Arimaa Forum - Print Page


    
      
        Arimaa Forum
        (http://arimaa.com/arimaa/forum/cgi/YaBB.cgi)
      

        Arimaa >> Site Discussion >> Option to hurry up the clock?
        
(Message started by: Boo on Aug 28^th, 2012, 3:54am)

Title: Option to hurry up the clock?
Post by Boo on Aug 28^th, 2012, 3:54am

In one of my postal games I have accumulated time reserve of ~100days, which is almost equal to the game time remaining. However, if I want to benefit from this situation, I have to wait for 20 days for 5 times (because of 21d move limit). I find it somewhat annoying making a decision in a couple of minutes and then being forced to sit for 3 weeks.
Would it be possible to have an option to decrease my time reserve by 20 days also decreasing the game time remaining by the same amount? It could significantly speed up the game. Winning the game by not playing is no fun at all.

Title: Re: Option to hurry up the clock?
Post by rbarreira on Aug 28^th, 2012, 6:32am

IMHO, if there's anything that needs fixing here it's the fact that the time control typically used on postal games has a game time limit which is not that big (only guaranteeing 90 moves).

To me, adding an option to make it more convenient to exploit this flaw in the time control seems like the wrong way to go.

PS: When I created the cyborg_briareus games I deliberately chose no game time limit at all.

Title: Re: Option to hurry up the clock?
Post by Fritzlein on Aug 28^th, 2012, 1:16pm

Yes, a human vs. human game should never reach the maximum game time unless the board position is stalemated. If the game stalemates, then Arimaa is flawed. If the game doesn't stalemate but the total game time is reached, then the time control is flawed.

I used to object to the Postal Mixer time control every year because I expected this would happen, but year after year it never happened, so I stopped objecting. Now that there is a real case of someone winning by running out the clock, it may be time to rethink the time controls after all.

Title: Re: Option to hurry up the clock?
Post by browni3141 on Aug 28^th, 2012, 2:35pm

on 08/28/12 at 13:16:10, Fritzlein wrote:

If the game stalemates, then Arimaa is flawed.

I disagree. We already know that stalemates are possible, so I'm not 100% sure what you mean. I guess you mean if we got good enough that stalemates became very frequent? I guess if that's the case then I must agree, but I doubt we will ever reach that level of play. The "drawing" margin of Arimaa is likely much smaller than chess, since "draws" almost never occur in real play between humans. Even if perfect play results in a stalemate, we will never be good enough to achieve it, so I don't think it makes Arimaa flawed.

Title: Re: Option to hurry up the clock?
Post by ocmiente on Aug 28^th, 2012, 2:48pm

on 08/28/12 at 13:16:10, Fritzlein wrote:

Yes, a human vs. human game should never reach the maximum game time unless the board position is stalemated.

This particular game is not stalemated.

Title: Re: Option to hurry up the clock?
Post by Fritzlein on Aug 28^th, 2012, 5:50pm

on 08/28/12 at 14:35:19, browni3141 wrote:

We already know that stalemates are possible, so I'm not 100% sure what you mean.

I was conflating two ideas, so it isn't surprising that I confused you.
(1) I prefer a drawless game. I would happiest if stalemated positions would never arise in practice. If they do arise in practice, it is a flaw in Arimaa (IMHO) in proportion to how often they arise. It's not the worst flaw a game can have, but it is important enough that I hope it never happens in Arimaa. I know, however, that many people don't mind draws, and therefore disagree with me about this point.
(2) There is a question of how to handle games that can't be decided on the board in a reasonable amount of time. What should we do when a practical draw occurs? The repetition rule might take millions of moves to kick in, so that doesn't necessarily prevent all draws in practice. There are many options. The option currently in place is a time cutoff and a winner decided by piece count. I don't want any games to be decided this way. This rule is only for emergencies, because we can't let games go on forever, but if the time rule ever gets invoked we want to figure out what went wrong and prevent it from happening again.

