Power: just a few notes

[RobertJasiek]

The Power part of the 2-in-1 book Power / Brilliance has a few minor aspects I'd like to mention:

1) P. 79 says: "In general, when a fight on the side occurs, it is usually better to run away than to try to live on the side." P. 81: "[...] achieves easy access to the centre, usually a sign of success in a side fight [...]". P. 83: "[...] Karigane is ignoring usual go theory about the preference for running away [...]" So Fairbairn tries hard to convey the thought that in a fight, usually, running to the center conforms to go theory. The reader, however, is left alone wondering what that go theory is. Giving related advice in First Fundamentals, I can cite the relevant principles to explain that go theory: a) Avoid premature endgame. b) Choose the bigger space. / Move to the wider direction. c) Increase your eyespace or run out. [When the other major life-making options of partitioning eyespace or connecting to another live group are not available.] Among Black's (c) options, running out meets (a) and especially (b).

2) P. 112 says: "In a one-move approach ko, where four moves are needed in total [to win and dissolve the ko], the value [per move] drops to one quarter." Now I recall a paper by Bill Spight, which I can't find at the moment, explaining a relation between approach kos and fibonacci numbers. This lets me wonder: is 1/4 right or would it be 1/5? Why?

3) P. 59 says: "[...] with adjournaments but no sealed moves". P. 127 says: "[...] Shusai [...] had the option of spinning out a day's play in a difficult position with the right to adjourn. These were the days before sealed moves." Only one of the two sealing move statements can be correct; the other must by a typo. Which?

4) Mochikomi is so often stressed that really every reader will understand its relevance:)

[HermanHiddema]

3) P. 59 says: "[...] with adjournaments but no sealed moves". P. 127 says: "[...] Shusai [...] had the option of spinning out a day's play in a difficult position with the right to adjourn. These were the days before sealed moves." Only one of the two sealing move statements can be correct; the other must by a typo. Which?

Both pages claim there are adjournments, but no sealed moves. I don't see the contradiction?

[RobertJasiek]

In "These were the days before sealed moves.", "days", by context, refers to those "days" in the preceding sentence. To remove such ambiguity, it could have said "That was the age...".

[HermanHiddema]

Oh, I see, you think that "before" in this context can mean that the day (of play) preceded a sealed move, i.e. the day's play was ended with a sealed move. That is not a meaning that any native English speaker would ever find in the sentence, hence it would never occur to John that the sentence is in any way ambiguous. And it isn't, to anyone with sufficient fluency in English, but I can see how it could confuse foreign readers.

[Bill Spight]

The Power part of the 2-in-1 book Power / Brilliance has a few minor aspects I'd like to mention:

1) P. 79 says: "In general, when a fight on the side occurs, it is usually better to run away than to try to live on the side." P. 81: "[...] achieves easy access to the centre, usually a sign of success in a side fight [...]". P. 83: "[...] Karigane is ignoring usual go theory about the preference for running away [...]" So Fairbairn tries hard to convey the thought that in a fight, usually, running to the center conforms to go theory. The reader, however, is left alone wondering what that go theory is.

Jowa advised: Don't run. Later Korschelt gave the rule of thumb: One eye and access to the center, which he presumably learned from Shuho. And, of course, there is the general advice not to get fenced in. Obviously, none of these are universal truths. You cannot build a consistent go theory from them without qualification. However, human beings are quite good at reasoning with unspecified exceptions, so, as pieces of advice, these are still good.

2) P. 112 says: "In a one-move approach ko, where four moves are needed in total [to win and dissolve the ko], the value [per move] drops to one quarter." Now I recall a paper by Bill Spight, which I can't find at the moment, explaining a relation between approach kos and fibonacci numbers. This lets me wonder: is 1/4 right or would it be 1/5? Why?

The value of a ko, particularly the value of approach kos and 10,000 year kos, etc., depends upon the ko threat situation. I call the ko along with the threat situation the ko ensemble. Berlekamp came up with the idea of komaster, and showed how it could be used to derive the value of what he called hyperactive kos, such as approach kos and 10,000 year kos.

In the approach ko that JF refers to, suppose that in one play Black can win the ko, while White, if allowed to, could take the ko, convert it to a direct ko, and then win the ko. If White is komaster, then the average value per move is, as the book says, 1/4 the difference between a Black win and a White win. However, if Black is komaster, the average value is 0.

Now, often neither player is komaster. Berlekamp later came up with the idea of a Neutral Threat Environment (NTE) where each player has equal and opposite ko threats to the other player. Berlekamp, Bill Fraser, and I explored such environments. A Neutral Threat Environment has two flaws. First, the exact structure of the threats matters. Second, it is not very much like typical ko threat environments. Still, it is better than nothing. We finally opted for an NTE with the simplest form of ko threats. Then I discovered that, for the simplest family of approach kos, the value of the move in such an NTE is the difference between winning and losing the ko times one over a Fibonacci number.

For a direct ko (a zero move approach ko) the ratio is 1/3. For a one-move approach ko it is 1/5. For a two-move approach ko it is 1/8. For a three-move approach ko it is 1/13. Etc.

There is a saying that a three-move approach ko is no ko. Think about how difficult it would be for the player who takes the ko to be komaster. According to the rule used by the author, the value of a ko move would be 1/6 the difference between winning and losing. In most cases that would be way off. The saying is surely closer to the truth. I expect that the NTE ratio of 1/13 is probably too large, as well. But who knows? There is still room for research here.

