It is currently Sat Apr 17, 2021 12:57 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 192 posts ]  Go to page Previous  1 ... 6, 7, 8, 9, 10
Author Message
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #181 Posted: Tue Oct 27, 2020 10:16 am 
Lives in gote

Posts: 341
Liked others: 4
Was liked: 19
Rank: 1er dan
RobertJasiek wrote:
Read variations as far as necessary, then for each major branch possibly apply theory.


Yes Robert OC. But taking aside professionals or strong amateurs who is able to find in a real game the miai value of G1? No doubt you are able to find yourself this miai value but what about not strong amateurs?

In the other hand, because I assume all G2, G3 ... very simple gote, a clever weak go player can solve the problem without any error by reading the simple branch of the yose and by asking for the value of the komi. Why not using all technics available?

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #182 Posted: Wed Oct 28, 2020 1:07 am 
Honinbo

Posts: 10674
Liked others: 3594
Was liked: 3339
Gérard TAILLE wrote:
RobertJasiek wrote:
Read variations as far as necessary, then for each major branch possibly apply theory.


Yes Robert OC. But taking aside professionals or strong amateurs who is able to find in a real game the miai value of G1? No doubt you are able to find yourself this miai value but what about not strong amateurs?

In the other hand, because I assume all G2, G3 ... very simple gote, a clever weak go player can solve the problem without any error by reading the simple branch of the yose and by asking for the value of the komi.


That's your assertion. I have to say it sounds like wishful thinking in actual games. Simple gote are easy to construct, but until near the end of the game are not so easy to find in actual play. Your example a few posts back is a case in point.

Quote:
Why not using all technics available?


And that sounds like a misunderstanding of how strong players use the heuristics of move evaluation. Do you think they fall back on heuristics when they do not have to?

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #183 Posted: Wed Oct 28, 2020 4:15 am 
Lives in gote

Posts: 341
Liked others: 4
Was liked: 19
Rank: 1er dan
Bill Spight wrote:
That's your assertion. I have to say it sounds like wishful thinking in actual games. Simple gote are easy to construct, but until near the end of the game are not so easy to find in actual play. Your example a few posts back is a case in point.


Yes Bill, actual games are not so simple.
If I manage to give a reliable miai value to each area fine I will use them to choose my move and that way I will obviously avoid many mistakes.
If I do not hesitate between two areas say G1 and G2 then OC no problem.
But what can I do if I hesitate between two areas G1 and G2 and I am not able to give enough accuracy to G1 and G2 miai values to compare them?
How can the answer be "Read variations as far as necessary" help me?

OK let's take an example where one player tries a common furikawari strategy.

Click Here To Show Diagram Code
[go]$$B White to play
$$ -----------------
$$ | . . . B B O . |
$$ | . . . B W . . |
$$ | . . . B W . X |
$$ | . . . B W X . |
$$ | . . . B W . . |
$$ | . . . B W . . |
$$ | . . X W W . . |
$$ -----------------[/go]


Well apply the theory and tell me what will be the white move.
OC assume all marked stones invulnerable to assure independancy.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #184 Posted: Wed Oct 28, 2020 8:19 am 
Honinbo

Posts: 10674
Liked others: 3594
Was liked: 3339
Gérard TAILLE wrote:
But what can I do if I hesitate between two areas G1 and G2 and I am not able to give enough accuracy to G1 and G2 miai values to compare them?
How can the answer be "Read variations as far as necessary" help me?

OK let's take an example where one player tries a common furikawari strategy.

Click Here To Show Diagram Code
[go]$$B White to play
$$ -----------------
$$ | . . . B B O . |
$$ | . . . B W . . |
$$ | . . . B W . X |
$$ | . . . B W X . |
$$ | . . . B W . . |
$$ | . . . B W . . |
$$ | . . X W W . . |
$$ -----------------[/go]


Well apply the theory and tell me what will be the white move.
OC assume all marked stones invulnerable to assure independancy.


