K_Brown wrote:
... In this sense, everyone is already making the best move ...
Not any sense that I am advocating. I try (not very hard) to maximize the probability of achieving a goal, but, of course, failure is an option.
Define "Best Move"
Sort:
K_Brown wrote:
… Are you saying that there is no way to define a best move in human play. ...
We can "define" "best move" about as well as we can define anything, but that is somewhat different from identifying the best move for a particular position, player, and opponent.
K_Brown wrote:
… Postmortem analysis will tell the human if he made the best move ...
I see no reason to believe that we can dependably do more than guess at what would have maximized the probability of achieving a goal in this or that position. Of course, we can be confident for some positions.
K_Brown wrote:
… Their memory will tell them the best move when that position happens again in a later game (this is what i'm talking about when I say that chess will become a game of memory). … What I'm saying is that best move in human play will be based off of memory, not style. ...
As long as brains remain as they are, players will, most of the time, face positions that they do not remember. If a position is not remembered, a player may well want to consider his or her own abilities as well as those of the opponent when trying to maximize the probability of achieving a goal.
K_Brown wrote:
… If it is a complicated position and they find that the best move is harder to play against a human opponent, they will happily play an inferior move because they are more comfortable with it. ...
I would think that it would be fairly rare for a player to reject a move while believing that it would maximize the probability of achieving a goal.
K_Brown wrote:
… Just because it is easier for a human to play an inferior move, doesn't somehow make it the best move. ...
I have not written that an inferior move is best just because it is easier for a human to play. For one thing, I would fear that "inferior", "best", and "easier" all have different meanings for you and me.
K_Brown wrote:
… By outside analysis, I mean analysis that has taken place outside the game. ...
It seems to me that outside-the-game position analysis can reasonably consider the abilities of the player and the opponent.
Oct 28, 2018
0
#169
Ahptoemiz escreveu:
Assuming it’s true that for any given position there is an objectively “best” move (or alternatively a co-equal set of best moves) then how do we define this in all instances?
In some instances defining best is trivial. If only one move delivers mate in one and all other moves draw then, obviously, the move that mates is best. And if there are multiple moves that mate in one while all others allow a draw then any of the mating moves are best. But what if we have a position that has two mating sequences, one a mate in one and one a mate in three, is the mate in one “better” just because it involves fewer moves? The scoreboard will still read the same after the game is over.
And granting that a faster win is better (I think most would) What about drawn positions? Can we distinguish “better” between two moves that both should result in a draw? If you say no then how could one differentiate between, say, different openings when, in all likelihood, no opening leads to a forced win for white and black should always equalize (regardless of whether it’s on move five or 25) and then draw.
Regarding drawn positions let me answer my own question and propose that any move that maximizes the potential for one’s own side to achieve the best possible result is best. So, for instance, if I play a move that my opponent can answer with eight possible moves and five will draw and three will be a win for me, that is objectively better than if I played an alternate move that allows six moves that will draw and two that will win for me.
Thoughts?
'concrete' best move are moves like "mate in two" instead "mate in four". Subjective best move are "white is wining, and keeps winning until win the game". If white, in some moment, lost the advantage, in that move white make a mistake, or if white lost a half of your advantage, white make a innacuracy. For example, if white end the oppening with black better, white make some mistakes. If black equalize, white make some innacuracy!
Oct 28, 2018
0
#170
Daniel1115 wrote:
I have not seen you stating any facts. Therefore you are not stating 1+1=2. Its more like saying that 0.99999..999 (to infinity)=1. Its more of an opinion and how you look at it.
Also, you are very defensive. What kinda spongey meant (I am assuming) is that nobody plays perfect, therefore practical chances are important in any practical position (a real game). You only play perfect with engine assistance.
Not to be a jerk, but .999999999 repeated for all eternity does equal 1. It’s not an opinion. It’s a mathematical proof, and a very important one too.
Oct 28, 2018
0
#171
kindaspongey wrote:
K_Brown wrote:
… Their memory will tell them the best move when that position happens again in a later game (this is what i'm talking about when I say that chess will become a game of memory). … What I'm saying is that best move in human play will be based off of memory, not style. ...
As long as brains remain as they are, players will, most of the time, face positions that they do not remember. If a position is not remembered, a player may well want to consider his or her own abilities as well as those of the opponent when trying to maximize the probability of achieving a goal.
I think this sums up your side very well. I agree with this completely. When the human brain inevitably fails at trying to remember every single position, they will have to find the subjective "best move" which can vary person to person.
As long as we agree that while there will be subjective best moves in these cases, there will also be an objectively best move, in all applicable positions, in the theoretical future.
Most of the time, these will probably vary. As our understanding of chess increases, maybe we will close the gap a little on this. I am inclined to believe that we are approaching a plateau on how much better humans can get at chess. They may have thought the same thing in the past too though.
K_Brown wrote:
… there will also be an objectively best move, in all applicable positions, in the theoretical future. …
That sounds to me like it could be a departure from current language usage. (Depends, I suppose, on what you have in mind by "applicable".) Of course, it is possible that language usage will change to what you describe.
K_Brown wrote:
Daniel1115 wrote:
K_Brown wrote:
Give me a post # to address first of all, because I didn't see it.
Second, you have no idea what I'm saying or you would understand the analogy because the analogy itself is solely based on a hypothetical future where tablebase 32 has been developed. That is determined with context clues, which will never be easy to demise by someone who constantly takes people out of context...
