Recommend
3 
 Thumb up
 Hide
131 Posts
[1]  Prev «  1 , 2 , 3 , 4 , 5  Next »  [6] | 

Abstract Games» Forums » General

Subject: Is it possible to have a pure abstract strategy game where no side is favored and it doesn't draw? rss

Your Tags: Add tags
Popular Tags: [View All]
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
Markus Hagenauer wrote:
There is only one last thing left to be said:

DO NOT FEED THE TROLL!

Who is trolling?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmb
docreason wrote:
Now, a question I would have is: If the you can change the rules during play, in a way that is fair and balanced, is the initial conditions of not favoring a side, and also not ending a draw, possible?
By "you" do you mean the players, by mutual consent, changing the rules?

Then the answer is course "no" because the player who's currently winning has no incentive to agree to change the rules.

If by "you" you mean some outside referees/spectators/God/whatever changing the rules, then it's no longer a pure abstract strategy game, as you've injected some element of nondeterminism/randomness/chaos beyond the players' control.

If you mean something else, please elaborate.

Quote:
I do ponder this to, because there is one noted abstract strategy game designer who insists on none of his games ending in a draw.
Sure; and his games of course favor one player in theory (like all drawless pure abstract games, e.g. Hex), but that really says nothing about whether they favor one player noticeably in practice or not.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
docreason wrote:
Now, a question I would have is: If the you can change the rules during play, in a way that is fair and balanced, is the initial conditions of not favoring a side, and also not ending a draw, possible?

I do ponder this too, because there is one noted abstract strategy game designer who insists on none of his games ending in a draw.
I fear your questions become rather exotic, although in this case I suspect you're really in the dark. "If you can change the rules during play…" - isn't there a proverb to the effect that you can't?

The weirdness of the question is obviously related to the support you seek in your ongoing debate with Voldemort. Is he awakening? surprise

(P.S. You might want to check the back of your head...)
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Herb
United States
Georgia
flag msg tools
mbmbmbmbmb
docreason wrote:
On a practical level, even Connect 4 doesn't seem to work out to be a win for the first player all the time. I have done tournaments in it, and seen cases where the player going second won all 3 games. And I am not even talking about my simultaneous version of it, but the regular version.

The strategy to play Connect 4 perfectly is so complex that a human can't keep track of the whole game tree and wonders away from the "true path" of perfect play.

A sufficiently large computer can be programmed to a perfect game. See:
"A Knowledge-based Approach of Connect-Four - The Game is Solved: White Wins"
Victor Allis, 1988

http://www.connectfour.net/Files/connect4.pdf

As I remember the game is strongly solved in that if the human is playing white and makes a mistake, the computer will find the winning play sequence.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
herace wrote:
As I remember the game is strongly solved in that if the human is playing white and makes a mistake, the computer will find the winning play sequence.
I'll quote wiki here

"The strongest sense of solution requires an algorithm which can produce perfect play (moves) from any position, i.e. even if mistakes have already been made on one or both sides."

Not that I disagree about anything you say about Connect4, other than that's not quite the same as what you seem to say. "Strongly solved" means the program can find the truth in any legal position and execute the win.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Herb
United States
Georgia
flag msg tools
mbmbmbmbmb
christianF wrote:
herace wrote:
As I remember the game is strongly solved in that if the human is playing white and makes a mistake, the computer will find the winning play sequence.
I'll quote wiki here

"The strongest sense of solution requires an algorithm which can produce perfect play (moves) from any position, i.e. even if mistakes have already been made on one or both sides."

Not that I disagree about anything you say about Connect4, other than that's not quite the same as what you seem to say. "Strongly solved" means the program can find the truth in any legal position and execute the win.
We're in violent agreement, though I'll quibble that your last sentence isn't exactly correct either.

In Connect 4, the first player should win. So a computer playing first should never lose. Such a program would only weakly solve the game.

"Strongly solved" for Connect 4 means that a computer program would play perfectly even when playing second. So no matter where the first player leads the game, if the first player makes a mistake (opens a winning line of player for the computer), then the computer would find that winning line of play.

The point is that if the computer is playing second, then the computer can not win unless the first player makes a mistake.

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
herace wrote:
"Strongly solved" means the program can find the truth in any legal position and execute the win.
herace wrote:
We're in violent agreement…
Yes, a refreshing deviation.
herace wrote:
...though I'll quibble that your last sentence isn't exactly correct either.

