Gote move vs sente move in yose

[RobertJasiek]

How do you continue the analysis to find what is the best first white move?

Suppose the reverse sente count R, and we do have a gote option and a sente option indeed with their resulting counts A and B, respectively, where, since White is the creator, A ≤ B < R.

Calculate / determine the
- gote move value of the gote option,
- follow-up move value of the sente option,
- possibly sente move value of the sente option,
- possibly gote-sente-difference B - A.
- temperature T, and
- possibly second-largest environmental move value T1.

Third, apply the theory in [22], [23] of Bill Spight and me to decide whether to play elsewhere, locally with the gote option or locally with the sente option.

https://www.lifein19x19.com/viewtopic.p ... 45#p143245

[Gérard TAILLE]

How do you continue the analysis to find what is the best first white move?

Suppose the reverse sente count R, and we do have a gote option and a sente option indeed with their resulting counts A and B, respectively, where, since White is the creator, A ≤ B < R.

Calculate / determine the
- gote move value of the gote option,
- follow-up move value of the sente option,
- possibly sente move value of the sente option,
- possibly gote-sente-difference B - A.
- temperature T, and
- possibly second-largest environmental move value T1.

Third, apply the theory in [22], [23] of Bill Spight and me to decide whether to play elsewhere, locally with the gote option or locally with the sente option.

https://www.lifein19x19.com/viewtopic.p ... 45#p143245

Robert, you know perfectly that on many occasions in the past I tried to use your theory or Bill's theory but unless the position is very simple I always failed to give the correct answer for two reasons : firstly the method is too complicated and needs too much time to be used in a practical game (see what you have listed yourself in your post under "Calculate / determine the") and secondly the local position is rarely clearly defined for applying a pure theoritical method and you will be always able to critisize (and you will be right!) any assumption for simplifying the analysis.
As an example:

$$W White to play $$ ---------------------------- $$ . . . . . . . 1 . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$W White to play $$ ---------------------------- $$ . . . . . . . 1 . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

If I claim white is sente I am quite sure you will contredict me by saying it will depend on the exact configuration of black on the upper side and of course you are right aren't you?

No Robert, unless you show yourself, on this practical example, how you can apply your method, I consider it is a valuable pure theorical approach but completly inapplicable in practice (unless for quite obvious positions OC). It is up to you to prove you can yourself apply it with the position I proposed in this thread. I am unable to help you on this point.

[RobertJasiek]

We can make an agreement on what assumptions not to verify (such as B shall be sente and gennan's earlier value -9 shall be correct, which I have not verified).

However, to apply theory instead of just guessing(!) which move might be right, it is mandatory to calculate the aforementioned values at least approximately. (Depending on the case, it can be sufficient to calculate only some particular ones.) This calculation is what takes the time because the local values depend on sequences and positions, which must be constructed and annotated carefully to avoid big evaluation mistakes, as you could already see so far. If it shall be clear enough to convince everybody at least as an approximation, for every such complicated shape example, it takes hours to compose, annotate and use for calculation the sequences and positions.

Time that I do not have. I am more than busy with work and private matters. The time for doing these things preceding application of the theory must come from you other users here.

When the necessary values will have been calculated at least approximately, then application of the theory only takes minutes so then I will be able to take the time and do it.

If you absolutely want me to do all the work of every overly complicated example you show, you need to pay me for my spent time so that I do not starve. Seriously. Time I cannot spend on work loses income. It is 2 pm here and I have worked all day before my breakfast now. This is how little time I have.

Why must you often start from too complicated positions to show off the real game world? Can you not start from a simplified example for which application of the theory is fast, let us apply it, then slightly modify the position to add some first complication, let as discuss necessary approximation, then iterate until eventually reaching real game complications? You act like life and death book problems are beneath you because they are not as complicated as real game life and positions. We cannot solve everything and even quickly just because it is realistic.

If you want guesswork solutions, use the method of reading and counting for your whole board position: emulate the gote option and see a count of a follow-up quiet position; do so also for the sente option; compare! No too-time-consuming theory needed! Approximation only but you have declared to prefer it anyway.

[Gérard TAILLE]

We can make an agreement on what assumptions not to verify (such as B shall be sente and gennan's earlier value -9 shall be correct, which I have not verified).

However, to apply theory instead of just guessing(!) which move might be right, it is mandatory to calculate the aforementioned values at least approximately. (Depending on the case, it can be sufficient to calculate only some particular ones.) This calculation is what takes the time because the local values depend on sequences and positions, which must be constructed and annotated carefully to avoid big evaluation mistakes, as you could already see so far. If it shall be clear enough to convince everybody at least as an approximation, for every such complicated shape example, it takes hours to compose, annotate and use for calculation the sequences and positions.

Time that I do not have. I am more than busy with work and private matters. The time for doing these things preceding application of the theory must come from you other users here.

When the necessary values will have been calculated at least approximately, then application of the theory only takes minutes so then I will be able to take the time and do it.

