wear the same sort of distressed blue jeans. After a while that style becomes boring and people want something different.

Probably the win/loss statistics.

From everything I've ever heard about professionals, that's just not how they think. They play to win, not for entertainment.

Scientists search for truth and businesses try to maximize profits, yet there are fads in scientific theories and in management. Just because there's an objective goal doesn't mean there can't be fads.

I think one should distinguish between fads and learning. Sanrensei was an experiment that doesn't seem to have worked out that well. It's playable, but probably suboptimal. On the other hand, various openings are popular, then fall out of style without professional judgment really crystalizing the they're bad.

And lastly, is the low chinese really gone? I thought it was still played, if not as much as at its heyday.

_________________
Occupy Babel!

From GoGoD, since 2000 there have been an even 1200 games played using the low Chinese, with White winning 51.1%, but there have only been 241 san-ren-sei games with White winning 52.3%. The curious thing about san-ren-sei was that overall the percentages always seemed to favour White, except that Takemiya seemed to do reasonably well. If he was excluded, the percentages favoured White even more.

Best wishes.

_________________
No aji, keshi, kifu or kikashi has been harmed in the compiling of this post.
http://www.gogod.co.uk

The luxury of being an amateur is that we can play any of these openings, and I personally think we'd be remiss not to do so.

Of course, I'll still pay attention to modern pro openings.