In Connect 4, the first player should win. So a computer playing first should never lose. Such a program would only weakly solve the game.

"Strongly solved" for Connect 4 means that a computer program would play perfectly even when playing second. So no matter where the first player leads the game, if the first player makes a mistake (opens a winning line of player for the computer), then the computer would find that winning line of play.

The point is that if the computer is playing second, then the computer can not win unless the first player makes a mistake.
Isn't your interpretation a special case of mine? The truth of any position in the tree means, trivially, that it is one side'd turn, and, non-trivially, that the computer knows the optimal way to play either side. Playing for the losing side that would mean that it will prolong play as long as possible against perfect play, and, indeed, switch to a win if counterplay is less than perfect. At least that's how I understand it.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Herb
United States
Georgia
flag msg tools
mbmbmbmbmb
herace wrote:
"Strongly solved" means the program can find the truth in any legal position and execute the win.

I didn't write the above, you did.

This is true "Strongly solved" means the program can find the truth in any legal position

The problem is here and execute the win.

The truth may be that there isn't a winning line of play, so the program can't execute the win.

In your the last paragraph of your new post Playing for the losing side that would mean that it will postpone play as long as possible against perfect play, isn't correct either. Playing perfectly doesn't mean choosing the longest line of play. Ideally a computer would choose between "equal" lines of play randomly so that the opposing player couldn't learn a particular line of play.

So even if the computer is playing second (hence should be loosing) in Connect 4, it should try different lines of play to see if the first player can be taken into an unfamiliar game in which the first player would make a mistake.

On the flip side let's consider the computer playing second and finding a winning line of play. Perfect play would mean that the computer knows all the winning game trees and how they interconnect. So the computer could take different winning lines of play to make it harder for the first player to discover where the mistake was made that let the second player find the winning line of play.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
herace wrote:
herace wrote:
"Strongly solved" means the program can find the truth in any legal position and execute the win.

I didn't write the above, you did.
Yes, sorry, my mistake.

herace wrote:
This is true "Strongly solved" means the program can find the truth in any legal position

The problem is here and execute the win.

The truth may be that there isn't a winning line of play, so the program can't execute the win.
Apart from the question about alternative lines with an equal outcome, the last proposition puzzles me. If the position is legal and a draw is excluded, then a program which claims to have solved the game strongly should be able to execute the win. Am I missing something?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Herb
United States
Georgia
flag msg tools
mbmbmbmbmb
"Strongly solved" means the program can find the truth in any legal position and execute the win.

The above sentence implies to me that a computer program which has been programmed with the complete "strongly solved" solution could always find a winning line of play.

So if the computer is playing Connect 4, and playing second, then there is no winning line of play unless the first player makes a mistake. If the first player makes a mistake, then a perfect playing computer program would discover the mistake and exploit some winning line of play to the end of the game.

I understand what you meant which is correct. I was just quibbling with you about your sentence construction.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmb
christianF wrote:
herace wrote:
The truth may be that there isn't a winning line of play, so the program can't execute the win.
Apart from the question about alternative lines with an equal outcome, the last proposition puzzles me. If the position is legal and a draw is excluded, then a program which claims to have solved the game strongly should be able to execute the win. Am I missing something?
Should be able to execute the win, if a win is possible from that position.

E.g. consider a Hex position where black has almost connected their 2 sides and needs to place only one stone to connect & win, and there are 2 equally winning places to place their one stone to connect and win (i.e. miai). It is white's turn. A computer which can correctly analyze this position (it's white's move) nonetheless cannot execute the win (it's white's move, but white cannot win).

If it is black's turn in this position, then of course the computer can execute the win.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
russ wrote:
A computer which can correctly analyze this position (it's white's move) nonetheless cannot execute the win (it's white's move, but white cannot win).

If it is black's turn in this position, then of course the computer can execute the win.
Yes, execute the win for the winning side of course. I'm acutely aware now of the degree of clarity required.
2 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
Muemmelmann wrote:
That is why some abstract players prefer games like SET

And the likes of Backgammon with the doubling cube? By making the decision tree probabilistic and uncertain, by adding uncertainty, it is an end around, but I do see an ideal of getting as close as possible to the ideal of no draws and no side favored, so a game would have depth to it.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
docreason wrote:
... but I do see an ideal of getting as close as possible to the ideal of ...
I'm trying to work it out, hold on …
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
christianF wrote:
docreason wrote:
Now, a question I would have is: If the you can change the rules during play, in a way that is fair and balanced, is the initial conditions of not favoring a side, and also not ending a draw, possible?

