Author |
Topic: Rules Discussion (Read 1252 times) |
|
The_Jeh
Forum Guru
Arimaa player #634
Gender: 
Posts: 460
|
 |
Rules Discussion
« on: Jan 31st, 2008, 11:58pm » |
 Quote  Modify
|
I love discussing rule particularities, so I am going to do so.
on Jan 30th, 2008, 10:47am, Fritzlein wrote:Even with the rabbit extermination rule, there is a possibility the players will shuffle pieces endlessly without accomplishing anything. Technically the repetition rule will end such games eventually, so they aren't drawn by rule, but in practice the sun might burn out before the game ends. The practical solution for live games, at least at present, is to have a time cutoff after which the ArimaaScore formula determines the winner. But what if the scores of the two players are tied? Omar has foreseen even this infinitesimal possibility, and ruled that Silver wins if the score is tied after time cutoff.
At one time we feared that optimal play would produce indefinite piece shuffling, but now it seems that optimal play will either pull opposing rabbits or advance friendly rabbits voluntarily, so there is little chance of drawn out human games. Bots are another matter. They are still dumb enough to potentially get caught playing aimlessly forever. Thus Claude needs to have some move cutoff built into his testing program.
I think what Claude's latest post is saying is that the bots he is testing now are so dumb and aggressive (not smart and defensive) that some bot will lose all its rabbits before the cutoff comes into play. |
|
Since time (and the scoring system, in my opinion) is arbitrary, and won't work, for example, in casual games with no maximum game time enforced, what would be your opinion of a rule that states, beginning the count on turn 100w, the game immediately ends if there have been 101 turns (50.5 moves) with no captures, and is lost by the player who just played? I make it an odd number of turns so that if one captures a piece or advances a rabbit, the burden of making the next capture rests with the opposite player. With a rabbit extermination rule in place, there will always be capturable material available. The first burden rests with gold, similar to the already-in-place rule that silver wins if the score is tied. If gold is up in material when the count begins, he has less of an excuse for not capturing or winning, especially since he has a whole 50.5 moves with which to work. Besides, he can suicide if desperate.
And, if you want to eliminate all draws, here would be my proposal for the standard turn sequence:
1. Check immobilization
If you have no legal move, the game immediately ends, and your opponent wins.
2. Make move
3. Check your own goal
If you have a rabbit in your goal, the game immediately ends (without proceeding to step 4), and you win.
4. Check opponent's goal
If your opponent has a rabbit in his or her goal, the game immediately ends, and your opponent wins.
5. Check opponent's rabbits.
If your opponent has no rabbit left, the game immediately ends, and you win.
6. Check Threefold Repetition Rule
If you have just caused the same position with same player to move to occur for the third time, the game immediately ends, and your opponent wins.
7. Check 101-turn rule
If you have just made the 101st consecutive turn with no captures (and it was made on or after 150w), the game immediately ends, and your opponent wins.
|
| « Last Edit: Feb 1st, 2008, 12:42am by The_Jeh » |
 IP Logged |
|
|
|
arimaa_master
Forum Guru
Arimaa player #2010
Gender:
Posts: 358
|
|
Re: Rules Discussion
« Reply #1 on: Feb 1st, 2008, 2:35am » |
Quote Modify
|
on Jan 31st, 2008, 11:58pm, The_Jeh wrote:I love discussing rule particularities, so I am going to do so.
Since time (and the scoring system, in my opinion) is arbitrary, and won't work, for example, in casual games with no maximum game time enforced, what would be your opinion of a rule that states, beginning the count on turn 100w, the game immediately ends if there have been 101 turns (50.5 moves) with no captures, and is lost by the player who just played? I make it an odd number of turns so that if one captures a piece or advances a rabbit, the burden of making the next capture rests with the opposite player. With a rabbit extermination rule in place, there will always be capturable material available. The first burden rests with gold, similar to the already-in-place rule that silver wins if the score is tied. If gold is up in material when the count begins, he has less of an excuse for not capturing or winning, especially since he has a whole 50.5 moves with which to work. Besides, he can suicide if desperate.