If you absolutely want me to do all the work of every overly complicated example you show, you need to pay me for my spent time so that I do not starve. Seriously. Time I cannot spend on work loses income. It is 2 pm here and I have worked all day before my breakfast now. This is how little time I have.

Why must you often start from too complicated positions to show off the real game world? Can you not start from a simplified example for which application of the theory is fast, let us apply it, then slightly modify the position to add some first complication, let as discuss necessary approximation, then iterate until eventually reaching real game complications? You act like life and death book problems are beneath you because they are not as complicated as real game life and positions. We cannot solve everything and even quickly just because it is realistic.

If you want guesswork solutions, use the method of reading and counting for your whole board position: emulate the gote option and see a count of a follow-up quiet position; do so also for the sente option; compare! No too-time-consuming theory needed! Approximation only but you have declared to prefer it anyway.

I fully agree with you Robert : applying theory for such position is far too time consuming. I perfectly understand you do not want to lose time applying the theory in such position and you have to understand that I also do not want to apply theory for such position, for exactly the same reason. In this context can you simply avoid suggesting applying this theory you do not accept yourself to use because it is too time consuming? In practice you can hardly take more than 5' to decide whether you will choose a sente option rather than a gote one in a local position.

Note that in my very first post of this thread I did not suggest to apply this inapplicable theory, I was only looking for your view as go player (not as yose theoritician) when you have to choose between the gote move "a" and the sente move "b".

BTW Robert, assume you are really able to stricly apply the theory. Assume you calculate that a sente option will gain 7 points while a gote option will gain (only) 13 points. You will guess that the sente move should be the best option but you might be wrong in practice because the environment is rarely ideal.

As a theory, be sure I consider it as very valuable, but in practice it is for me completly inapplicable.

In practice I guess each player has developped her own evaluation method and for sure I will never say that my own approach for the calculation is the best one.

[RobertJasiek]

can you simply avoid suggesting applying this theory you do not accept yourself to use because it is too time consuming?

This is your wrong interpretation of my choice. My choice is a) to not always spend much time whenever somebody else (in this case: you) wants me to spend much time and b) to spend much time when I want to, and can, do so on my own (E.g., during playing my own game and I have enough thinking time left. E.g., in a recent tournament game with ca. 90 minutes basic time left, I spent 60 minutes on one position's large, game-deciding life and death problem. I tried, did a pretty good job at reading but still failed - but if I had succeeded at completing the reading correctly to its deepest ends only then involving rare tesujis, I would have won the game.).

assume you are really able to stricly apply the theory. Assume you calculate that a sente option will gain 7 points while a gote option will gain (only) 13 points. You will guess that the sente move should be the best option but you might be wrong in practice because the environment is rarely ideal.

In the simplest case of a high temperature, the creator chooses the gote option or tenukis so we might ignore the sente move value. In that case, all we need of the sente option is (an approximation of) the follow-up move value to verify that indeed the temperature is high. Determining this and (an approximation of) one initial move value (the one of the gote option) should be possible within 5 minutes.

each player has developped her own evaluation method

LOL. Like my endgaming was as a low dan confusing gote and reverse sente. It was my own method but the error was at least the factor 2. Or just guessing move values - was as bad.

2 EDITs

[Gérard TAILLE]

[quote="RobertJasiek"]In the simplest case of a high temperature, the creator chooses the gote option or tenukis so we might ignore the sente move value. In that case, all we need of the sente option is (an approximation of) the follow-up move value to verify that indeed the temperature is high. Determining this and (an approximation of) one initial move value (the one of the gote option) should be possible within 5 minutes.[quote]

Here I recognise some similariry with the way I calculate but I am not sure to understand what you mean exactly.
You mentionned a "high temperature". This temperature is related to what? The local position or the environment?

[RobertJasiek]

Unless I explicitly say otherwise, I always mean the environment's ambient temperature.

[Gérard TAILLE]

In the simplest case of a high temperature, the creator chooses the gote option or tenukis so we might ignore the sente move value. In that case, all we need of the sente option is (an approximation of) the follow-up move value to verify that indeed the temperature is high. Determining this and (an approximation of) one initial move value (the one of the gote option) should be possible within 5 minutes.

each player has developped her own evaluation method

OK Robert. Let's assume the ambiant temperature is high. If you decide to play in the local position that means that the value of a move in this local position is also high and more or less equal to the ambiant temperature right? In that case, is it a good understanding that you prefer the gote move (=> in that case the value of this gote move is about equal to the ambiant temprature) and you simply ignore the sente move? If not what do you mean exactly in your post? My view is that I have to compare the sente value to half the value of the gote move but you seem to introduce a new idea I have never heard before.