Possibilities to prevent it from happening again include:
(A) Changing the time control
(B) Changing the formula for who wins when time cutoff is reached
(C) Declaring a draw if time cutoff is reached
and many more ideas...

In this case it seems like (A) is by far the most appropriate answer. I don't think we need to start allowing draws in Arimaa, and I don't have a better formula for what to do at time cutoff. But the position isn't a true stalemate. It isn't a practical draw; there just haven't been enough moves permitted to play to a natural conclusion.

By the way, if the game ends with 9 pieces on each side, ocmiente wins, because he most recently had 10 pieces. I can only assume that Boo made a capture in the last two moves that I can't see due to spectator delay, or else Boo would have no incentive to run out the clock.

Title: Re: Option to hurry up the clock?
Post by Fritzlein on Aug 28^th, 2012, 6:03pm

on 08/28/12 at 03:54:37, Boo wrote:

Would it be possible to have an option to decrease my time reserve by 20 days also decreasing the game time remaining by the same amount? It could significantly speed up the game.

You can guess from my post that I don't want this feature to be implemented, because I hope it will never be used. I don't want any games to be decided by running out total game time. Of course, I don't blame you if you win this way. You should be allowed to win any way within the rules. It is the rules that are the problem. The solution is not to add a new client feature, but to change the rules instead so that no one will ever need this client feature.

Quote:

Winning the game by not playing is no fun at all.

I agree. It is no fun to win by not playing. Fortunately, if you want to have fun you have a different option: win by playing! ;D

Title: Re: Option to hurry up the clock?
Post by ocmiente on Aug 28^th, 2012, 7:29pm

on 08/28/12 at 17:50:53, Fritzlein wrote:

By the way, if the game ends with 9 pieces on each side, ocmiente wins, because he most recently had 10 pieces. I can only assume that Boo made a capture in the last two moves that I can't see due to spectator delay, or else Boo would have no incentive to run out the clock.

That's funny. Before I read this (and confirmed that you are correct by reading the rules) I thought that Boo would have won the game if time ran out with the game state as it is at move 89s (which is what spectators can see now - no spoiler here).

Title: Re: Option to hurry up the clock?
Post by browni3141 on Aug 28^th, 2012, 8:26pm

You guys do realize that the current position is not hidden, right? We don't need to see the board position to know what it is. Perhaps you should have also covered up the move list.

Title: Re: Option to hurry up the clock?
Post by Boo on Aug 29^th, 2012, 1:58am

Quote:

Yes, I knew that. Actually I thnik ocmiente cannot avoid captures for 5 moves, so it is a good spot to start slowplaying. And if he does hold everything, it is OK also. Having 1day to 40days is still a big handicap as I can choose to end the game in 2 moves or in 40 moves.

Quote:

If the game doesn't stalemate but the total game time is reached, then the time control is flawed.

I don't think time control is flawed. It is a rule that adds another aspect to the arimaa strategy - time management. Giving a time reserve handicap of 1 day to 100 days is like putting yourself in an awkward position - opponent can choose when to end the game, and you can not. So you should either be ahead materially or play fast.
Why is it flawed? It is a rule like many others. You don't say that rules are flawed, when somebody wins by advancing rabbit to the 8th rank.

Quote:

Possibilities to prevent it from happening again include: <...>

D) Remove time limit and add a rule like "If there is no rabbit move forward step for 25(?) moves by each side, the game is decided by material advantage." instead.
However I am not sure if it is a good idea to remove time limit. Consider playing a live WAC 2min game with 8h time limit. Now you at least know maximum amount how long the game can last. If the rule gets removed - are you ready to play 24h non-stop?

Quote:

I used to object to the Postal Mixer time control every year because I expected this would happen, but year after year it never happened, so I stopped objecting. Now that there is a real case of someone winning by running out the clock, it may be time to rethink the time controls after all.