And, if you want to eliminate all draws, here would be my proposal for the standard turn sequence:
1. Check immobilization
If you have no legal move, the game immediately ends, and your opponent wins.
2. Make move
3. Check your own goal
If you have a rabbit in your goal, the game immediately ends (without proceeding to step 4), and you win.
4. Check opponent's goal
If your opponent has a rabbit in his or her goal, the game immediately ends, and your opponent wins.
5. Check opponent's rabbits.
If your opponent has no rabbit left, the game immediately ends, and you win.
6. Check Threefold Repetition Rule
If you have just caused the same position with same player to move to occur for the third time, the game immediately ends, and your opponent wins.
7. Check 101-turn rule
If you have just made the 101st consecutive turn with no captures (and it was made on or after 150w), the game immediately ends, and your opponent wins. |
|
It seems fine to me.
|
|
IP Logged |
|
|
|
99of9
Forum Guru
Gnobby's creator (player #314)
 
Gender:
Posts: 1413
|
|
Re: Rules Discussion
« Reply #2 on: Feb 1st, 2008, 3:30am » |
Quote Modify
|
Given that every suicide extends the game by 50 moves, that could lead to an unplayably long game. I'd suggest reducing the number, but in principle I like the idea that pieces will start to thin out by suicide, making the chance of a goal more and more possible.
|
|
IP Logged |
|
|
|
clauchau
Forum Guru
bot Quantum Leapfrog's father
Gender:
Posts: 145
|
|
Re: Rules Discussion
« Reply #3 on: Feb 1st, 2008, 5:44am » |
Quote Modify
|
It's good to have another subgoal that surfaces when the position doesn't seem to evolve otherwise. Now, I would'nt like to rely on an exact count of moves. I'd rather have subgoals in the same spirit as others, for example you may also win if one of your piece has completed a circuit around a stronger piece of your opponent's, or if more of your pieces are next to traps than your opponent's, or ...
I also would add :
5b. Check your own rabbits.
If you have no rabbit left, the game immediately ends, and your opponent win.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: Rules Discussion
« Reply #4 on: Feb 1st, 2008, 9:11am » |
Quote Modify
|
I like the idea of placing the burden of progress on the players alternately. The limit of 50 moves to make a capture seems fair and is guaranteed to end the game on a human time scale rather than an astronomical one. It would be quite reasonable for bringing some theoretical results within reach. Unfortunately, it still fails a test of practicality for live games.
For tournament games, we will likely calibrate our time control based on the expected length of a game that is fought to the death, say 45 moves on average. Any game that goes more than twice the expected length is likely to be a problem due to exhaustion or other time commitments. A practical tournament rule to decide games than aren't progressing should be guaranteed end games at 100 moves, not hundreds.
Of course, there have been games lasting more than 100 moves which were not stuck and which did end over the board. This makes a game-ending rule for live tournaments a somewhat intractable problem in my mind.
I guess my favorite solution is one originally proposed by Mr. Brain: In addition to all the other time constraints, the total time allotted for the game is divided between the two players. If the game is supposed to end in three hours, each player has 90 minutes for all their moves total, and the first to overstep the time loses. This enforces a mad blitz scramble at the end of drawn-out games, so if the game would naturally have taken 120 moves to complete, the players will force in all the moves in to finish within the time control, or they will blunder and end it sooner.