[gennan]

I'll try to use these pages from senseis: https://senseis.xmp.net/?MiaiCounting and https://senseis.xmp.net/?MiaiCountingWithTrees

I put positions A, B and C in a tree, taking left branches as black moves and right branches as white moves (but I collapse some sequences).

$$W A $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$W A $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

$$W B: count = 2 $$ ---------------------------- $$ . . . . . T T 6 2 5 S S . | $$ . . . . . T T 4 3 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$W B: count = 2 $$ ---------------------------- $$ . . . . . T T 6 2 5 S S . | $$ . . . . . T T 4 3 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

$$W C: count = -5 $$ ---------------------------- $$ . . . . . 6 5 7 S S S S . | $$ . . . . . 0 4 3 9 1 O S . | $$ . . . . . . . X O 8 X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$W C: count = -5 $$ ---------------------------- $$ . . . . . 6 5 7 S S S S . | $$ . . . . . 0 4 3 9 1 O S . | $$ . . . . . . . X O 8 X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

Code: Select all

     A
    / \
   B   C

Filling in counts:

Code: Select all

     A
    / \
   2  -5

Backtracking counts and calculating the temperature (between parentheses) of position A:

Code: Select all

   -1.5 (3.5)
    / \
   2  -5

What I further understand from those pages:

If the ambient temperature < 3.5, black should respond locally like B (or an equivalent move) with count 2.
In that case we need to prune C from the move tree, so we get:

Code: Select all

     2 ()
    /
   2

However, if the ambient temperature > 3.5, black should play a bigger move elsewhere, allowing white to move to C with count -5.
I also think the ambient temperature should not be too high (like >= 4.5), otherwise white should not even be playing in this area to begin with.

An alternative for position A is position D:

$$W D: count = 2 $$ ---------------------------- $$ . . . . . . . 1 . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$W D: count = 2 $$ ---------------------------- $$ . . . . . . . 1 . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

So it seems to me that D is equivalent to A when the ambient temperature <= 3.5.

But if 3.5 < ambient temperature < 4.5, is A better for white than D (because A would then have count -1.5)?

Edit: Changed 'miai count' to just 'count' after Robert Jasiek's remark about terminology below.

[RobertJasiek]

Let's assume the ambiant temperature is high. If you decide to play in the local position that means that the value of a move in this local position is also high and more or less equal to the ambiant temperature right?

Since you do not know my theory, why have you praised it?;)

In the case of a high temperature T, compare it to the gote move value Mgote to decide whether to play locally the gote option!

T ? Mgote

Simple, isn't it?:) The proof has been quite an ordeal though...! I have needed to exclude the sente option.

My view is that I have to compare the sente value to half the value of the gote move but you seem to introduce a new idea I have never heard before

You can profit from my theorem that only the gote option matters! (If the temperature is high.)

[RobertJasiek]

gennan, those webpages are for ordinary local endgames - they are not for local endgames with gote and sente options!!! The latter have a player's two alternative moves from a position to either the gote follower or the sente follower. Such as

Code: Select all

        A
      / /\
     / /  \
    G  S   R

There is no such thing as a miai count. Modern endgame theory aka miai counting is the field of studying (per-move) move values aka miai values, counts and gains.

[gennan]

There is no such thing as a miai count.

Thanks, I changed 'miai count' to count' in my previous post.

[gennan]

So can we say that this position has count 2, because white has a sente sequence to reach that count and 'sente gains nothing'?

$$ F: count 2 $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$ F: count 2 $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

I try to calculate the miai value of black's reverse sente move:

$$B G $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$B G $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . 1 O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

$$B H: count 16 $$ ---------------------------- $$ . . . . . T T T T T . 7 . | $$ . . . . . T T T T 1 Q 5 6 | $$ . . . . . . . X Q 3 X O . | $$ . . X , . X . X X X 2 O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$B H: count 16 $$ ---------------------------- $$ . . . . . T T T T T . 7 . | $$ . . . . . T T T T 1 Q 5 6 | $$ . . . . . . . X Q 3 X O . | $$ . . X , . X . X X X 2 O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

$$B J: count 9 $$ ---------------------------- $$ . . . . . T T T T . . S . | $$ . . . . . T T T T 1 O 4 . | $$ . . . . . . . X Q 3 X O . | $$ . . X , . X . X X X 2 O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$B J: count 9 $$ ---------------------------- $$ . . . . . T T T T . . S . | $$ . . . . . T T T T 1 O 4 . | $$ . . . . . . . X Q 3 X O . | $$ . . X , . X . X X X 2 O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

Code: Select all

           F
          / \      
         G   D
        / \ 
       H   J 

Code: Select all

           2
          / \      
  (3.5) 12.5 2 
        / \
      16   9 

Would this mean that black's reverse sente move (G) has miai value 10.5, so white should play his sente before the temperature drops to near 10.5?

[RobertJasiek]

So can we say that this position has count 2, because white has a sente sequence [...]

Go Diagram
F: count 2
Click Here To Hide Diagram Code

$$ F: count 2 $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |

Click Here To Show Diagram Code

[go]$$ F: count 2 $$ ---------------------------- $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . X O . X O . | $$ . . X , . X . X X X . O . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . O . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . . . . | $$ . . . . . . . . . . O . . | $$ . . . , . . . . . , . . . |[/go]

What sente sequence?

[gennan]

I mean the large monkey jump (D in my post from about 3 hours ago).