Maybe, but AlphaGo needs to play more games, right now it's #2 on goratings and Ke Jie is #1.

After the Chunlan semifinal against Park Yeonghun on 12/22, Ke Jie will continue his crazy month with the start of Mingren preliminary on 12/24; the first opponent there will be Li Xuanhao. If he wins, the next match will be two days later against the winner of Gu Zihao and Fan Yunruo.

Ke Jie got through the first two rounds of Chinese Mingren and into the quarterfinal.

Ke Jie's second game was an interesting moyo and invasion one: as white he gave black a big moyo and built his own, black invaded and ended up living but at enough cost to an outside position that Ke came out ahead.


Also Shi Yue vs Zho Ruiyang was a very interesting game that seemed to be AlphaGo inspired, I posted about that here: viewtopic.php?p=214957#p214957.

Yeah, that ko to survive was costly for black, as it ended up giving white not one, but two free moves and costed its bottom group.

What would have happened if black played m16 at m3 instead to defend the bottom group? I imagine white would make some reduction into the top side moyo using the aji of those stones in the centre, but where and how much of a reduction? Losing that lower group was big, and ending in gote at o5 seemed pretty miserable to me.

White could also make more territory at the top if black was forced to defend further reduction of the moyo after w's initial move.

Next month's schedule for Ke Jie: Mingren quarterfinal/semi/final, Tianyuan (will get the entire bracket played over 6 days), and CCTV Lunar New Year Invitational (an exhibition tournament with Ke Jie/Park Junghwan/Iyama Yuta).

Could someone list some of Ke Jie's greatest games to look closer at?

Cheers

I decided to do a quick check on how Ke Jie compares to Park Junghwan for the past three years (I included also Shin Jinseo who is only 16 years old and rose sharply each past year). For reference, I included Lee Chango stats during the same age interval as Ke Jie (17-19 years old) and Lee Sedol stats during his best three years (so far at least) when he won three world championships.

All data is taken from Mr. Kin site (Win-Loss pages), except for Lee Chango which is from SL:



Ke Jie won three more championships then Park during these years but noticing the marginal difference between them makes me wonder if The Reign of Ke and Park is not more proper...

Also, looking at the Lee Chango stats, makes me marvel again at his dominance: he played in world championship finals for about 20 years (1992-2012). For that matter, Lee Sedol played in world finals for "only" fifteen years (2001-2015).

Ke Jie made a great start toward a legendary career, and for sure in five-ten years he will accomplish much more. As for now, I don't think that is fair to put him yet in the same discussion with Lee Chango (and Cho Hun-hyun, Lee Sedol, Gu Li etc) as the best ever Go player.

The thing is, you can't just look at the winrate. You need to look at the strength of the opponents as well. I have a much higher than 80% winrate on flyordie, because I sandbag 15kyus. I'm definitely not stronger than Lee Changho.

When you look at Shin Jinseo's games in the Chinese A league, his winrate is not all that great, because the competition is fierce. While Park Junghwan does quite well in the Chinese A league, his winrate is also slightly "inflated" by the relatively weaker opponents in the korean domestic matches.

On the other hand, pretty much all of Ke Jie's opponents are top 50 in the world, at least. Top 30 even. All of whom are probably at least as strong in absolute terms as Lee Changho at his best.

And with a full day to go, the hyperbole champion of the year is settled.

_________________
Occupy Babel!

goratings.org is not perfect, but at least it's a good starting point for discussion.

Lee Changho's top historical rating there is 3559; if someone today had that rating, they would be the #3 human in the world, behind Ke Jie and Park Junghwan. Ke Jie's current rating of 3634 is the highest that it's ever been, and is appreciably higher than Le Changho's peak rating.

Ke Jie has played 88 games that goratings.org knows about in 2016; the average rating of his opponents was 3433.5. Someone with that rating would be the #31 human in the world. 24 of his 88 games were against players outside of the current rating cutoff for the #50 human in the world.

Isn't this the same position as in Chess?
Lee Changho is Garry Kasparov (longer reigning time)
While Ke Jie is Magnus Carlsen (higher peak rating point, current world no.1 player)
I bet in Chess forums also have this kind of thread 'The Reign of Magnus Carlsen'.

But aren't ELO ratings artificially inflated when there are a lot more pros and more recorded games? I raised a comparison in another thread where I found Xie He got to a higher ELO rating in 2006 than Cho Hunhyun ever had in his career, even though by that time Xie He's best result in international tournaments was a 3rd place finish, and he only made to the quarterfinals two other times. When there were only a few other worthy challengers and all you could do was consistently beating players 100-200 points lower like Cho and LCH did, you couldn't really improve your ELO much.

I'm not saying Lee Changho can dominate like he did if he were born 15 years later, but his ELO would definitely be higher even by winning fewer tournaments.

I think that's not a bad start, dfan. Like you said, goratings can't be fully relied upon.

I would point out that until Lee Changho got older and his level of play fell, only Lee Sedol (maybe Gu Li?) had a claim to be as strong as he was (and this was when Lee Changho had already been on top of the world for a decade, so I don't know if it was his peak). And Lee Sedol is still, at 33, one of the very best players. So unless you argue that Lee Sedol has been getting stronger as he enters his 30s, there's at best a few players who could match Lee Changho in his prime.

What is true is that today there is probably deeper competition than Lee Changho had.

_________________
Occupy Babel!

It is hard to say. There are both forces that could cause inflation and forces that could cause deflation, and one doesn't know a priori whether they will balance. There's a intuitive perception among chess players that Elo ratings have inflated over time (after all, today's players have higher ratings), but studies have shown that the relationship between rating and move quality (as evaluated by computer) has remained relatively constant over the last 100 years. None of this necessarily has any relation to historical goratings.org ratings, though.

Goratings uses the ELO scale, but whole history rating. I'm not sure if it's subject to inflation per se. As I understand it, the problem is that it introduces new players with a prior of 3000. So if you have a pool of players and the range of strengths increases, the top and bottom of the range will expand, even if the top player is no better. The converse happens if the weakest players get better but the strongest doesn't.

Sure enough, the pool of players included is much bigger than it was in 1990 (look at the historical ratings).

_________________
Occupy Babel!

Ke Jie had been sick the last few days and had to withdraw from Mingren quarterfinals.

I don't know about "greatest ever", but in December 2015 he played against Lee Sedol in the MLily Cup finals. It's a best of 5, and it's a great series. You can find (long) English commentary from Mungwan Kim 9p here:
https://www.youtube.com/playlist?list=P ... Dqnctwl6L_

Thanks