The sudden-death time rule has some drawbacks: the quality of moves gets lower towards the end, players with slow Internet connections are punished, some games will end on time rather than over the board, etc. However, it seems more "natural" than other proposals. The time cutoff is indeed arbitrary, but no more arbitrary than a 50-move rule (why not a 37-move rule?), and definitely less arbitrary than the ArimaaScore formula, which is in the rules but thankfully has never decided a tournament game.
|
|
IP Logged |
|
|
|
woh
Forum Guru
Arimaa player #2128
Gender:
Posts: 254
|
|
Re: Rules Discussion
« Reply #5 on: Feb 1st, 2008, 11:01am » |
Quote Modify
|
An other suggestion:
after his 1st - 60th move the player's most advanced rabbit must have reached at least the first row from his side of the board (this is always true)
after his 61st - 120th move the player's most advanced rabbit must have reached at least the second row from his side of the board
after his 121st - 180th move the player's most advanced rabbit must have reached at least the third row from his side of the board
end so on
after his 421st - 480th move the player's most advanced rabbit must have reached at least the 8th row from his side of the board
meaning that no game could last beyond move 421w
The idea is this:
the goal of the game is to reach the other side of the board and one has to advance a least one row every 60 turns
This would also mean: if one's most advanced rabbit is pushed back, it must advance (or an other rabbit) on the next move or the game is lost. The same holds when the most advanced rabbit is captured.
|
| « Last Edit: Feb 1st, 2008, 11:03am by woh » |
IP Logged |
|
|
|
The_Jeh
Forum Guru
Arimaa player #634
Gender:
Posts: 460
|
|
Re: Rules Discussion
« Reply #6 on: Feb 1st, 2008, 12:02pm » |
Quote Modify
|
on Feb 1st, 2008, 9:11am, Fritzlein wrote:
The sudden-death time rule has some drawbacks: the quality of moves gets lower towards the end, players with slow Internet connections are punished, some games will end on time rather than over the board, etc. However, it seems more "natural" than other proposals. The time cutoff is indeed arbitrary, but no more arbitrary than a 50-move rule (why not a 37-move rule?), and definitely less arbitrary than the ArimaaScore formula, which is in the rules but thankfully has never decided a tournament game. |
|
If it is defined as an intrinsic rule, the 101-turn rule would be no more arbitrary than the threefold repetition rule, or that rabbits cannot move backwards, or every other rule invented by Omar. Time, and the score because it depends on time, can vary from game to game, so they are arbitrary in the sense that they can be chosen on a game-to-game basis. Time is an element of the players' world, while the 101-turn rule would be part of the game's world. I am not saying we should get rid of time; on the contrary, I think allotting each player a finite amount of game time is quite sensible for match play. But adding a loose intrinsic progression-forcer could not hurt general games.
Woh, your idea is very interesting. I'll have to ponder it a while.
|
| « Last Edit: Feb 1st, 2008, 12:16pm by The_Jeh » |
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: Rules Discussion
« Reply #7 on: Feb 1st, 2008, 3:24pm » |
Quote Modify
|
on Feb 1st, 2008, 12:02pm, The_Jeh wrote:| If it is defined as an intrinsic rule |
|
I don't see what this clause clarifies. A sudden-death time control could also be defined as an "intrinsic rule" by making time part of the game state. Then time would be part of the game's world. (Unless you define intrinsic properties of the game as precisely those that don't change with time, which is sort of a circular way to argue. Then saying time isn't intrinsic to the game is no more than saying that time is time.)
I don't think the intrinsic/extrinsic vocabulary contributes much to the discussion, since all proposals are currently extrinsic to Arimaa as now defined and all could become intrinsic to Arimaa if added.
Quote:| the 101-turn rule would be no more arbitrary than the threefold repetition rule, or that rabbits cannot move backwards, or every other rule invented by Omar. |
|
Well, perhaps arguing about which rule is "arbitrary" doesn't contribute to the discussion either. Yes, all the rules are arbitrary. We could change any rule of Arimaa and still have a game, albeit a different game than before.
I think the starting point of the discussion is that Arimaa is a good game as it stands, and whatever rule we add to fix a perceived flaw should have as minimal an impact as possible. We don't want to warp the game with any rule change, and require players to play differently than they do now, except in extreme situations. Both your rule proposal and woh's rule proposal would indeed have a minimal impact.