OK, so we're assuming a larger board on which the marked stones are invulnerable, you just set this example up on the small board for convenience. :)

Gérard, when you say, "Apply the theory", I guess you mean calculate the full thermographs of these positions or at least the miai values of the original plays. Why is that the thing to do?

I think I have already said this in our recent discussions, but you use the right tool for the job. I think we agree on that. :)

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #185 Posted: Wed Oct 28, 2020 10:23 am 
Lives in gote

Posts: 341
Liked others: 4
Was liked: 19
Rank: 1er dan
Bill Spight wrote:
Gérard, when you say, "Apply the theory", I guess you mean calculate the full thermographs of these positions or at least the miai values of the original plays. Why is that the thing to do?

I think I have already said this in our recent discussions, but you use the right tool for the job. I think we agree on that. :)


Sorry Bill, basically I reacted against the "Read variations as far as necessary" of Robert because I see it could too difficult for middle level players as me.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #186 Posted: Wed Oct 28, 2020 12:33 pm 
Oza

Posts: 2874
Liked others: 16
Was liked: 3971
Quote:
I reacted against the "Read variations as far as necessary" of Robert because I see it could too difficult for middle level players as me.


It's not just middle-level players. It's probably all social players, which is what nearly all of us are. And if you were prepared to put that amount of effort into reading you would gain far, far more from reading out tesujis and tsumegos. Why eat lettuce for lunch when you can have steak?

But if you really do want endgame lettuce in your diet, there is a compromise which some people (even pros) use. You can memorise the value of lots of standard endgame positions. The average Japanese or Chinese person bright enough to play go well learns at least 3000 kanji, which can have up to 50 strokes. Treating moves as strokes and numerical value as meaning, you can easily learn 1,000 endgame positions - as some people do. 10 a day for about three months. Voilà!

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #187 Posted: Wed Oct 28, 2020 12:59 pm 
Tengen

Posts: 5286
Liked others: 0
Was liked: 726
Gerard, please explain your shorthand (once more)! There can sometimes be shorthands, e.g., if the maximum value of an endgame region is too small so playing in another region must be more valuable. - Have you read my proofs now and formed an opinion?

Learning 100,000 endgame values by heart is possible in principle presuming correct available values. Such, and application of all theory, do not make tactical reading in (early) endgame superfluous, like in life and death.

John, what is a social player - do non-social players exist!?! - In the endgame, tesuji and tsumego are the opposite of steak because almost all endgame moves must be distinguished by values or tactical reading.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #188 Posted: Wed Oct 28, 2020 1:52 pm 
Lives in gote

Posts: 341
Liked others: 4
Was liked: 19
Rank: 1er dan
RobertJasiek wrote:
Gerard, please explain your shorthand (once more)! There can sometimes be shorthands, e.g., if the maximum value of an endgame region is too small so playing in another region must be more valuable. - Have you read my proofs now and formed an opinion?


Click Here To Show Diagram Code
[go]$$B White to play
$$ -----------------
$$ | . . . B . B B O . |
$$ | . . . B B B W . . |
$$ | . . . B . B W . X |
$$ | . . . B B B W X . |
$$ | . . . B W W W . . |
$$ | . . . B W W W . . |
$$ | . . . B W . W . . |
$$ | . . . B W W W . . |
$$ | . . X W W . W . . |
$$ -----------------[/go]


Robert, the problem is that I do not know what do you suggest in order to find what white has to play in the example above.
It suspect I misunderstood completly your process to find here the white move and I do not want to disagree on something I am not sure.
Please explain your process on this example.
Be sure I will forget my proposal if your process is efficient in real play. So let's begin by understanding your approach.

Concerning double sente and your proof I have no more problem:
With you defintion "double sente" do not exist. With my defintion double sente exist but they are only a particular case of gote without any interest for theory. At least it is clear.
BTW I tried to undestand your proof but, because I basically agree with you if you take your defintion, and because your notation is quite hard to masterise, I gave up. Anyway I am pretty sure it is correct in your context.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #189 Posted: Wed Oct 28, 2020 2:44 pm 
Tengen

Posts: 5286
Liked others: 0
Was liked: 726
I have meant your discussion with Bill, where you have suggested that a fast approximation instead of CGT perfect analysis can be possible if all one wants is to achieve winning. Can you please explain your method without forcing me to solve a complex example puzzle, for which I lack time?