I do ponder this too, because there is one noted abstract strategy game designer who insists on none of his games ending in a draw.
I fear your questions become rather exotic, although in this case I suspect you're really in the dark. "If you can change the rules during play…" - isn't there a proverb to the effect that you can't?

The weirdness of the question is obviously related to the support you seek in your ongoing debate with Voldemort. Is he awakening? surprise

(P.S. You might want to check the back of your head...)

The question isn't one of whether or not you can, although the norm is not to. The question is about doing it in such a way that the process doesn't favor one side or another, if this would result in a game that no side is favored, and also there isn't a draw. When you chop it like you did, you reach a different conclusion, and see a need for mental health treatment that isn't needed :-P. (Of course, in my case, the need for this or not would be based on other things :-P).

Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf. And I also see with Symple and Sygo, the ideas that you can change an aspect of the rules of play.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
docreason wrote:
Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf. And I also see with Symple and Sygo, the ideas that you can change an aspect of the rules of play.
Winning the right to change the rules doesn't mean that the rules can be changed during a game (unless of course that is the rule change). I wonder also to which extent the rules might be changed - how recognisable should the new rules be to still call the resulting game Gipf?

I'm not aware that any such a phenomenon exists in either Symple or Sygo. The latter had one set of rules from the onset. Symple underwent a considered change of rules in that compulsory placement was introduced some time after the game was first published, but surely you don't refer to that?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmb
docreason wrote:
The question isn't one of whether or not you can, although the norm is not to. The question is about doing it in such a way that the process doesn't favor one side or another, if this would result in a game that no side is favored, and also there isn't a draw.
You're speaking in confusing riddles...

If you mean theoretically no draws and not favoring one (idealized perfect) player, then we already know that it's mathematically impossible. You might as well be trying to prove that pi is a rational number.

If you mean practically/empirically no draws and not favoring one (mere human) player, then we already know that there exist plenty of such games, e.g. Arimaa or Hex on a sufficiently large board or Go or many others.

Quote:
Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf.
That's a pretty fast and loose characterization of GIPF with subgames. No rules change during play. A challenge is just a subgame played to determine whether a player has the right to activate a "potential" or not.

It's essentially analogous to resolving a battle in a strategic game on a side tactical map with a tactical subgame (as happens in various classic wargames, e.g. Titan or StarForce 'Alpha Centauri': Interstellar Conflict in the 25th Century). There's no rule changes going on.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
herace wrote:
docreason wrote:
On a practical level, even Connect 4 doesn't seem to work out to be a win for the first player all the time. I have done tournaments in it, and seen cases where the player going second won all 3 games. And I am not even talking about my simultaneous version of it, but the regular version.

The strategy to play Connect 4 perfectly is so complex that a human can't keep track of the whole game tree and wonders away from the "true path" of perfect play.

A sufficiently large computer can be programmed to a perfect game. See:
"A Knowledge-based Approach of Connect-Four - The Game is Solved: White Wins"
Victor Allis, 1988

http://www.connectfour.net/Files/connect4.pdf

As I remember the game is strongly solved in that if the human is playing white and makes a mistake, the computer will find the winning play sequence.

What I can say will happen, when you have a community of a sufficient number of players who communicate and study each other games, is that there is a collective wearing out of ways to play. Opening book get established, and it heads towards conformity. Without this, even a game that is solved, like Connect Four, can still end up having enough depth on a personal level, that you can keep playing it regularly.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
christianF wrote:
docreason wrote:
Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf. And I also see with Symple and Sygo, the ideas that you can change an aspect of the rules of play.
Winning the right to change the rules doesn't mean that the rules can be changed during a game (unless of course that is the rule change). I wonder also to which extent the rules might be changed - how recognisable should the new rules be to still call the resulting game Gipf?

I'm not aware that any such a phenomenon exists in either Symple or Sygo. The latter had one set of rules from the onset. Symple underwent a considered change of rules in that compulsory placement was introduced some time after the game was first published, but surely you don't refer to that?

