19th Nongshim Cup

[Uberdude]

Anyway, third stage February 26 - March 2, 2018 at Shanghai. VS. my favorite player Iyama Yuta if Japan really win this there'll be no doubt of his international performance for a decade.

I think Iyama (#3 3587) could get a win or two, but little chance to plough through the rest of the competition Nie Weiping style. To do that he'd have to beat the following. I've put rating and win % from goratings in brackets, I'm sure ewan/by78 will come along to gleefully tell us how goratings overranks Iyama:
Dang Yifei (#46, 3402, 74%)
Kim Jiseok (#6, 3536, 57%)
Ke Jie (#1, 3640, 42%)
Shin Jinseo (#4, 3566, 53%)
Park Junghwan (#2, 3630, 44%)
Win all: 4%. (a bit more than the 3% you get for 0.5^5 for all matches even).

Mamumamu ratings (at end Oct):
Iyama (#27, 9.80)
Dang (#43, 9.57, 56%)
Kim (#13 10.058, 42%)
Ke (#1, 10.86, 24%)
Shin (#5, 10.29, 37%)
Park (#2, 10.70, 27%)
Win all: 0.6%

[pookpooi]

Yes, I know the difficulty. That is why if he can do it, no one will asking about this topic ever again. (And we can close iyama Yuta world ranking discussion thread forever)

[kimidori]

Well, if Iyama can really win it, I guess someone will open a thread saying that he indeed deserves the #1 spots on mamu or Dr.Tail Bae ranking

Anyway I tend to agree with Uberdude regarding to his chance. Japan still hasn't scored any win this year, so now all the pressure is on him. He is somehow the slight favourite to stop Dang Yifei, and if he lives up to the expectation, then he can think about getting his revenge on Kim Jiseok (who beat Iyama the last time they met at the Nong Shim Cup). After it will be extremely tough.

[pookpooi]

Sina analysis establish Dang Yifei as hero because he has to win otherwise Chinese team will failed because Ke Jie is currently in 'poor state'. I agree with them that Ke Jie has poor form now, but I think he can recover when around February.

[xiayun]

Ke Jie's "poor form" is a bit exaggerated. Yes, he has lost 7 of his last 10, but all of them against quality opponents. The lowest ranked player he lost too since October is Jiang Weijie, a former world champion who is still in his mid-20s. Ke Jie has definitely been underperforming against his expected wins recently, but it's not unusual to see a top player go through such a stretch. He himself had a stretch earlier this year where he lost 4 out of 7 and 6 of out 13, before going on that 22-game winning streak.

[idontgetit]

his poor form is for the whole year, I think. Compared to '15, he's doing worse in the A leagues, he got 0 int'l titles (possibly 1, compared to 3 in '15), it is... kinda disappointing.

Plus, I don't know, he just doesn't give off a really safe, sturdy feeling. Like, if it was Lee Changho in his prime, even if the other country still had 3 or 4 players, whatever. Stone buddha is still going to win it for us.

But Is Ke Jie going to plow through Iyama/Shin/Park? I hope so, but I really doubt it.

As for Iyama, I don't know why pookpooi keeps on wanting people to recognize him as the world no.1 or something. I mean I think it's pretty clear his rating is slightly inflated on goratings. but really, in a relatively fast tournament like the nongshim cup, all of the players in the top 50ish would have close to a 50% chance of winning against each other. A lot more of a luck/"flow" element in it; like it depends on the conditions of the person that day, and whether they are too nervous, or excited, or focused or whatever.

[WindCaliber]

his poor form is for the whole year, I think. Compared to '15, he's doing worse in the A leagues, he got 0 int'l titles (possibly 1, compared to 3 in '15), it is... kinda disappointing.

Plus, I don't know, he just doesn't give off a really safe, sturdy feeling. Like, if it was Lee Changho in his prime, even if the other country still had 3 or 4 players, whatever. Stone buddha is still going to win it for us.

But Is Ke Jie going to plow through Iyama/Shin/Park? I hope so, but I really doubt it.

Your information is incorrect. Counting by when he actually won the final game, he won 2 international titles in 2015, 2 in 2016, and he won the Limin Cup this year. Thus, he is on track to win 2 a year for 3 years in a row. I say he is on track because he is facing Zhou Ruiyang in the finals of the Xin'ao cup, and he's currently 8-1 against Zhou. As for his record this year, it's actually not bad. Up until the past couple of games, his record was significantly better than last year, and is still slightly better even with the recent losses. Let's not forget his 22-game winning streak earlier this year.

That being said, with Korea fielding their #1, #2, and #3 players, I would say that they are clearly the heavy favorites.

[Uberdude]

WindCaliber, your information is incorrect too . It is Peng Liyao 5p rather than Zhou Ruiyang that Ke will face in the Xinao final. Of games in go4go Ke has 3-1 record plus Peng is lower rated so Ke is still heavy favourite.

