Poll: Who was the best player in 2013?

[wineandgolover]

Okay, even I feel like Dr Bae Taeil is piling on now.

45'th for Iyama? Ouch. Good thing he's rich.

Trout, where did you find this? I did a cursory Google check before posting mine. Thanks.

(side note: my current favorite player to review, Ke Jie, is up to 26th. sweet!)

[trout]

It is from Cyberoro Korean site.

http://www.cyberoro.com/news/news_view. ... =3&cmt_n=0

[Mef]

Therefore, I do not think that a counting statistic based on international play (number of titles won) is a good metric for evaluating his strength relative to his international peers.

Thanks for replying, Mef. Out of interest, do you have a better metric?

I understand your argument, and it makes some sense. But really it smacks as more of an excuse than a justification. IMHO, you can't be considered the strongest if you can't, or choose not to, play and beat the big dogs.

Out of interest, does anybody know if Iyama played in the big 2013 international tournaments? Was he beaten, or merely absent?

Well, I never intended it to be a justification for why I think Iyama was the best (personally I don't think that), I just felt that it might not be giving those who do think he is the best a fair shake (because I'm sure they have their reasons).


As far as a metric for judging him...I can't say that I have one in hand, but I can think about ways I would go about it, unfortunately for many the data just might not be there. The rating system mentioned by hyperpape is one possibility, and certainly is grounded in a reasonable starting point though I think this will have large error bars for players with only limited sample sets.

Ideally we could come up with some kind of condition-normalized performance factor to rate him relative to other Japanese players, then evaluate the his performance relative to those who have a large number of international games. You could attempt to build some kind of projection system based around this...but it may be challenging as there might not be a good comparison for Iyama amongst Japanese players with large numbers of international games.

To continue the tangential analogy discussion:

But I think your baseball analogy isn't very useful for two primary reasons.
1) Baseball is a team sport, which by its nature is different than individual sports and games like go.
2) Today's era in go doesn't correspond with today's era in baseball at all. In baseball, you have several countries that play well, but only one that is dominant. Many, if not most, of the best players eventually go to MLB to prove themselves, and, yes, to earn the big bucks.

This is one of the reasons I tried to use pitchers as the basis for comparison. Pitchers performances are highly individual based and most modern pitcher evaluations will remove external effects (team defense, park factors, etc). A pitcher is a one on one contest much like go, there is perhaps a weakness in the analogy in that pitchers do not directly compete with each other instead they are evaluated by proxy. When evaluating how a pitcher is expected to perform, they use rate stats to assess his past performance (something like FIP, fielding independent pitching), regressed statistics to predict how he should have performed (xFIP - FIP but normalized for places the pitcher may have gotten lucky, like having a large number of flyballs but low number of home runs), and counting statistics trying to estimate overall value (WAR or Wins Above Replacement - it compares both rate stats and the total amount of time played in a season so that you get a total value produced). In theory you could try to derive similar metrics like this for go, normalizing wins based upon the context. Blitz wins might not count as much as wins with long timesettings, wins in later rounds may count more, wins may have more or less value based upon the state of a series (Is it easy to put away an opponent when they are facing elimination? or is that a game where the person who's ahead tends to take a breather and give up a loss?). Ideally putting everything into a normalized context could help you project expected performance, and maybe find places where players got lucky.

The other amusing reason I picked this analogy is because baseball pitching has this literal exact dilemma going on right now (which I alluded to in my previous post). There is a pitcher from the Japanese Leagues who was absolutely dominant last year and they need to project how well he will do when he makes the transition to MLB. Masahiro Tanaka went 24-0 and posted very strong "traditional" baseball stats over 212 innings pitched (2.29 FIP if you care to look at that...lower number is better, like an ERA). Of course the challenge is projecting how moving to bigger ballparks with stronger hitters and a ball with more bounce will affect his performance. Luckily there are people who run the numbers and do the projections, and they think he'll end up being somewhere around the 25th-35th best pitcher in the league.

The overall analogue to go is quite striking. Japan's baseball leagues are good, but generally considered not quite as strong as the US. There is only limited mixing between the two groups, but there are some comparisons that could be made for reference. Hiroki Kuroda is sometimes given as a good comparison for Masahiro Tanaka, he was about 22nd or 23rd in WAR last year depending on who you ask. If we use Dr Bae Taeil's ratings and we figure Iyama falls somewhere in the 20-50 range of best players, that's actually about what they're expecting from Tanaka as a pitcher both from projections and from comparisons...it's actually a little funny how these things seem to parallel. Of course the one problem we have moving over to go is that there really isn't a good recent comparison for Iyama that we can use to assess outcomes in international play, the last time someone was this dominant in Japan the international go scene was also dominated by Japan. Perhaps he will make the jump and we can use him as the baseline for future players.

