Life In 19x19

You can do bouldering outside too, but it's indeed always close to the floor :


(the person behind is here to redirect you to the mat when you fall)

It's often more technical than "traditionnal" rock climbing. And far much shorter.

I'd say that, in go term, bouldering is akin to solving a tsumego

sorin wrote:

ez4u wrote:He refused to give his guess (he was grinning when he refused, clearly realizing that I was trying to pin him down. )

Thanks for trying!

I asked this question on Cho Hye-Yeon 9p forum on Facebook, and she replied unequivocally: 1 point is worth more then 100 minutes of thinking time for her, and she would make this trade-off at the beginning of the game, if given the chance:

sorin wrote:

sorin wrote:

ez4u wrote:He refused to give his guess (he was grinning when he refused, clearly realizing that I was trying to pin him down. )

Thanks for trying!

I asked this question on Cho Hye-Yeon 9p forum on Facebook, and she replied unequivocally: 1 point is worth more then 100 minutes of thinking time for her, and she would make this trade-off at the beginning of the game, if given the chance:

So she would pay at least 100 min. time for 1 pt. Would she pay 1 pt. for 100 min.? My feeling is that the more time you have, the less per minute it is worth.

Bill Spight wrote:So she would pay at least 100 min. time for 1 pt. Would she pay 1 pt. for 100 min.? My feeling is that the more time you have, the less per minute it is worth.

I read her "1 point means everything to me" as "will not give away 1 point, no matter how much is the increase in thinking time".

(Of course, this is about a possible trade at the beginning of the game.)

I read her "1 point means everything to me" as "will not give away 1 point, no matter how much is the increase in thinking time".

Even if you read it correctly, it may not be right in the first place - self-delusions can creep in.

I say this on the basis of a long career in journalism where most news journalists, already under deadline pressure, invariably chose to leave writing the story until the very last minutes. We felt that this way we worked better. Quite possibly we did - one rational explanation is that the longer you delay, the more of a moving story you can incorporate. The intense focus you need also enables you to omit extraneous details, which is almost always good for a news story. You type faster. Above all, you feel r e a l l y good at the end of it. I have no recollection of anyone ever missing a deadline. You had to have your rush.

On occasions we'd be called on to write a feature or opinion piece, with much more time and no special deadline. That was purgatory. We'd be listless, and that would manifest itself by pecking away at the typewriter instead of pounding it. And many missed deadlines would occur.

I read years later that what is going is that, with a looming deadline, your reaction to it causes a pleasurable spike in noradrenaline in the brain, to which you can become mildly addicted. We all know of thrill seekers. The big difference in front-line journalism is that we have behind us an army of editors and sub-editors who catch our mistakes.

But mistakes do occur. One I still cringe at (but only mildly) is when, writing at very high speed a breaking story at a trade conference in Uruguay, I realised I couldn't remember the first name of the Uruguayan trade minister. I shouted out to the newsroom, "What's Iglesias's first name?" and got the instant reply "Julio" and that's what went into my story - instead of Enrique. I still finished on the usual high, which incidentally doesn't feel like a sugar rush - it's more like an icy calm at the time, and the pleasure comes afterwards. By which time your mistakes are in the safely forgotten past. The reason I remember that incident particularly is not because I made a mistake - I'm sure I made many and never bothered noticing. It was because on that one occasion the night sub-editor didn't catch my mistake and I was serenaded on my return. (For the teeny-boppers, Julio Iglesias was a big singing star of the time.)

I now know that chess blitz players experience a similar noradrenaline high, so it seems likely that go players do, too. They of course do not have anyone in reserve to catch their mistakes, but I would speculate that that makes their rush even more intense and thus even more pleasurable, closer to what the extreme-sports dare-devils must feel.

When you get used to that feeling, as blitz players must because of the sheer number of games they play (and you can play more in any given time), I can easily imagine, from my own experience, how wearisome extra time can feel.

I am therefore positing that go players who prefer to rush than dawdle may not be truly aware of either their mistakes or the size of them. And I further posit that they do make such mistakes.

Thanks Sorin for pinging Cho Hye-yeon - the more data points the better.

