Nothing against Iyama of course but now that he has achieved the grand slam of Japanese titles, I kind of want other players to win some titles, especially "new faces" like Ichiriki. I feel this can only make the go scene more interesting.

I think such odds which assume games are independent events are not particularly useful in a multi-game title match where psychological factors, form, momentum and such like come to the fore.

As for the game, I am puzzled by Takao's peep of move 71. I can see that if Iyama answers it then it is a teensy bit good exchange (because if white wins the ko his connection becomes wasted, but black's move has little value too as white wedging into the kosumi achieves little), but it gives Iyama a chance to magnify the ko. The result in which Iyama doesn't start the ko but plays an atari to create bigger threats (saving a wasted move inside if he loses, good technique to remember), Takao resolves the ko and Iyama gets 2 stones in the centre and sente is very difficult for me to judge, but if it is ok/good for Iyama why did Takao peep because the left side fighting seemed quite ok for him (maybe my judgement wrong here)? I thought Iyama's peep of 56 was strange because it wasn't sente as he hadn't made the atari connect exchange earlier, so if he wanted that peep why did hs solid connect instead of atari for 48? He didn't want to give black the choice to take corner?

Yeah. Honestly Iyama should lose all his titles and start trying his luck in international titles. Then we can know whether he's top class or not after all.

If he only starts play international tournaments when he already lost all his titles, I guess he would be no longer competitive at that moment, like Cho U now. But I do hope that the young guns could manage to take some titles from him (Ichikiri and Yu are actually good candidates, their recent international performances are pretty good), so we no longer see the usual "schedule conflict" reason anymore.

Yeah, if he starts losing his titles because he's getting too old. Which is why I want him to lose his titles now, as unlikely as it may be.

Ichiriki Ryo is a beast. He should take some of the burden (big 7 titles) off of Iyama's shoulders

Not useful how?

I agree that the actual objective odds of a win vary quite a bit based on factors that a rating model doesn't use.[0]

If we knew the players' sleep levels, stresses, facts about which openings they prepared, we could do better. I also think there is such a thing as momentum, but I'm not sure we could make reliable judgments about it even if we spent lots of time shadowing the players in the days before the game. Players are streaky, but I'm not sure whether we have the ability to say Iyama's last two wins are signs that he's on a good streak as opposed to reverting to normal behavior. Or if a player is on a streak of playing well, how do we predict when it will end? Maybe someone can do reliably make that prediction, but I know it's not me, and I'm not sure if it's anyone else who posts on these boards.

[0] In fact, aren't they always .9999999999 or .00000000001 or whatever? I suspect that quantum mechanics doesn't really introduce significant non-determinism into how a go match proceeds.

_________________
Occupy Babel!

I seem to recall reading (maybe Tournament Go 1992?) that if one player wins the first 3 games of a 7 game match and then the other the next 3 in most cases that 2nd comeback player keeps the momentum and wins the decider. However clicking through Mr Kin's site I find 2 exceptions: Iyama winning the 39th Kisei against Yamashita and Yamashita winning the 66th Honinbo against Hane, and only 1 conforming with Hane winning 63rd Honinbo against Takao. Maybe some kind forum member has stats going back further.

The Japanese seem to be really big on this "nagare". I personally don't believe in it, other than the placebo effects it might have.

I recall reading some time ago (years?) an article analyzing streaks on basketball free throws (i.e., if they occur more often than chance would predict, and by how much), but I don't remember what the conclusion was. Someone might be able to dig it up.

In any case, I don't think it totally applies here: go players might figure out the best way to adjust to a particular opponent, while basketball players don't have issues with a particularly tricky hoop.

@Uberdude Three examples of cases where someone came back from 3-0 are on the Honinbo page at Senseis. http://senseis.xmp.net/?Honinbo#8. I thought there was a complete list of all times it had happened somewhere on SL, but I can't find it. That doesn't tell us how many times were near-misses, though.

@uPWarrior: viewtopic.php?f=8&t=12617

_________________
Occupy Babel!

Thanks hyperape, looks like it was less than a year ago after all.

In Go and in other things, the past doesn't predict the future very well.

Cheers,
Vesa

What do you mean by "top class"? If by that you mean top 15 or 20 in the world, then there is probably no need for Iyama to play internationally to find out because he doesn't belong.

I missed you by78.

What do you mean by "pretty good"? Ichiriki recently won a game against Lee Sedol by 0.5, and that's about the only notable win I could find between Ichiriki and Yu against international competition from Korea and China. Looking through their records going two years back, they both lost consistently against middling competitors from China and Korea.

Keep looking.

(For the record, I won't bet on Ichiriki being competitive with players like Shin Jinseo or Li Qincheng, but he's young and still improving so we can't rule him out just yet).

_________________
Occupy Babel!

One swallow doesn't make a summer but more than one Japanese swallow can be seen swooping around at the moment.

In the Jastec Cup last month: Hirata beat Li Qincheng and Min Sang-yeon; Yu beat Tang Mengcheng, Song Chi-hun and Li Yixiang; Ueno beat Song Ronghui; Shibana beat Yang Bowei; Son beat Hwang Chae-yeon; and Ida beat Wang Yuanjun.

Not earth-shattering, but there's a little bit of a rumble there that didn't seem to exist a short while back. Maybe the longish time limits of the Jastec helped.

Li Yixiang is Taiwanese, as is Yang Bowei, and Song Ronghui is a female Chinese pro... The rest are middling competitors at best.

The Japanese will get my attention when they start winning some international tournaments such as Samsung Cup and LG Cup, where top Korean and Chinese pros participate.

Riiight, and there are still people looking for unicorns. You think they'll find them?

Well, perhaps I pay more attention on their wins whereas you pay more attention on their loss. I don't say that they are already on par with, or better than Li Quincheng or Shin Jiseo (like I never said that Iyama is stronger than Ke Jie or Park Junghwan, or Lee Sedol), but I would expect that if Ichikiri and Yu (and even the others like Kyo Kagen, Shibano or Hirata) can play more international tournaments in the near future, their level will be not far from their Chinese/Korean counterpart. If you want a list of wins:

- Ichikiri won 3 games in the Nongshim Cup last year (1 in the year before), got a 6-1 record in the Chinese C League. IF you still argue that those opponents are only "weak Chinese/Korean pros", he just went 1-1 with Yang Dingxin in 2016 (a top 20 Chinese, 34th on goratings), and kicked out Li Quincheng in the Samsung Cup few weeks before his game with Lee Sedol.

- Yu lost a bad game to Lee Sedol few months ago, but he won to both Chinese and Korean participants in the Asian New Star match recently (Dong Mengcheng is a strong guy who just kicked out both Shi Yue and Gu Li just before the match). Yu also beat Peng Liyao and Lee Donghun (no.2 Korean youngsters) in the LG Cup last year, before failing to the runner-up Park Younghun.

About the chance of a Japanese win a Samsung or LG Cup, well, as their young talents are now at best world top 50-100 level, I would say that the only guy who has a slim chance is still Iyama, but perhaps he needs to play like 10-20 tournaments to really have a chance to win one (so do most of the strong Korean/Chinese players outside top 5).