[WindCaliber]

WindCaliber, your information is incorrect too . It is Peng Liyao 5p rather than Zhou Ruiyang that Ke will face in the Xinao final. Of games in go4go Ke has 3-1 record plus Peng is lower rated so Ke is still heavy favourite.

Whoops, sorry for the hallucination!
I must have misread it.

[John Fairbairn]

That being said, with Korea fielding their #1, #2, and #3 players, I would say that they are clearly the heavy favorites.

Of games in go4go Ke has 3-1 record plus Peng is lower rated so Ke is still heavy favourite

We seem to quite different interpretations here, and if you look at Ke Jie's stats even more interpretations fly off the page. We also seem to have lots of mathematicians/statisticians on this forum (nothing's perfect )so perhaps someone can tell me how we are supposed to read stats like this. I suspect the answer is "with a salt mine full of NaCl" but still...

Some of the other stats to take into account for Ke Jie:

Against the two players who were so raved about very recently that a massively sponsored game was arranged between them (and also raved about on this forum very recently so that a huge book of the match appeared in English), Ke Jie has a massive plus: 8-0 against Gu Li and 11-3 against Yi Se-tol.

But against the players he has played more often he has much more modest scores: 11-8 vs Chen Yaoye, 10-8 against Shi Yue and 9-6 vs Tang Weixing. He even has negative scores against Qiu Jun and Jiang Weijie (not to mention Pak Cheong-hwan) and Tan Xiao has split him on 6-6.

He even has 0% against quite a few players. I'm relying on memory, but one of the features of Go Seigen's stats was that he only ever had negative scores against a very tiny number of players - much fewer than Ke Jie is recording, anyway.

So, how are we supposed to make sense of all this beyond a vague "Ke Jie seems a fairly strong player"?

In vaguely similar territory, it has long been a given among top pros that if they had to play God to save their lives, they would need four stones to feel totally confident of winning. It seems from the latest server reports that some pros are losing on four stones to a bot. Looks like pros may have got more of their understanding of the game wrong than they thought.

Never trust numbers, I say

[Uberdude]

That being said, with Korea fielding their #1, #2, and #3 players, I would say that they are clearly the heavy favorites.

Of games in go4go Ke has 3-1 record plus Peng is lower rated so Ke is still heavy favourite

We seem to quite different interpretations here,

In case it wasn't clear, I meant Ke is heavy favourite in the Xinao cup final against Peng Liyao (and therefore end 2017 with 1 or 2 international titles (depending if you count young-only Limin) which doesn't look so bad compared to what he won in 2015/16), not to win the Nongshim cup all by himself.

[xiayun]

Not that he cares, but Ke Jie also achieved his personal high in Goratings ranking this year and looked almost invincible at one point, when Gu Li jokingly said he was "half AlphaGo, half human". He just hasn't necessarily won the right games. To be fair, for the 5 international titles he has won, he could've easily lost to Qiu Jun and Lee Sedol, and as Ke Jie himself admitted, Tuo Jiaxi outplayed him for the better part of their best-of-3 title match, so luck always plays a factor.

[pookpooi]

It seems from the latest server reports that some pros are losing on four stones to a bot. Looks like pros may have got more of their understanding of the game wrong than they thought.

Never trust numbers, I say

They already expand the upper limit. I can't remember if Shin Jinseo or who said that with 6 stones handicap he'd bet everything including his life as well. While I can't imagine how he'll lose with that condition in the near future, but the lesson from AlphaGo is just like you said. Never say never!

[gowan]

Well we mustn't forget that in human-against-human games psychology and emotions play a big role. AI doesn't have these issues. The chess grandmaster Emmanuel Lasker was famous for playing the opponent not the board. In go matches the same sort of tactics are used. One player is known to favor the san-ren-sei opening as black so opponents might play it against him. Against an AI such tactics have no effect. In a face to face go game a player feeling herself behind might play moves she knows would likely not be correct but which will complicate the game, hoping the opponent will make a mistake in responding. Do top level AI players complicate the game when they are behind? Of course not, they are never behind against humans In chess again the best computer programs are able to defeat top level humans giving material handicaps. If high-ranked go pros are losing go games receiving a four stone handicap, think how much pressure the pro would be under in a six-handicap game.

[sorin]

It seems from the latest server reports that some pros are losing on four stones to a bot. Looks like pros may have got more of their understanding of the game wrong than they thought.

I don't think any pro lost to FineArt with 4 stones. Those are 9D online accounts, many can be 5-6D in Europe or so, not pros.

At the most recent US Go Congress, Andy Liu 1P from US said that he is losing against FineArt on Fox server on 2 stones consistently, but winning on 3 stones consistently.