[RobertJasiek]

Bill, can you please give (simplifying, where necessary) examples of a one move approach ko im case of a) Black komaster, b) White komaster, c) NTE to demonstrate the values 1/4, 0, 1/5?

There is a saying that a three-move approach ko is no ko.

I won a 13- or 14-move approach ko (whose semeai group covered the center and entered the sides) by means of many local ko threats. Less extreme many-step kos are not scarce in my games. So... theory must be enhanced!:)

[Bill Spight]

Bill, can you please give (simplifying, where necessary) examples of a one move approach ko im case of a) Black komaster, b) White komaster, c) NTE to demonstrate the values 1/4, 0, 1/5?

There is a saying that a three-move approach ko is no ko.

I won a 13- or 14-move approach ko (whose semeai group covered the center and entered the sides) by means of many local ko threats. Less extreme many-step kos are not scarce in my games. So... theory must be enhanced!:)

You had to take the ko 13 or 14 times, then ignore the opponent's threat (if he made one) to make an approach move, until finally you made a direct ko, which you won?

As for approach ko examples, there are some on SL, I think. Let me check.

[RobertJasiek]

You had to take the ko 13 or 14 times, then ignore the opponent's threat (if he made one) to make an approach move, until finally you made a direct ko, which you won?

Something like this, yes. My opponent's problem: he did not have noteworthy threats because the entire game hinged about the dragon's life. Only local threats counted. The dragon was pretty well connected and had enough liberties, but the surrounding stones had lots of ko date aji with little endgame loss.

[RobertJasiek]

P. 52f says: "plays Shuho on five stones [...] Becomes 1-dan" / "beats Shuei thrice on three stones and so is made 4-dan"

I conclude that Meiji ranks had nothing to do with today's pro or amateur ranks. Taking handicap 5 against a today's 9p means being about European 3d. Taking handicap 2 (because 3 is too much) means about (strong) European 6d. Now, does this mean that, in those times, 1p and 4p were today's amateur dans and only 5p+ would be today's professionals? Or does it mean that Meiji dan promotion was extraordinarily and unreasonably slow? I would not buy the third principle possibility of the top Meiji players being today's 12p+.

***

P. 54 refers to Shikatsu Myoki (tsumego) and Igo Shinzui ("unusually high-level set of commentaries"). Is there more information on this literature? What characteristics qualified for "unusually high-level"?

[hyperpape]

I conclude that Meiji ranks had nothing to do with today's pro or amateur ranks. Taking handicap 5 against a today's 9p means being about European 3d. Taking handicap 2 (because 3 is too much) means about (strong) European 6d. Now, does this mean that, in those times, 1p and 4p were today's amateur dans and only 5p+ would be today's professionals? Or does it mean that Meiji dan promotion was extraordinarily and unreasonably slow? I would not buy the third principle possibility of the top Meiji players being today's 12p+.

If "(strong) European 6d means "Ilja Shikshin, and only Ilja", then this may make sense, but is there really evidence that any other European amateurs are too strong to take a 3 stone handicap from a 9p? This means going by current standards for promotion to 9p. Some of the older professionals reached 9p during a time when that was not as difficult as it is today.

See for instance: http://www.361points.com/blog/tag/shusaku-cup/.

[RobertJasiek]

If "(strong) European 6d means "Ilja Shikshin, and only Ilja",

No. Ilya is 7d.

then this may make sense, but is there really evidence that any other European amateurs are too strong to take a 3 stone handicap from a 9p?

I win significantly more than I lose when taking H3.

Forget the myth that 1p == 1/3 handicap stone. Nowadays it is more like 1/4.

[HermanHiddema]

If "(strong) European 6d means "Ilja Shikshin, and only Ilja",

No. Ilya is 7d.

then this may make sense, but is there really evidence that any other European amateurs are too strong to take a 3 stone handicap from a 9p?

I win significantly more than I lose when taking H3.

Against top pros at their peak strength?

Forget the myth that 1p == 1/3 handicap stone. Nowadays it is more like 1/4.

Things change. It was also once 1/2 stone per pro rank. It will be different again the future.

Forget the myth that European 7d equals 1p.

[topazg]

I win significantly more than I lose when taking H3.

Can we have opponents and games for this please?

[RobertJasiek]

Can we have opponents and games for this please?

Games: sorry, did not record (except for the H1 game, which I have on paper only).

Results (as far as I remember, only H3 games):
- Saijo Masataka 8p/9p, I win about 4:2. (Not counting many lightning games.)
- Korean pro, I do not recall rank: I win 1:0.
- Simultaneous Kobayashi Satoru 9p: I win 1:0.

Huge life+death study:
- Saijo Masataka 8p and another 8p discussing against me and trying to kill my dragon: I win 1:0.

H1 game:
- Dozens of pros running around against 100+ amateurs: I win 1:0.

Note: Before I was strong enough to play H3 games, i.e., in the 1990s, my results with higher handicap were much worse because then I did not know well yet how to use the fixed handicap stone patterns.

[RobertJasiek]

Forget the myth that European 7d equals 1p.

I have long abandoned it. It is more like Japanese 5p/6p.