There is an issue with the 32 tablebase theory (I did not realise you assumed you were arguing assuming such a thing existed). It is not currently feasible, we would need massively improved technology, this could be something 50-100 years from now. I would say that assuming you have a 32 piece tablebase is the "easy" way to try to argue it, since it calculates every scenario from every possible position. The problem with taking this side is you are in essence saying that it is impossible to determine the best move without the 32 tablebase (if it is not than why do you use that in your argument). It is also equivalent to: "Whats the best move": "Well after calculating all possible lines, this line forces a win". The tablebase is better than any engine, since its never wrong and never incorrectly evaluates the position (i.e. as forced win, forced loss forced draw). However the resources to determine the result of ever position are gargantuan.
However there are some inherent flaws (which I think kinda spongey was touching on). How does the tablebase determine which moves that all lead to a forced draw with best play, are the best? How does it determine which moves from a forced loss position (assuming its against another 32 piece tablebase) are better than others. Your whole idea that it will give forced mate in X is flawed because the path chosen can easily be the one of least resistance.
Keep in mind the whole probability of victory argument does not work for the range of drawing moves since that would suggest that it is best to capture defended pawns, since the opponent only has 1 move (normally) of several that will punish this move. This example is equivalent to one where after move X made by tablebase, there is one move that forces a loss and 2 that would result in a forced win.
Very good! Now we are getting somewhere. By the way, I tend to edit as I brainstorm. I often post something only to realize that I am leaving something out. Please look at my last post which address #162. I admitted that.
Key points in my theory:
Not all positions may be subject to this "best move" criteria. That is where the argument started. My viewpoint is that the best move only depends on the position itself. Kindaspongey seems to believe that it is dependent on the player.
The objectively best move (in applicable positions) doesn't change from player to player in my opinion.
To say that a player gets to choose from a list of best moves inherently means to me that a best move doesn't even exist in that position. I'll try an analogy:
best move= fact
best moves= opinion
A fact can never be an opinion and an opinion can never be a fact. It is one or the other.
I think mate in x would still happen because I am talking about perfect play.
You did not answer to this part of my argument, which actually is a response to your response. How does the tablebase determine which moves that all lead to a forced draw with best play, are the best? How does it determine which moves from a forced loss position (assuming its against another 32 piece tablebase) are better than others?
A tablebase itself cant give you which are really better, someone will have to manually look through to see which position is better, using their subjective criteria. If you put together any "best move criteria" it will be subjective. Even Stockfish's criteria is not perfect, it loses games. If you are assuming we have an engine that has the perfect best move criteria, than that solves the issue but is not really realistic (considering that something similar to a 32 man tablebase will be needed to verify).
Oct 29, 2018
0
#174
That is what a big portion of the text you quoted me on is referring to actually....
Forced draw = no best move
Best move = fact
Best moves = opinion
Facts can’t be opinions. Opinions can’t be facts
I already said that...
And yes i’m assuming that we are in a theoretical future that has the perfect best move criteria when a best move does exist.
K_Brown wrote:
With best play chess is a forced draw. You still have not answered how you would determine what move to play(e.g. move 1, move 2 etc.). How would this best move criteria be developed. The best move(s) are only as factually correct as the criteria used to select them. How would you determine this best move criteria.
Keep in mind that defining the best move as the one that follows the best move criteria only leads you to the equivalent question of " what is the best move criteria"
That is what a big portion of the text you quoted me on is referring to actually....
Forced draw = no best move
Best move = fact
Best moves = opinion
Facts can’t be opinions. Opinions can’t be facts
I already said that...
And yes i’m assuming that we are in a theoretical future that has the perfect best move criteria when a best move does exist.
Defining it as such is just rephrasing the question.
Oct 29, 2018
0
#177
You keep saying that chess is a forced draw with perfect play. That not only goes against current understandings but also is a shaky theory. I’m not rephasing the question, I am answering it. You are asking a question that is flawed to begin with since no singular best move exists as I’ve had to say 3 times now.
K_Brown wrote:
You keep saying that chess is a forced draw with perfect play. That not only goes against current understandings but also is a shaky theory. I’m not rephasing the question, I am answering it. You are asking a question that is flawed to begin with since no singular best move exists as I’ve had to say 3 times now.
Um, let me prove my statement by contradiction. Assuming chess is not a draw with best play, how come we have not found any forced wins for whites starting move 1? In fact, we have not found any forced wins in any of the GM (2700) approved openings, otherwise no one would play them at that level. Computers have not found a forced win. There is a reason the common understanding of starting eval is 0.5 for white, not winning.
Well, how do you determine the "range of best moves" than. Everything I said applies to the range of best moves as well.
: ) > The Best Move.... is Pawn Forward. No lie.
Goodbye
And Goodluck.
Oct 29, 2018
0
#180
I don't understand what you mean by that at all. We haven't found any forced wins from move 1 because the amount of positions that have to be calculated is extremely vast. Why do you think we are only at tablebase 6 right now?
Range of best moves are most easily defined in a drawn position. There are often many moves that are evaluated as 0.00 Or we could say we are talking about a mate in 3 vs a mate in 3 utilizing different moves but achieving the same result in the same amount of time.
Let's use a bare bones example and say that a position consists of only two kings. Is there a best move in that position? No, because that is a position where a best move doesn't exist. It is drawn due to "insufficient mating material" no matter what the move is. Any time there is a position "range of best moves" then it fails to meet the criteria needed to have a singular "best move".
I have been speaking as if it was understood that certain positions (aka ones with a range of best moves) don't factor in to this conversation as they could be described as NA (non-applicable)
… I am saying that it is a theoretical plausibility that in the future we will be able to determine whether or not a player played the best move from any position. ...
As long as 'best' continues to refer to a human playing (without computer use) in such a way as to maximize the probability of achieving a goal, I see nothing that rules out variance from one player to another with the moves to do that.
... doesn't everyone think they are doing that every time they are playing? ...
Well, speaking for myself, I seem to often fail to maximize the probability of achieving a goal.