The revolution I saw with those games was that a player had an option to change a rule in the game, that impacted the game, for strategic reasons. Pretty much player interaction ends up governing whether or not there is any rules change. And I am sorry if I might of misunderstood what was going on.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
russ wrote:
docreason wrote:
The question isn't one of whether or not you can, although the norm is not to. The question is about doing it in such a way that the process doesn't favor one side or another, if this would result in a game that no side is favored, and also there isn't a draw.
You're speaking in confusing riddles...

If you mean theoretically no draws and not favoring one (idealized perfect) player, then we already know that it's mathematically impossible. You might as well be trying to prove that pi is a rational number.

If you mean practically/empirically no draws and not favoring one (mere human) player, then we already know that there exist plenty of such games, e.g. Arimaa or Hex on a sufficiently large board or Go or many others.

Quote:
Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf.
That's a pretty fast and loose characterization of GIPF with subgames. No rules change during play. A challenge is just a subgame played to determine whether a player has the right to activate a "potential" or not.

It's essentially analogous to resolving a battle in a strategic game on a side tactical map with a tactical subgame (as happens in various classic wargames, e.g. Titan or StarForce 'Alpha Centauri': Interstellar Conflict in the 25th Century). There's no rule changes going on.


The potential is a new piece added to the game, that a player decides whether or not to add to the game or not, and players play a subgame to decide this. The new potential is a new rule added to the game. It impacts how the game works, thus a new rule added.

In regards to this, what I see by having this happen is that the decision tree gets made a mess and hard to do, which is a good thing actually.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
christian freeling
Netherlands
flag msg tools
designer
Avatar
mbmb
docreason wrote:
christianF wrote:
docreason wrote:
Heck, the premise of the Gipf Series is that you play another game to win the right to change the rules in the game Gipf. And I also see with Symple and Sygo, the ideas that you can change an aspect of the rules of play.
Winning the right to change the rules doesn't mean that the rules can be changed during a game (unless of course that is the rule change). I wonder also to which extent the rules might be changed - how recognisable should the new rules be to still call the resulting game Gipf?

I'm not aware that any such a phenomenon exists in either Symple or Sygo. The latter had one set of rules from the onset. Symple underwent a considered change of rules in that compulsory placement was introduced some time after the game was first published, but surely you don't refer to that?

The revolution I saw with those games was that a player had an option to change a rule in the game, that impacted the game, for strategic reasons. Pretty much player interaction ends up governing whether or not there is any rules change. And I am sorry if I might of misunderstood what was going on.
The Symple move protocol and its embedded balancing mechanism is revolutionary indeed in that it is generic. Both Symple and Sygo use it, and there's also a 'disconnection' game called Multiplicity in the (applet) pipeline that naturally and seamlessly employs it.

But none of my games has the option to change a rule during play (I feel slightly ridiculous even mentioning that). I don't know where you got the idea, but not, I fear, by ever playing either Symple or Sygo or even contemplating the move protocol. I've encountered such notions that seem unhindered by any actual knowledge before, but I fear you're misguided. Not that I blame the source, because he's nuts. But you're there with your eyes open, aren't you?
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmb
docreason wrote:
The potential is a new piece added to the game, that a player decides whether or not to add to the game or not, and players play a subgame to decide this. The new potential is a new rule added to the game. It impacts how the game works, thus a new rule added.
When you promote a rook to a dragon in Shogi, you're adding a new piece to the game.

When you promote an ordinary piece to a king in Checkers, you're adding a new piece to the game.

But a new rule was not added. The rules were there all along, and known all along. No rules changed. The rules for Shogi and Checkers and GIPF with potentials are printed in their rulebooks.
5 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
russ wrote:
docreason wrote:
The potential is a new piece added to the game, that a player decides whether or not to add to the game or not, and players play a subgame to decide this. The new potential is a new rule added to the game. It impacts how the game works, thus a new rule added.
When you promote a rook to a dragon in Shogi, you're adding a new piece to the game.

When you promote an ordinary piece to a king in Checkers, you're adding a new piece to the game.

But a new rule was not added. The rules were there all along, and known all along. No rules changed. The rules for Shogi and Checkers and GIPF with potentials are printed in their rulebooks.