[Mef]

Okay, even I feel like Dr Bae Taeil is piling on now.

45'th for Iyama? Ouch. Good thing he's rich.

So, I slept on this and it occurred to me...Iyama's rating drop may in part be an artifact of the Japanese system. When you first win titles you must beat all other challengers. This means that in the years you are acquiring titles you probably have ~75-85 winning percentage. To maintain the status of holding titles, you simply need better than 57%.

One might expect Iyama to have a higher rating when he's winning 75% of his games against the field compared to 60% of his games against a very limited top tier.

In contrast, most Chinese and Korean tournaments the winner comes up through the bracket each time. This means a top Korean/Chinese player gets to hold that 75% against everyone each year.

It may be that (paradoxically) losing a couple his titles may allow Iyama to climb back up the international ratings.

[hyperpape]

If so, that would be a damning indictment of his system. Such a system would be deeply deficient.

Luckily, I think it's not true. Iyama's win percentage in 2013 was still 70%.

[Mef]

If so, that would be a damning indictment of his system. Such a system would be deeply deficient.

Luckily, I think it's not true. Iyama's win percentage in 2013 was still 70%.

I don't think it would be an indictment of the system, so much as yet another manifestation of a classic problem in ratings-- How do you take limited data and accurately evaluate relative performance between two largely isolated populations?

Conceptually this problem is similar to something like "Who is the stronger player, Honinbo Shuei or Go Seigen?". The difference is that this is a political/geographical separation rather than a temporal one.

[hyperpape]

What I'm talking about is simpler: it's not about international comparisons at all. I'm saying that on your explanation, the ratings system thinks 57% against a title-challenger is worse than 70% against a lesser player. And getting that right is the entire point of a ratings system. That's not really anything to do with international comparisons.

[moboy78]

I cannot believe that people are actually saying Iyama Yuuta is the best player of 2013. That's like going back to the 1960s or 1970s and asking the same question and hearing someone answer Cho Namchul or Kim In. Sure, they did very well domestically, but put them up against someone like Rin Kaiho and the answer of who would be the better player would be answered pretty quickly.

[lemmata]

The Korean pros are almost unananimous in declaring Shi Yue as the most difficult Chinese opponent. In that sense, Shi Yue as the strongest player in the world seems to be a reasonable hypothesis. Pros certainly do not play enough games for any rating system to be anything better than a rough approximation, but the current ranking does not seem to be too bad even if the ranking methodology has issues.

Iyama's ranking receives a positive bonus due to playing championship games in tournaments that have huge prize totals. Dr. Bae's system gives bonuses for playing in title matches and also weights games based on the total prize pool of the tournament. In fact, Iyama's games receive the maximum possible bonuses in those dimensions. Iyama's ranking is low not because of the few bad results he has had in international play. In fact, I would say that he didn't play all that badly in international match-ups. His ranking is low because his Japanese opponents were crushed by Chinese and Korea players as a group.

One might claim that Iyama might be underrated because he does not play too many international matches. While he himself has not played too many international matches, his Japanese opponents have played a nontrivial number of international matches as a group. Many Japanese pros admit that Japan lags behind China and Korea at the moment. This does not mean the absolute skill gap is huge. For pros, being 1 to 2 points (points/moku NOT stones!!) weaker is more than enough to explain that difference.

One might also claim that Iyama's ranking suffers too much from the performance of his Japanese peers, but he is not dominating his peers to an extent that would make us believe that. Cho Hunhyun in his heyday may have been underrated, but Cho was one stone stronger than the second strongest player in Korea and perhaps two stones stronger than most Korean pros. Cho beat Seo Bongsoo in more than 2/3 of their match-ups. We cannot attribute that level of dominance to Iyama.

I would not be surprised if Iyama is actually the 10th strongest player in the world. That would certainly be strong enough to win an international tournament. However, I would be quite surprised if he was the best player in the world. Iyama might be underrated, but he's not number one.

[Mef]

I would not be surprised if Iyama is actually the 10th strongest player in the world. That would certainly be strong enough to win an international tournament. However, I would be quite surprised if he was the best player in the world. Iyama might be underrated, but he's not number one.

