Cards or app for miai-value based endgame practice?

[Gérard TAILLE]

Area scoring: any drop, but exceptions would occur more easily in the 2 points and below range of move values.

OK let's try, as in territory scoring, a drop egal to 1/2.
I see how to build simple gote without follow-ups with the value 1, 2, 3, 4 ...
but how do you build simple gote without follow-ups with the value 0.5, 1.5, 2.5, 3.5, 4.5 ...?

[RobertJasiek]

If you want to represent it on the board for area scoring, multiples of 1 seem necessary while 1/2 is impossible. Do you see a necessiety for 1/2?

[Gérard TAILLE]

If you want to represent it on the board for area scoring, multiples of 1 seem necessary while 1/2 is impossible. Do you see a necessiety for 1/2?

I agree with you, multiples of 1 is necessary.
BTW it was just a question for all those (like me) who use area scoring.
I am not against a drop equal to 1. I just note that the anomalies you mentionned earlier occur more easily when the value of the drop is higher.

So far my understanding is that, with different definitions, the miai value according to thermography or CGT and the move value according to your theory or according to mine seem identical (no counter example identified).

The main differences of the different approaches concern the use of an (ideal?) environment.

Basically, I use only one environment I qualify as a "sufficiently rich environment". The basic idea behind is the following.
Though no ko is introduced, this "sufficiently rich environment" is a kind of average of all realistic environments encounter in practice. Considering a local position, the idea is to put this local position into this environment and to identify at which temperature the players MUST play locally to reach the best result. That way if temperature corresponding to position P1 is higher that temperature corresponding to position P2 then I conclude that there is a good chance that you must play in position P1 before playing in position P2.
OC it may not be true in the real environment in which position P1 and P2 appear but anyway it is good indication for the player who can complement the analysis by a reading process.

I understand you calculate a move value without the need of an environment which is fine. Knowing that in a real game an environment is almost never an ideal environment (with your definition) I do not see what is the purpose of your ideal environment. Is it simply a pure theorical use (as example?) or do you propose to real players a new tool, as complement of the move value itself?

[RobertJasiek]

The ideal environment is like an average of real positions. It enables theorems as exemplified before. We can also frequently apply such advice as approximation for the timing of non-simple local endgames.

You can divide the anomaly values by 10 and replace 5.3 by 5 1/3. I have constructed such examples, which are not exactly simple but almost so with a basic ko follow-up.

[Gérard TAILLE]

I do not use ideal environment to define count and move value! I need not. I define via tentative counts and tentative move values - gote or sente values being larger, equal or smaller. Local evaluation independent of environments! Environments for value definitions are overkill for practical application. Counter-examples (without kos) to the method of making a hypothesis are so hard to find that none is known on a go board so far.

Sure my definition of "move value" of a position uses explicitly what I called a "sufficiently rich environment" but it appears really that the calculation of this "move value" depends ONLY on the subtree of the position and not the environment itself!
IOW I completely agree with you that the local evaluation has better to be independent of environments.
The real difference between us is that you defined the "move value" by its calculation where I tried to define the "move value" by the need of go players (I mean a good tool to compare positions in the context of more or less representative environment).
Assume you choose "naïve move value" considering all nodes in the subtree as gote nodes (that way the recursive calculation is very simple OC). Obviously, it is a possible calculation and a possible definition for a "move value" but you will be right to say that this definition is unuseful and another calculation will be better. What does that mean? That means that you have really in mind some properties of the "move value" and if these properties are not fulfilled, you’re not happy with your calculation.
My question is the following: what are these properties you used to build your calculation?
For me these properties may be considered as the real definition and the calculation a consequence.

[RobertJasiek]

I am unsure what properties you are asking for other than those already mentioned.

[Gérard TAILLE]

I am unsure what properties you are asking for other than those already mentioned.

If you do not see properties other than those already mentioned, then it is good news isn't it? Let me remind you the two following ones:

1) Considering a "sufficiently rich environment" the "move value" of a position P is the temperature under which it is better to play locally
2) This “move value” depends only on the tree but not the environment itself.

OC, you recognize that properties 1 is my own definition of “move value”.
I understand your definition is the calculation itself (I do not know it because I saw only some examples) but I am pretty sure that the definitions are equivalent. I guess also that your “move value” is equivalent to the “miai value” used by Bill but here again I do not know. Did you ever proved this last point?

[RobertJasiek]

For my move value, count and gains, 1 no and 2 yes.

CGT values also depend on infinitesimals so not necessarily equivalent to not considering them, as typically Bill has done and I do.

If we calculate mod inf (modulo infinitesimal), there surely will be many cases of same move value as miai value. However, I do not know if this is always so and I am far too cautious to predict either. My study of CGT reversal using difference games my traversal and Bill's / my method of comparing the opponent's branches for tree simplification has made me aware of the existence of pathological examples (which I do not recall by heart but see [22] for thermography versus traversal using my method of making a hypothesis) for these for which the three methods produce different values or behaviours. Not all we wish to be equivalent is.

Bill has used different tools, incl. miai value versus move value. For move value and count, he mentioned relations between different move values and the counts in a clause of a sentence as conjectures. I have proved them but he already applied them.

[Gérard TAILLE]

For my move value, count and gains, 1 no and 2 yes.

