Welcome, Guest. Please Login or Register.
Mar 28th, 2024, 9:23am

Home Home Help Help Search Search Members Members Login Login Register Register
Arimaa Forum « Essay by Christian Freeling on inventing games »


   Arimaa Forum
   Arimaa
   Off Topic Discussion
(Moderators: christianF, supersamu)
   Essay by Christian Freeling on inventing games
« Previous topic | Next topic »
Pages: 1 ... 53 54 55 56 57  ...  78 Notify of replies Notify of replies Send Topic Send Topic Print Print
   Author  Topic: Essay by Christian Freeling on inventing games  (Read 520586 times)
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
Re: Essay by Christian Freeling on inventing games
« Reply #810 on: Dec 25th, 2011, 2:20pm »

10.
on Dec 24th, 2011, 7:57am, christianF wrote:
6. Komi

This is a kind of pecuniary compensation for moving second in Go, or similar games based on acquiring points, blunt, practical and completely lacking in style. In Go it is employed for lack of an alternative. Thinking about that, I figured a 3-stone pie might work, say two blacks in an Orang Utan formation against a properly placed white, but I can't be the first to have suggested that or a dozen other alternatives for that matter, and all rejected no doubt.
 
It works, games get decided. Is it fair? For starters, to get anywhere near 'fair', the value should result from massive amounts of results between more or less evenly matched high level players. But even then, it is arbitrary. As the most natural thing in the world Wiki states "To prevent ties, komi always includes half a point. In Japan the standard Komi is 5.5 points, though in the U.S.A it is 6.5."
Who could imagine a draw between evenly matched grandmasters? It's so much better to have a situation where the same game is a win in the USA and a loss in Japan. This is a bookkeepers rule in a world where bookkeepers rule and where draws are ruled out, on practical grounds by tournament organizers, on cultural grounds by the Japanese and on traumatic grounds by the founders of the Church of Hard Finitude.
« Last Edit: Dec 25th, 2011, 3:06pm by christianF » IP Logged
MarkSteere
Forum Guru
*****




Finite games rule

   
WWW

Gender: male
Posts: 289
Re: Essay by Christian Freeling on inventing games
« Reply #811 on: Dec 25th, 2011, 3:46pm »

on Dec 25th, 2011, 2:20pm, christianF wrote:

the Church of Hard Finitude.

The Church of Hard Finitude gave us Oust.
 
-Leonardo
IP Logged
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
Re: Essay by Christian Freeling on inventing games
« Reply #812 on: Dec 26th, 2011, 10:56am »

Occam's Razor and Shogi wisdom
There's this Shogi proverb, "If you find a good move, look for a better one". The same holds for games. The wisdom of the advice is quite transparent, yet one tends to forget: good moves are tempting, and so are good games.
There's also Occam's Razor: never make unnecessary assumptions. As a game inventor I'd put that in N E O N.
 
By now it should be clear that in the case of Symple, both Benedikt and I forgot the first and violated against the second. The hunt was for the quintessential 'group penalty' game. The force behind it was the existence of several connection games that employed 'group penalty', but always based on awarding points to groups connected to the edges in certain ways, never on the size of the group itself. This smelled of unnecessary assumptions and eventually led to the birth of Symple, and everybody was happy except Mark Steere. The game was quintessential - no unnecessay assumptions - and the hunt was over.
 
This was sheer complacency, fueled by the fact that there didn't seem to be anything wrong with it, except ... a certain lack of drama. As J. Mark Thompson points out in his Defining the Abstact: "It should be possible for a player to recover from a weaker position and still win the game. Victory should not be achievable in a single successful blow; the suspense should continue through an extended campaign. Otherwise an early disadvantage makes the remainder of the game uninteresting".
"Like a big ship slowly heeling to one side", as I put it myself, characterizing Symple as 'interesting' and 'significant', all within the very relative importance of abstract games of course, but not 'great'.
At the same time I considered it a self explanatory game, so there's not much one can do about it: quintessential games, for better or worse. "are what they are". Therefore I didn't consider it a problem, let alone contemplate a fix.
 
'Group penalty' is an elusive theme with a terrible name. In its previous incarnations - *Star et al - it was a mix of static and dynamic connection, static in the sense of connecting to the edges, dynamic in the connecting of groups. Symple, instead of counting the number of edge-cells a group touches, simply counts the size of the group. Thus it replaces the static connection theme by 'territory' (i.e. groupsize), while the dynamic connection theme remains intact.
 
