hgm wrote: Starting from nonsense is a bad idea, and the fact that it might be simple nonsense does not improve that.
Simple and nonsense are 2 different concepts.
A simple SEE is not equivalent to a SEE with all pieces values at 1. At least it is not equivalent for anybody that knows what SEE is. A simple SEE is just a SEE that works as expected (and as described in literature), without additional fancy.
hgm wrote: Starting from nonsense is a bad idea, and the fact that it might be simple nonsense does not improve that.
Simple and nonsense are 2 different concepts.
A simple SEE is not equivalent to a SEE with all pieces values at 1. At least it is not equivalent for anybody that knows what SEE is. A simple SEE is just a SEE that works as expected (and as described in literature), without additional fancy.
So you argumentation is pure void content.
Simplify something more than it could bear, and you get nonsense.
If you cannot understand that, you wouldn't get very far developing a novel engine, as supposed to implementing a recipe for one from the literature. SEE means just what it says: "static exchange evaluation". That is the score benefit ('evaluation') from an optimal sequence of capture moves ('exchange') to the same square ('static', i.e. not considering the dynamic consequences of the disappearing pieces). With nonsense values for the pieces, as you advocate because it happens to be described in literature, and most literature you know happens to be about orthodox Chess, where a location-independent piece value happens to make sense, it would just be an example of "garbage in, garbage out".
You seem to suffer from the same problem as that infant of the psychological experiment, where they showed him the same book every time, telling him: "this is a book". Then they brought him to a book shelf, and asked him: "bring me all the books you see". Sure enough, the infant brought all the copies he could find of that same book, leaving all other books untouched...
hwiechers wrote:Including the PST values doesn't make sense to me.
Unless your SEE implementation handles xrays, pinned pieces, capture orderings and zwischenzugs, you're worrying about a small inaccuracy and ignoring much larger ones.
I use the SEE mainly for the ordering of capture moves, and it is not difficult to understand that using the PST will have influence on the order of nearly equal captures. In my case it costs hardly any extra time and for most positions the tree size gets somewhat smaller. Of course this behavior will be different for other engines.
In my engine the most time is consumed by probing the hash in quiescence and the evaluation function, so I'm not worried at all about a few extra instructions in the SEE.
hgm wrote: Simplify something more than it could bear, and you get nonsense.
In rhetoric theory there is a specific term to identify your answer: it is when you pretend that your opponent has said something (that he hasn't) and you attack that because you are not able to attack what he actually said. If I remember correctly it is called "Straw man" attack. An interesting attempt, although there are people much more well-versed than you in this art here on the forum. Our old timer professor is undisputed number one here
Actually your posts are more useful to me to study your way to escape from hopeless situations (and you are not bad in this regard, I have to admit), than for the specific technical content.
Yeah, obviously you did not learn what SEE actually means... (Oh that useless technical content! )
But never mind if this is all beyond you. I am sure most other readers will perfectly understand what I wrote. So there is no need to discuss it further.
hgm wrote:Yeah, obviously you did not learn what SEE actually means... (Oh that useless technical content! )
But never mind if this is all beyond you. I am sure most other readers will perfectly understand what I wrote. So there is no need to discuss it further.
I think that you are right.
It is obvious that you talk about games when
"it is known that location is all-important" and not about chess when location has little value relative to the identity of the piece.
To make things clear for marco
If it is known that the value of rook in the center is 900 and the value of rook in the corner is -100(it is better to be without it) then it is not logical to start with SEE that does not use the location of the rook in the first place(for the discussion 900 advantage means the same advantage as queen advantage in chess).
I do not know about games when it happens but I guess that you know more about non chess games than me.
hgm wrote: Simplify something more than it could bear, and you get nonsense.
In rhetoric theory there is a specific term to identify your answer: it is when you pretend that your opponent has said something (that he hasn't) and you attack that because you are not able to attack what he actually said. If I remember correctly it is called "Straw man" attack. An interesting attempt, although there are people much more well-versed than you in this art here on the forum. Our old timer professor is undisputed number one here
Actually your posts are more useful to me to study your way to escape from hopeless situations (and you are not bad in this regard, I have to admit), than for the specific technical content.
I guess you will be remembered as "just someone who is annoying as hell"?
HGM made a point. It apparently went over your head. You have been arguing over your head since then.
bob wrote: I guess you will be remembered as "just someone who is annoying as hell"?
HGM made a point. It apparently went over your head. You have been arguing over your head since then.
I won't be remembered.
I was talking of chess and I stated it clearly (no poker, etc). HGM argumentation were just bla bla. You even didn't reached bla bla level in your post.
bob wrote: I guess you will be remembered as "just someone who is annoying as hell"?
HGM made a point. It apparently went over your head. You have been arguing over your head since then.
I won't be remembered.
I was talking of chess and I stated it clearly (no poker, etc). HGM argumentation were just bla bla. You even didn't reached bla bla level in your post.
HGM talked about SEE not in chess and he explained that the same idea is correct also for developing a new chess engine(and he did not mean specifically SEE).
When you start a new chess engine
You do not start with piece values that are the same but from common knowledge about the value of pieces.
HGM did not say that in chess it is a bad idea to start with the normal SEE that most program use and later test changes to it.
)