I don't think it should be waited until something happens in practice. If it can happen theoretically, it will happen practically eventually. Sooner or later. It is better to discuss it immediately when a theoretical possibility is found.

Quote:

You guys do realize that the current position is not hidden, right? We don't need to see the board position to know what it is. Perhaps you should have also covered up the move list.

I didn't know postal games are viewable by others.

Title: Re: Option to hurry up the clock?
Post by ocmiente on Aug 29^th, 2012, 9:04am

on 08/29/12 at 01:58:34, Boo wrote:

I didn't know postal games are viewable by others.

Non-event postal games can be viewed by others with no restrictions. Event postal games (like this one) can be viewed, but spectators cannot see the current move. They see the game as it was 2 moves back. This game is now on move 91s, but spectators see it at move 89s.

The problem that browni3141 is referring to is that you posted an image in the original post of this thread that includes a move list. The current game position can be constructed using the position at 89s plus this move list. Please edit your original post to remove that image or edit the image so that the move list is not visible.

For those who are not aware of the delay, it was implemented so that spectators cannot comment on the current game state. It is especially useful for event games with a faster time control (which are time delayed, rather than move delayed) so that people can chat about the game in teamspeak and the chatroom without worrying about players listening.

Title: Re: Option to hurry up the clock?
Post by Boo on Aug 29^th, 2012, 1:02pm

Quote:

Please edit your original post to remove that image

Removed.

Title: Re: Option to hurry up the clock?
Post by Fritzlein on Aug 29^th, 2012, 1:38pm

on 08/29/12 at 01:58:34, Boo wrote:

I don't put all rules on an equal footing. It is necessary to have a time control to keep the game fair, since it is easier to come up with a good move given more time, but ordinarily time controls don't change which move is best.

Admittedly, time controls will always have some influence on optimal play. For example, there may be a strong defensive move which is a forced win in 52 turns according to game theory, but in a game at 15 seconds/move, it may create a higher a winning percentage to play an attacking move which is a forced loss in 37 turns, given that in practice it is harder to defend than attack when under time pressure. Thus the optimal move can be changed by any time control. But normally, for every time control in common use, a strong play is a strong play. That is as it should be. I like the time control to affect optimal play only very slightly.

To have a total game time cutoff can drastically change the correct move. For example, suppose that from an equal position I can capture a cat this turn in exchange for losing my camel in two turns This is normally a terrible exchange, but I might know that the game is about to end on time, so my opponent will only get the chance to play one more move before I run out the clock. Then a horrible move could become a winning move.

You could argue that every rule is of equal importance to every other rule, and that each given time control creates an Arimaa variant with its own strategy and its own winning condition. I don't like Arimaa variants, though, and I don't want to go about creating a new variant with each new time control. I don't like changing optimal play. I know that it is unavoidable to a small extent; time controls are a necessary evil. But I would like to minimize the impact of time control on optimal play.

Why do we like the current time control of the Postal Mixer so much that we are willing to let it distort optimal play relative to, say, a total game time of 360 days instead of 300? It is a question of priorities. We should gain something very important if it creates a high probability that ordinarily great moves become terrible and vice versa.

I believe that we can come up with a time control that meets the needs of fairness and meets the needs to the Postal Mixer to end in finite time but which has a much lower probability of creating a distorted ending than the current time control.

Title: Re: Option to hurry up the clock?
Post by rbarreira on Aug 29^th, 2012, 1:55pm

Indeed not all rules are created equal, and this is goes for other games too.

In chess no one would think about changing the way checkmate works, yet the 50-move rule has been controversial and in the recent past got changed back and forth a few times. Actually that rule has strong similarities with a game time limit...

Title: Re: Option to hurry up the clock?
Post by PatriciaLucas on Apr 17^th, 2014, 8:10am

I am agree with you #BOO and there is an other fact, that is every game have its own functionality and perception which make it differ from others.