But let's back up. What is the perceived flaw? We are not trying to fix the problem that Arimaa can be drawn out forever with no termination. The repetition rule already fixes that. Arimaa is already a drawless game, once rabbit-elimination is a loss. A rule that forces captures or forces rabbit advancement is not necessary to eliminate draws.
What is necessary is a way to terminate games within a humane time period. Without such a rule, live tournaments could become impossible or impractical. Neither the proposed capture rule nor the proposed rabbit-advancement rule meets this requirement. A game that might go twenty hours will be likely decided by exhaustion or other commitments long before the on-board conditions are met.
Let's compare solutions that guarantee either that the game ends, say, by move 120, or guarantees that the game ends in four hours. When this need is met, let's see which means of meeting it has the least impact on the course of the game. So far the sudden-death time control is the lowest-impact solution I have seen discussed. But perhaps some accelerated version of the capture rule or rabbit-advancement rule would have a lesser impact on the ordinary course of the game than a time scramble would, and would still guarantee a short enough cutoff to have a practical use.
|
| « Last Edit: Feb 1st, 2008, 3:36pm by Fritzlein » |
IP Logged |
|
|
|
The_Jeh
Forum Guru
Arimaa player #634
Gender:
Posts: 460
|
|
Re: Rules Discussion
« Reply #8 on: Feb 1st, 2008, 4:57pm » |
Quote Modify
|
By intrinsic, I was referring to rules that affect the actual gametree. Extrinsic rules I define as not affecting the gametree. So, the goal was to change the gametree such that games terminate within a reasonable number of moves, while still keeping the bulk of the game the same as we all have come to know and love. That is all I meant.
What if, during moves 101-105, we define both the 7th and 8th ranks as the goal ranks, 106-110 the 6th rank and beyond, 111-115 the 5th and beyond, etc. We apply the standard turn sequence. So if both gold and silver have a rabbit in the 7th rank at the beginning of move 101w, gold wins if he still has a rabbit in the 7th rank at the end of the turn. Or, if you prefer silver, we can begin with move 101b.
|
| « Last Edit: Feb 1st, 2008, 5:11pm by The_Jeh » |
IP Logged |
|
|
|
Janzert
Forum Guru
Arimaa player #247
Gender:
Posts: 1016
|
|
Re: Rules Discussion
« Reply #9 on: Feb 1st, 2008, 5:25pm » |
Quote Modify
|
I like that adding an absolute time limit for each player, directly addresses the problem (longest possible time for the game), doesn't change the game mechanics (from the current rules or within a single game as it progresses*), is inherently flexible (e.g. a blitz tournament could still have a longer absolute time if desired for long games to play out) and is fair to both players (the current time cutoff could conceivably be reached by one player squandering time while the other is playing quickly).
Janzert
* To put it in The_Jeh's intrinsic-extrinsic terms, in this case at least I would prefer an extrinsic solution to this extrinsic problem.
|
|
IP Logged |
|
|
|
Fritzlein
Forum Guru
Arimaa player #706
Gender:
Posts: 5928
|
|
Re: Rules Discussion
« Reply #10 on: Feb 1st, 2008, 5:27pm » |
Quote Modify
|
Intuitively I like moving the goal line nearer better than requiring rabbit advances or requiring captures. Rather than imposing a loss for failure to act, it makes a win available with easier and easier action. Also it feels very similar to the original and primary objective.
To follow up Janzert's reasoning, the solution of changing the goal is appropriate if the goal is too hard to reach regardless of time control, but a time-based solution for a problem that is essentially about time seems appropriate.
Fortunately, the number of moves to goal in Arimaa games does not appear to be stretching longer and longer as we get better at playing. Games at all levels seem to move forward rather than stagnating. We are lucky to be discussing a problem that exists only in theory so far.
|
| « Last Edit: Feb 1st, 2008, 5:34pm by Fritzlein » |
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print