Not to disagree with you in any way, simply to add some ideas to the discussion --

The title of this thread is "Who was the best player in 2013" not "Who do we think the strongest player in the world is?"

I point this out because there is an important distinction here. In an analogy that will surprise no one, this is another problem that comes up frequently in baseball: The difference between who performed the best in a given year vs. who is the best player in a given year. These might sound like this same, but they are not quite the same thing. Again, in baseball for baseball it's easy to imagine the difference -- Think about a player who hits 25 home runs...now think about a player who hits 25 game-winning home runs. Both have demonstrated equal hitting skill, but one has situationally had a much better performance.

The same could be said for go -- imagine a player who wins 3 major international titles but loses all games outside of those tournaments. If you go 15-85, is this a poor year (because you lost 85% of your games) or a great one (because you won 3 international titles)? In the same vein, I think it could be quite possible to have the best player of a given year not necessarily be the one who is considered the strongest. At its core we are talking about two separate ideas: 1: Who had the best results of the year and was most successful now vs. 2: Who do we believe has the highest true talent level, and thus in future years is predicted to be most likely to have success?

[moboy78]

I think that if we are arguing who the best player is based on number of important wins and total wins, then maybe Iyama would come out on top. It is important to remember that Iyama only has to defend many of his titles, and doesn't have to beat every other Japanese player in the league or tournament that makes him the challenger for a title. But in international tournaments, the previous winner has to fight his way back to the finals, making any sort of continuous victories in the same tournament day more impressive than Iyama's victories. It takes more skill for pros to beat a continually changing group of people year after year than it does for Iyama to beat the same ten people who play him in domestic title matches and continually lose to him.
I'm not saying that Iyama's victories aren't impressive and that he isn't a strong player, I'm merely saying that his wins are just against people he's shown he's better than and do less to salvage Japan's reputation on the international scene than the wins of other players do for their own country and records.

[lemmata]

...The difference between who performed the best in a given year vs. who is the best player in a given year...

I would agree that these are meaningful distinctions. Dr. Bae's ranking tries to approximate something closer to the former. I believe that his ranking considers results from the past three years with older results being weighted less.

My own view of the differences falls roughly (very roughly!!) along the following lines.

1. Who had the best skill in 2013: Who, if he played every player exactly 10000 times under identical conditions, would have the highest winning percentage?
2. Who performed the best cumulatively in 2013: Who, in the games he actually played, had the highest number of "quality-adjusted" (based on strength of opponents) wins?
3. Who performed the best on average in 2013: Who, in the games he actually played, had the highest number of "quality-adjusted" (based on strength of opponents) wins divided by the number of games he played.

It is difficult for me (and many others) to rank Iyama higher than 10th under any of those definitions.

Now, if we asked who had the best financial performance in 2013, the answer would be Mr. Iyama Yuta for sure.

[macelee]

Check the latest Bailing Cup schedule

http://www.lifein19x19.com/forum/viewto ... =13&t=9960

Can you guess who is not there?

[gowan]

I would argue that the best player is not necessarily the one who wins the most tournaments or the most money. All these rating systems are flawed and unreliable, especially in attempting to compare members of groups that do not play each other very much. In any case I think the best player is the one with the deepest understanding of the game. There seems to be wide acceptance of the statement that Go Seigen was the best player in the world from approximately the late 1930's to the 1950's or thereabouts. But how many tournaments did he win? Winning a long match is different from winning a tournament. Go Seigen made many innovations in joseki, opening theory, etc. Winning a tournament requires freedom from blunders but great creativity seems to me to be accompanied by frequency of blunders, consider a great player like Fujisawa Hideyuki.

[cdybeijing]

I would argue that the best player is not necessarily the one who wins the most tournaments or the most money. All these rating systems are flawed and unreliable, especially in attempting to compare members of groups that do not play each other very much. In any case I think the best player is the one with the deepest understanding of the game. There seems to be wide acceptance of the statement that Go Seigen was the best player in the world from approximately the late 1930's to the 1950's or thereabouts. But how many tournaments did he win? Winning a long match is different from winning a tournament. Go Seigen made many innovations in joseki, opening theory, etc. Winning a tournament requires freedom from blunders but great creativity seems to me to be accompanied by frequency of blunders, consider a great player like Fujisawa Hideyuki.

Great creativity is one thing, certainly worthy of admiration, but if you blunder regularly how can you be considered the best player in the world?