Magicwand wrote:
1 is 6 points gote for black. Depending on the circumstances, would it not be better to play
at 
instead, assuming the follow-up monkey jump is sente?
| Life In 19x19 http://www.lifein19x19.com/ |
|
| basic endgame theory that actually is useful II http://www.lifein19x19.com/viewtopic.php?f=11&t=3048 |
Page 1 of 1 |
| Author: | Magicwand [ Sat Jan 29, 2011 10:50 am ] |
| Post subject: | basic endgame theory that actually is useful II |
Magicwand wrote: value of a = 1 point gote = 1 value of b = 1 + 1/2 point gote = 1.5 logic: after 1 point is played they both share 1 point endgame so add 1/2 and same logic for below will yield..as below. value of c = 1 + 1/2 + 1/4 point gote = 1.75 value of d = 1 + 1/2 + 1/4 + 1/8 point gote = 1.825 value of e = 2 point gote so point e is bigger than a,b,c,d ok Bill has gave some more intresting info about my diag. thank you very much. point value i gave above is a point value of point swing. (if black plays and if white plays) although i didnt give full explination on every example..(i didnt want to insult your intelligence) it is pretty much self explanitory. now as i planned i will go to next ending game. so on third line hanne and connect value is 6 point. (bill will count that as 3 point) i will continue with this thread later when i find the time.. what i am doing here is trying to lay foundation for more complicated end game. as i said before i am writing from my head so it may contain mistakes. please feel free to let me know if you find any. |
|
| Author: | cdybeijing [ Fri Mar 04, 2011 3:15 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
Magicwand wrote: 1 is 6 points gote for black. Depending on the circumstances, would it not be better to play at instead, assuming the follow-up monkey jump is sente?
|
|
| Author: | topazg [ Fri Mar 04, 2011 3:22 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
cdybeijing wrote: Magicwand wrote: 1 is 6 points gote for black. Depending on the circumstances, would it not be better to play at instead, assuming the follow-up monkey jump is sente?I don't think so. Firstly, the monkey jump isn't sente, secondly White can only play here first in gote, so I'm not sure why Black would want to play here without maximising his points. The monkey jump would be double gote, so it would probably be better to play the shorter jump, but even with that, it's only one point better than this sequence. |
|
| Author: | Magicwand [ Fri Mar 04, 2011 3:25 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
cdybeijing wrote: Magicwand wrote: 1 is 6 points gote for black. Depending on the circumstances, would it not be better to play at instead, assuming the follow-up monkey jump is sente?no. monkey jump here is not sente here. also if white answer as below..black is not gaining any point. i guess bill can explain this better ... |
|
| Author: | Bill Spight [ Fri Mar 04, 2011 4:23 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
cdybeijing wrote: Magicwand wrote: 1 is 6 points gote for black. Depending on the circumstances, would it not be better to play at instead, assuming the follow-up monkey jump is sente?Yes.  You do not even have to assume that the monkey jump is sente. We can show this with a difference game. ( See http://senseis.xmp.net/?DifferenceGame ) By mirroring we set up a position that is even. Then in one region Black makes one of the plays to compare and in the other region White makes the other. Then by play we may be able to see if the resulting position favors one player or the other. In this case it favors neither player, because the player with the move wins the difference game. That means that sometimes one play is better, sometimes the other play is better. Edit: Here is an example. White can win by 1 pt. with the descent at "a". The hane at "b" loses.
|
|
| Author: | topazg [ Fri Mar 04, 2011 4:32 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
@TMark, John, or anyone with lots of pro games and database-fu: Is there any instances where this approximate position has occurred in pro games (i.e. enough stones locally that the hane is gote for both plays, and the descent + monkey jump is gote twice) where the pro has descended instead of played the hane? |
|
| Author: | daal [ Fri Mar 04, 2011 4:49 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
Bill Spight wrote: cdybeijing wrote: Magicwand wrote: 1 is 6 points gote for black. Depending on the circumstances, would it not be better to play at instead, assuming the follow-up monkey jump is sente?Yes. It seems to me that your difference game proves exactly the opposite. Playing at loses in every instance, no? (Or do you mean "yes" as in "yes, it would not be better?")Edit: nevermind, I was misreading the minus signs. 
|
|
| Author: | Bill Spight [ Fri Mar 04, 2011 5:07 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
I have edited my earlier post to add an example.
|
|
| Author: | topazg [ Fri Mar 04, 2011 6:31 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
Bill Spight wrote: I have edited my earlier post to add an example. Many thanks Bill, very interesting. What about: Is the descent ever better than the hane here? |
|
| Author: | mw42 [ Fri Mar 04, 2011 8:29 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
topazg wrote: Is the descent ever better than the hane here? Hmm, no doubt Bill will come to correct me, but the situation in his example is such that white is leading by 1. With the top and bottom areas being equal, they are miai, so to speak. Therefore, white's best play is to always claim the miai. So, black's hane would be countered by white's hane. |
|
| Author: | Bill Spight [ Fri Mar 04, 2011 8:58 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
topazg wrote: Bill Spight wrote: I have edited my earlier post to add an example. Many thanks Bill, very interesting. What about: Is the descent ever better than the hane here? Both players can pass and count the score.  Edit: Sorry. By Japanese '89 rules the board as is is a whole board seki.  White needs to settle the boundaries.