It is hard to design any kind of experimental test with pros, in an actual setting that matters anyway. I wonder if the bots can help:
- Have a strong AI play itself thousands of times, with Bot 1 getting time = X, and Bot 2 getting time = 2X. (or if clock time is the wrong way to think about this, you can halve the playouts allowed for Bot 1, or whatever the proper lingo is)
- I assume Bot 2 will win more since it can "think" more. We can then manipulate the komi to favor the less thinking bot until each is winning 50% of the time.

This would give an mapping of time to points. I realize it isn't directly translatable to humans, but we'd at least know something about the potential magnitude that we didn't know before.

Who wants to call DeepMind?

The annoying problem with most of these kinds of experiments is that all of the Zero-style bots assume a particular komi value and are not capable of understanding other ones, so you can't make a good points-vs-time comparison.

Even if that were not the case, I think that the graph of strength as a function of thinking time could look very different for humans and bots.

Over on the Leela Zero issues page, people have done experiments to see how much thinking time affects strength for various generations of the network; unfortunately, I wasn't able to find them at the moment after a few minutes of looking. Time is definitely very important to them! On the other hand, they are trained expecting to have time to search, they're not capable (yet) of allocating their main time between moves intelligently, etc.

sorin wrote:I asked this question on Cho Hye-Yeon 9p forum on Facebook, and she replied unequivocally: 1 point is worth more then 100 minutes of thinking time for her, and she would make this trade-off at the beginning of the game, if given the chance:

Adding a link to the question/answer: https://www.facebook.com/groups/7181373 ... 981380012/

Kirby: who were the Korean players in the two games you chose? Were they of similar strength to Takao and Iyama, or stronger/weaker?

@hyperpape: the games chosen by Kirby are the following.





so one of the Korean players is stronger than the two Japanese. The other Korean player only has 1 game on goratings so it's hard to know his strength.

I was transcribing a game last night which was out of the ordinary in one respect: it had an awful lot of time sujis.

So here were players willing to give up ko threats and other aji just to have more time to think - on multiple occasions.

To my eye the game itself was not complex tactically. It was strategically complex in that White chose to sacrifice a huge group of about 30 stones, but it was not a simple trade. He clearly had to evaluate several other local positions in combination to assess whether he got sufficient compensation for his huge sacrifice. And I thought it was significant, too, that once he made that assessment the game ended by the rather large margin of W+7.5. We would not normally class that as a close game, but in terms of earlier decisions it was obviously a close-run thing.

There is room for debate about both the assessment itself and the final interpretation of "closeness", but it seems pretty clear that both players felt the need for more time and were willing to give something up for it.

Would Lee Sedol 9p or especially Rui Naiwei 9p have a different opinion? I heard Lee Sedol wanted more time against Alpha-Go.

Shorter limits seem to be the best for ascertaining skill between players; it allows for more games and may give bigger differences in winrate for small differences in strength. Granted, it is of course not a foolproof method as longer games emphasise different skillsets to shorter games. More on that later.

My confusion is that one would expect at the very least for the professionals that have proven their strength twice to get to play in many matches with long time limits.

Another point is that while shorter limits shower better how a go game feels to the casual mind-sports-are-boring-arien, it becomes more difficult for commentators to tell which moves might actually work or not and speak to us their conclusion the closer a game gets to mach one. For the pros playing, it 's like hunting for treasure with a storm on the way; 'X marks the spot' becomes 'sonic X marks the spot'. For all levels, longer time limits do a better job of teaching, I think.

At the end of the yose, it's longer time limits that offer the most connection between pros and amateurs. Trying one's very best with the skill one has.

Is there a correlation for Yi Ch'ang-ho between playing times and money? I hope he has still a sensible income from playing.

I'd hope Lee Changho invested a fair chunk of his substantial winnings from his glittering career and would get a tidy income from that.

bayu wrote:Is there a correlation for Yi Ch'ang-ho between playing times and money? I hope he has still a sensible income from playing.

I don't think Lee Changho can make a living income from playing, maybe only top few players can do that. The only way to make an income from playing is to win titles, or to at least be in top 4 or so.
I am pretty sure though he can make a good living income from other sources, mostly from teaching and from books, since he is famous and very well respected for his past achievements.