Oops, your answer is quite surprising. It will be an advantage if your "move value" had property 1 but you claim it is not the case!
I do not know your calculation (I am just guessing) but I can hardly believe your move value has not this property. Without any counter example I prefer to say that your move value has really this property.
BTW seeing you studied in detail the difference between your “move value” and Bill’s “miai value”, can you show us an example of position where the two values differ.

[RobertJasiek]

seeing you studied in detail the difference between your “move value” and Bill’s “miai value”, can you show us an example of position where the two values differ.

Uh, it is not Bill’s “miai value”, which was introduced by Conway and reintroduced by Berlekamp and probably others. (I do not know Guy's contribution.)

I have not studied practical differences of move value and miai value carefully. Ad hoc, I only know of Francisco Criado's counter-examples for move value as for the method of making a hypothesis versus thermography.

Counter-example 1:

A = {B|-4}, B = {5|C,D}, C = {-2|-6}, D = {-1|-9} in a rich environment.

Counter-example 2:

P = {Q|-3}, Q = {8|R}, R = {S|-1}, S = {7|U}, U = {1|-3}.

For explanations why these are counter-examples, see [22].

[Gérard TAILLE]

seeing you studied in detail the difference between your “move value” and Bill’s “miai value”, can you show us an example of position where the two values differ.

Uh, it is not Bill’s “miai value”, which was introduced by Conway and reintroduced by Berlekamp and probably others. (I do not know Guy's contribution.)

I have not studied practical differences of move value and miai value carefully. Ad hoc, I only know of Francisco Criado's counter-examples for move value as for the method of making a hypothesis versus thermography.

Counter-example 1:

A = {B|-4}, B = {5|C,D}, C = {-2|-6}, D = {-1|-9} in a rich environment.

Counter-example 2:

P = {Q|-3}, Q = {8|R}, R = {S|-1}, S = {7|U}, U = {1|-3}.

For explanations why these are counter-examples, see [22].

As I already said the reference [22] is not accessible via this site, so I cannot access to the explanations.

Anyway let me try to analyze Counter-example 2:
Node U : count = -1 move value = 2
Node S : count = 3 move value = 4
Node R :
Black to move : as Go player I can see that the exchange RS – SU reverses => after RS – SU black must continue towards the score +1
White to move => score -1
Eventually node R : count = 0 move value = 1
Node Q: count = 4 move value = 4
Node P : black PQ – QS is sente => node P : count = 0 move value = 3
What is your “move value” and what is the “miai value” ?

[Gérard TAILLE]

Counter-example 1:

A = {B|-4}, B = {5|C,D}, C = {-2|-6}, D = {-1|-9} in a rich environment.

My analyze of counter example 1
Node D : count = -5 move value = 4
Node C : count = -4 move value = 2
Node B: count = 0 move value = 5
Node A:
Here as a go player I see that AB-BC as well as AB-BD reverses => black must continue after either of the sequences => white has better to choose AB-BC => black reach only the score -2 in gote
Finally node A : score = -3 move value = 1
Here again what is your “move value” and what is the “miai value” ?

[RobertJasiek]

As I already said the reference [22] is not accessible via this site

It is always the same thread where you find these references:
Book References in Other Threads
https://www.lifein19x19.com/forum/viewt ... =17&t=8765

Anyway let me try to analyze Counter-example 2:
[...]
Node R :
Black to move : as Go player I can see that the exchange RS – SU reverses => after RS – SU black must continue towards the score +1

The method of making a hypothesis does not use reversal but compares move value to gains in a long sequence. Here, it detects a 5-move traversal sequence but it is wrong.

node P : count = 0 move value = 3

Correct but also for wrong reasons. You must not be even more naive than me by even disregarding the possibility of long sequences...! We need to apply thermography to justify these values!

My analyze of counter example 1
[...]
Node B: count = 0 move value = 5
Node A:
Here as a go player I see that AB-BC as well as AB-BD reverses => black must continue after either of the sequences => white has better to choose AB-BC => black reach only the score -2 in gote
Finally node A : score = -3 move value = 1

You need to take into account the rich environment and enriched scores. Then start afresh for my method of making a hypothesis.

[Gérard TAILLE]

In your post viewtopic.php?p=277835#p277835 you showed two counter examples for which you claim your "move values" are not equal to "miai values".
Just for your information, and without great details on my calculation, I gave you my own results for these two examples but that it is not, for the time being, the point.
Please can you complete your post by giving us your "move values" and the "miai values" for both examples?
That way the discussion on the different calculation methods would be far clearer. IOW I will then be able to give you the best answer to your previous post viewtopic.php?p=277840#p277840

[RobertJasiek]

Counter-example 1:

The problem of the method making a hypothesis is that, at low temperature T = 1,5, both options C and D are fulfilled, for option C giving the initial count -3 and move value 1 but for option D giving the initial count -2,5 and move value 1,5. A contradiction arises, which does not arise for thermography and the initial count -3 and move value 1.

Counter-example 2:

Method of making a hypthesis: initial count -1 and move value 2 derived from a 5-move sequence.

Thermography: initial count 0 and move value 3. Note that my description of thermography speaks of move value and I dare to do so because I ignore infinitesimals. I suspect that miai values with infinitesimals might differ from move values, but I do not know examples.