|
|
| Author: | topazg [ Fri Mar 04, 2011 9:34 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
mw42 wrote: topazg wrote: Is the descent ever better than the hane here? Hmm, no doubt Bill will come to correct me, but the situation in his example is such that white is leading by 1. With the top and bottom areas being equal, they are miai, so to speak. Therefore, white's best play is to always claim the miai. So, black's hane would be countered by white's hane. I agree. I suppose what I'm aiming it is whether or not in a practical game example the descent would be a better choice, under "standard" play up to that point. In the example Bill gave, Black has a very odd shape at the top, and it gives the impression that Black has just descended in the same fashion. If I was Black, I'd have hane'd, so White could do on the other side, we'd end up in exactly the same position. It may be that this is an example that could crop up frequently in normal games, it may be that the position is artificially contrived due to Black having played an unusual previous move. I'm not trying to be critical, your problem examples are always fab Bill, I'm just trying to fathom whether in "basic endgame" situations, what probability there is that the descent is the right move (hence asking for professional games with this situation in - there are plenty with the hane, but I haven't seen the descent before in one). |
|
| Author: | Bill Spight [ Fri Mar 04, 2011 11:07 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
topazg wrote: I suppose what I'm aiming it is whether or not in a practical game example the descent would be a better choice, under "standard" play up to that point. In the example Bill gave, Black has a very odd shape at the top, and it gives the impression that Black has just descended in the same fashion. If I was Black, I'd have hane'd, so White could do on the other side, we'd end up in exactly the same position. It may be that this is an example that could crop up frequently in normal games, it may be that the position is artificially contrived due to Black having played an unusual previous move. I'm not trying to be critical, your problem examples are always fab Bill, I'm just trying to fathom whether in "basic endgame" situations, what probability there is that the descent is the right move (hence asking for professional games with this situation in - there are plenty with the hane, but I haven't seen the descent before in one). Making realistic problems of this sort when the temperature is this high is a challenge. Anyway, here is another. |
|
| Author: | Bill Spight [ Fri Mar 04, 2011 12:47 pm ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
OK, to talk about this kind of play in general, let me start with losing sente. A losing sente is quite simply a move that is made with sente that takes a loss (absolutely or on average). OC, normally a losing sente is a mistake. However, it is well known -- or should be --, that a losing sente can be a good sacrifice, made to prevent a gain by the opponent when she gets the move.Now, the position in question is classified as a gote where the play gains 3 points. In that context, the descent is normally played with sente because it raises the local temperature, since it threatens the monkey jump. (As cdybeijing indicated. ) But, as Magicwand indicated, it is not a losing sente, since the response returns us to the same count that we started with. There is no sacrifice at all (on average). Once we have gotten this far, we can see that it is a possible play. In the first example I showed the descent was not played with sente. The second example is more typical, where there is another gote play that gains 3 points. In such a case, playing the descent normally gets the last 3 point play (i. e., one that gains 3 pts.). That is so when there are an odd number of 3 point gote elsewhere on the board. Then playing the descent typically dominates playing the hane and connect. (Because zero is an even number, I expect that the descent dominates a significant percentage of the time, but less than half the time.) Why, then, do the pros not seem to play the descent so often? There are a couple of reasons, I think. First, they learn that this is a 6 pt. gote. (It gains 3 pts. for White or 3 pts. for Black. ) Second, they correctly regard play values as mere guides, and use them to assist their reading, not to determine their play. So they will start out regarding it as a gote, and (I guess) only try the descent if the hane does not work out. Now, even though the descent often dominates, my educated guess is that it will make a difference between winning and losing once in 400 or 500 games. Since it matters so seldom, I think that the pros choose the prominent play the vast majority of the time.OK, here's the twist. Suppose that there are an even number of 3 pt. moves elsewhere on the board and White should therefore play the hane in order to get the last one. Black may be able to prevent that beforehand by playing the descent herself. (Remember, it raises the local temperature. Black does not have to wait until the temperature drops to 3 to play the descent.) Now, if it is foreseeable that one player, say, White, will get the move at temperature 3, then normally either White should play the descent in order to get the last 3 pt. play, or Black should play the descent beforehand to prevent that. I think that that is normally foreseeable, and that most of the time in a real game the descent is the normal play.  (And if it is not foreseeable, if you go ahead and play the descent before the temperature drops to 3, all you risk losing, as a rule, is a potential ko threat or the chance that your opponent will make a mistake.)That will be the case in a gote position when each player has a non-losing sente. (OC, not all third line hane and connect positions meet that criterion.) |
|
| Author: | cdybeijing [ Sat Mar 05, 2011 7:01 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
You guys should clearly listen to me more. I'm an expert at finding gote plays
|
|
| Author: | topazg [ Sat Mar 05, 2011 8:02 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
Bill, you rock. Seriously, keep posting this stuff! |
|
| Author: | mw42 [ Sat Mar 05, 2011 8:15 am ] |
| Post subject: | Re: basic endgame theory that actually is useful II |
To topazg, with love: So now you can't say you haven't seen a pro game with the descent.
|
|
| Page 1 of 1 | All times are UTC - 8 hours [ DST ] |
| Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |
|