As to my read then possibly apply theory suggestion, it is a three stage process:

1) read and develop tactical variations
2) at the reading end of each variation, assess that position by assigning the obvious values or applying theory to calculate them
3) combine to determine the initial position's values

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #190 Posted: Wed Oct 28, 2020 2:53 pm 
Oza

Posts: 2874
Liked others: 16
Was liked: 3971
Quote:
John, what is a social player - do non-social players exist!?!


It's well known that it's risky to make jokes in a foreign language. It is also fraught to try to make nice distinctions. I assume you are mixing up with something like sociable. Social, of activities, refers to something you do communally for pleasure as opposed to work. Expectations are lower. You are not expected to be an expert or an obsessive. Indeed, to be so would be considered non-social, and maybe even anti-social in certain circumstances.

Quote:
- In the endgame, tesuji and tsumego are the opposite of steak because almost all endgame moves must be distinguished by values or tactical reading.


My purview is much bigger than the endgame. If (as a social go player) you can only spend an hour a week on go, you will normally gain most if you spend 55 minutes at least on the opening and middle game.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #191 Posted: Wed Oct 28, 2020 3:09 pm 
Tengen

Posts: 5286
Liked others: 0
Was liked: 726
Opening, middle game and endgame all are difficult. A player can improve by studying some of these. However, a player neglecting one will rather soon stop improving.

Endgame is very important because very many moves have endgame aspects. Professionals consistently tell amateurs to study their weakest aspects: life and death reading, endgame. Amateurs in general - not only the experts. Many excuses can be invented not to study enough or seriously, and creating fear of expert-like work is such an excuse.

Top
 Profile  
 
Offline
 Post subject: Re: How evaluate double sente moves ?
Post #192 Posted: Wed Oct 28, 2020 3:14 pm 
Lives in gote

Posts: 341
Liked others: 4
Was liked: 19
Rank: 1er dan
RobertJasiek wrote:
I have meant your discussion with Bill, where you have suggested that a fast approximation instead of CGT perfect analysis can be possible if all one wants is to achieve winning. Can you please explain your method without forcing me to solve a complex example puzzle, for which I lack time?

As to my read then possibly apply theory suggestion, it is a three stage process:

1) read and develop tactical variations
2) at the reading end of each variation, assess that position by assigning the obvious values or applying theory to calculate them
3) combine to determine the initial position's values



Click Here To Show Diagram Code
[go]$$B White to play
$$ -----------------
$$ | . . . B . B B O . |
$$ | . . . B B B W . a |
$$ | . . . B . B W . X |
$$ | . . . B B B W X . |
$$ | . . . B W W W . . |
$$ | . . . B W W W . . |
$$ | . . . B W . W . . |
$$ | . . b B W W W . . |
$$ | . . X W W . W . . |
$$ -----------------[/go]


My point is that this position is not at all a complex example puzzle. This position is unbelievable simple.

You imagine :w1: at "a" and :b2: at "b" and you count the score of the game which is here obvious even for a weak player.
if this sequence give the win to white then white plays "a" otherwise white plays "b". I guess that a lot of players use such method and I am surprised to see it is not part of the theory your are defending but I explained already why probably it happenned.

What you have to understand is that this method is not general but only specific for positions with one or maybe two complex areas and all other very simple. It is simply an adding tool which cannot replace the more general approach with miai values. In these restraint contexts where it can be applied this method is very efficient and give a correct move without error (no heuristic). Note this method can be quite easliy applied by players trained to count the game in order to know who is winning.

I have no problem at all if you say it could not be an interesting method because it is not a genaral one ... because it is true!

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 192 posts ]  Go to page Previous  1 ... 6, 7, 8, 9, 10

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: dust and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group