Thing is that whether or not the new piece can be added or not, depends on how the other game is played out. The rules change in the game under Gipf, which is different than you see in Shogi or promotion of a piece in chess. Unless Shogi, there is a player controlled condition that determines if a new piece may or may not be added to a game.

The subgame structure with Gipf series opens up other things than something fixed. In short, the rules are not set with the game. One could then also agree to add other pieces to the game of Gipf by adding different rules.

This approach is far more akin to a collectible card game where new components not in the game originally can be added later.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Russ Williams
Poland
Wrocław
Dolny Śląsk
flag msg tools
designer
badge
Avatar
mbmbmb
docreason wrote:
russ wrote:
docreason wrote:
The potential is a new piece added to the game, that a player decides whether or not to add to the game or not, and players play a subgame to decide this. The new potential is a new rule added to the game. It impacts how the game works, thus a new rule added.
When you promote a rook to a dragon in Shogi, you're adding a new piece to the game.

When you promote an ordinary piece to a king in Checkers, you're adding a new piece to the game.

But a new rule was not added. The rules were there all along, and known all along. No rules changed. The rules for Shogi and Checkers and GIPF with potentials are printed in their rulebooks.

Thing is that whether or not the new piece can be added or not, depends on how the other game is played out. The rules change in the game under Gipf, which is different than you see in Shogi or promotion of a piece in chess. Unless Shogi, there is a player controlled condition that determines if a new piece may or may not be added to a game.

The subgame structure with Gipf series opens up other things than something fixed. In short, the rules are not set with the game. One could then also agree to add other pieces to the game of Gipf by adding different rules.

This approach is far more akin to a collectible card game where new components not in the game originally can be added later.
I sincerely believe you're somehow fooling yourself with semantics or with the different emotional feels of the different games (i.e. it is a cool unusual mechanism that in GIPF+Potentials you play a separate subgame on a separate board). But GIPF+Potentials is still a well-defined game. No rules are being changed during play! The fact that players play out a subgame to see if a potential is activated is not changing the rules of the game - that subgame happens EXACTLY BECAUSE the rules of the game define that the subgame happens.

If you really think some rule is being changed during play, then I request that you to tell me specifically which rule of GIPF+Potentials is being changed during play. I've played GIPF+Potentials; I am certain that we did not change rules during play. The burden of proof is on you, as you're making an extraordinary claim.


A subgame being used to resolve something in the main game is not changing rules of the game. It has an effect on the main game. Just like playing out a tactical battle as a subgame in Titan or StarForce 'Alpha Centauri': Interstellar Conflict in the 25th Century has an effect on the main game, but is in no way changing the rules during play.

The printed rules remain the same and continue to be in effect.

===

Look, suppose you played the cliched Chess variant (let's call it "Random Chess") where instead of automatically capturing a piece, you instead flipped a coin, or played Rock-Scissors-Paper, to see which piece survived. Does that mean that the rules of the game Random Chess are being changed during play, merely because you play out a subgame (albeit a quick trivial random one) to see if you your capture move succeeds or not? Of course not.

===

Or maybe you're mixing up the idea of changing rules during play with the idea of playing a variant (i.e. playing GIPF+Potentials instead of just playing regular GIPF - agreeing beforehand that additional rules concerning potentials and subgames are in effect). Playing GIPF+Potentials is playing a variant whose rules are already established before the first move; those rules are different from regular GIPF, but that "change" is made before the game starts, not during play.

It would be like asserting that playing Chess960 is somehow changing the rules during play... when it's simply agreeing to change the standard Chess rules to other rules (Chess960) before play begins.
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Richard Hutnik
United States
Albany
New York
flag msg tools
designer
Avatar
mbmbmbmbmb
Gipf Series was an example of what I was talking about. The idea is changing the environment of the game, by what is not present, by playing a subgame to see if such a change gets added. Rather than argue over whether or not a new piece coming into game is considered a new rule, let's make the change an actual change to the rules during play, that wasn't in the game originally. Such a rules change happens by playing a subgame to win the right to do so. In this case, is it possible doing this to end up with a game where neither side has a theoretical advantage and also there are no draws?

In short, theoretically, it it possible to have a pure abstract strategy game where no side has an advantage, and also no draws, in the cases where the rule set is infinite and not finite?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
[1]  Prev «  1 , 2 , 3 , 4 , 5  Next »  [6] |