BTW a sente position can be simply a position for which a part of the mast is colored, without being a double sente position.

This is a good candidate for a definition in thermography. (However, why at all? Why not just call such local endgames "purple"?) I have not studied purple examples carefully yet but recall that they are infrequent or related to rare kos.

Since thermography is pretty well developed, such a definition has a good chance of being useful. I have, however, used thermography so infrequently that I cannot say ad hoc if typical candidate shapes, such as doubly ambiguous, qualify. (In such a theory, there might be no necessity for the extra type doubly ambiguous but it might be put in the double sente category.)

I studied thermography in the past but I do not either use it. Anyway if it is a good mean to exchange avoiding ambiguity why not.
May be I can see a better wording. Considering purple color is both a blue and a red color:

A position is a black sente position if a part of the mast is blue.
A position is a white sente position if a part of the mast is red.
A position is a gote position if it is neither black sente nor white sente.
A position is double sente if it is both black sente and white sente.

I did not take into account ambiguous position. If it is a real need OC I can try to complete the definitions.

Thermography, which uses thermographs algebraically or graphically, is powerful and I do not want to discourage you from enhancing it. However, it is inapplicable by players in practice. For application to a typical example, I need hours. General principals for fast application to whole classes of positions are too rough or otherwise missing. This is very different from theory for the temperature and a few move values or counts, for which application time is fast or reasonable.

As I said clearly in my previous post I do not use thermography in practice. I completly agree with you it is inapplicable in practice. I use it here only to show you clearly what means "sente", "gote, or "double sente" for me.
Because the defintion I proposed are simply based on the definition of "sente" the challenge is only to find a simple defintion of "sente" immediatly understandable by any go player not necessarily aware of the huge amount of work made in the theory field.
Do you have yourself such simple defintion of "sente" position (if it is not a secret in your books OC)?

There has also been my naive definition (uh, even in writing): one player has a sente sequence and the opponent has a gote sequence. Might be good enough for some players but when Bill pointed out its informality, I dug deeper.

I agree with you Robert this naive defintion is not satisfactory : in this defintion of a "sente position" you use the wording "sente sequence" but what is a "sente sequence"? Do you need an environment to define a "sente sequence"? If yes don't you have in addition to analyse at which temperature of the environment a sequence can be sente?

It is, however, the most complicated that might be immediately understandable by any go player:)

I think if you try to define a "sente position" without taking into account an environment, then this defintion as a good chance to be complicated and more or less unanderstandable for go players. That is a basic reason why I created my approach based on a rich enviroment. With such environment everything becomes far easier indeed.

With algebra for every go player?! You need to hide solving systems of equations when selling it as easy for every go player!

No Robert you are confusing the defintion of "sente position", which could be quite simple, with the "proof" that a given position is really a "sente position".
For a simple position it is easy to see if a given position is really a "sente position" according to the defintion, but for a difficult position it may be complicated to prove if this position is sente or not. In any case however, the defintion itself can be simple.

Similarly if you take a given game G, you can define the score of G as being the best score the players can reach. Because the god plays may not be obvious you see that though the defintion is quite simple and understanbable by any go player, finding this score can be quite difficult indeed.

If in your method the defintion is also the proof then I understand clearly why this defintion can be quite complicated!

LOL, no.

Good news Rboert.


Seeing that without knowing it, you pretended that my defintion cannot be simple, I had to conclude that yours is probably quite complicated for a go player. Fine if it not the case (!). I am sure the readers of yours books know perfectly that your defintions are quite simple and, respecting your works, I will not ask you to give us your own correct and formal defintion of a sente position.

The reason why my methods of classifying positions or move decisions are comparatively simple is: they are about calculating a value or comparing two values. Even if application must be interative because variations are, at every step, one compares values. Algebra beyond that occurs in the proofs but go players never need to understand the proofs to apply the theory.

Does that mean that the "sente position" DEFINITION needs calculating one or two values you have to compare?

Of course.

That is the point Robert. If you need first a calculation to DEFINE if a position is sente or not, then providing the position is quite difficult, you simply cannot DEFINE what means the "position is sente" because the needed calculation may be too complex.

Take my other example where I define the score of a game G as being the best score the players can reach. Calculating the score may be difficult but the DEFINITION itself needs no calculation and as such appears simple and understandable by each go players.
It is the same for defining a "sente position". If you need a preliminary calculation to define what a sente position is, then this defintion appears to me complex and may even not be understandable by each go player.

No.

The definition is general for a class of positions. For a particular position, the definition is APPLIED by determining the values.

Look to your naive definition:
"one player has a sente sequence and the opponent has a gote sequence".
Ignoring that a "sente sequence" and a "gote sequence" are not defined, this defintion is quite good. It is quite well understandable by any player and this defintion do not imply explicitly any more or less long calculation.

In another post you, taking into account only simple positions you defined the status of sente position as:
"local sente := MSENTE > MGOTE"
Here it is quite different. The MSENTE and MGOTE values being explicitly compared in the defintion itself then you cannot avoid calculating them to apply the defintion. It is not as understandable for a go player who have to learn what means MGOTE and MSENTE and must learn how to calculate them.
In addition you already explained that this defintion cannot be used for difficult positions (with various sente and gote options for both players). That means that this defintion is not general.

You see the difference? For me, to reach the best defintion, you have only to change the wording of you naive defintion, without introducing explicitly values you have to calculate.

What is your related suggestion of some, IYO, possibly better definition modifying my naive one?