The trap
The trap we fell into was this: we considered 'territory' in a Go-like perspective because 'shape' was a key concept: surround vacant points in such a way that no advantageous invasion is possible. If you can no longer add to the score, pass. In the end, after successive passes, completely surrounded territory is yours. No need to fill it in while the other player keeps passing. The game was devoid of any zugzwang.
In retrospect, that was the source of its lack of drama: if you could accumulate small advantages it was often possible to hold on to them because the tactical means to turn the tables were quite modest. So games would peter out in small strategic victories.
 
Then came Luis with his question regarding compulsory movement, and though it referred to a tournament technicality, an agreed double pass in an even score, I couldn't help considering it. And then it dawned on me that we had made unnecessary assumptions because in the initial concept, points were awarded in two ways, a positive and a negative one:

1. One point for every stone on the board.
2. Minus P points for every group on the board.

There's no mention of vacant points, other than their implicit fuction. We introduced them by considering territory in a Go-like perspective and in consequence 'vacant territory' as inherent in the concept. Next we had of course to regulate it. We abandoned Occam's Razor.
 
Out of the trap
There's no need for any regulation if moving is compulsory. That's a simplification, because you have to explicitly rule moving either way, whether compulsory or optional, but the vacancy problem now regulates itself - the hallmark of a quintessential game.
 
Hot and cold
This is a fashionable distinction to make ... afterwards. In hot games you'd always want to move, say Hex, in cold games you'd sometimes rather not, say Oust. In terms of inventing I don't care one way or the other.
But clearly, compulsory movement in Symple suddenly leads a hot game into a cold finish. If growing options have run out, compulsory movement requires a new group to be started, bringing 1-P points. Not a position one would volontarily enter, though no more dramatic than seeing an intended connection being cut by the opponent either. Moreover, you can see it coming from fairly far away and take precautions, such as going easy on the growing, especially in the concave sections of one's groups, to save growing options for the endstage and avoid being forced to invade.  
Here's another trap to avoid: considering the new rule in terms of the old strategy. Benedikt was sceptical about the change and raised the question whether it "would be a tool or a randomizer", painting an image where a player at the end of the game suddenly would notice the problem. That, in my opinion, is not how it works: strategy has to be modified to meet the requirements of a now predictably cold and suddenly quite dramatic endgame. The same holds for another objection:
Quote:
"So, the game will come to a phase between growing and forced invasion where both players fill in their respective territories one stone a move each. Even on small boards this will be tedious."

Symple is hot up to and including the middle game where the first shadows of cooling may take 'shape', literally. In unbalanced games, the outcome isn't of any interest. In balanced games a 'reverse race' as Benedikt describes is dramatic in tactical terms, but it's not as if all groups are surrounded at the outside and all have vacant points on the inside. In a balanced game it's a limited phase involving a limited number of groups, and the tactical conclusion of a prior strategy. The new rule may lengthen the avarage game by maybe ten moves each. That's still less than 50. If length be the measure, how 'tedious' is Go in comparison?
« Last Edit: Dec 26th, 2011, 11:17am by christianF » IP Logged
MarkSteere
Forum Guru
*****




Finite games rule

   
WWW

Gender: male
Posts: 289
Re: Essay by Christian Freeling on inventing games
« Reply #813 on: Dec 26th, 2011, 11:17am »

on Dec 26th, 2011, 10:56am, christianF wrote:

This was sheer complacency, fueled by...

...a thirteen month hail of champagne corks.
IP Logged
clyring
Forum Guru
*****



Arimaa player #6218

   


Gender: female
Posts: 359
Re: Essay by Christian Freeling on inventing games
« Reply #814 on: Dec 26th, 2011, 12:38pm »

on Dec 26th, 2011, 10:56am, christianF wrote:
Quote:
"So, the game will come to a phase between growing and forced invasion where both players fill in their respective territories one stone a move each. Even on small boards this will be tedious."

 
Symple is hot up to and including the middle game where the first shadows of cooling may take 'shape', literally. In unbalanced games, the outcome isn't of any interest. In balanced games a 'reverse race' as Benedikt describes is dramatic in tactical terms, but it's not as if all groups are surrounded at the outside and all have vacant points on the inside. In a balanced game it's a limited phase involving a limited number of groups, and the tactical conclusion of a prior strategy. The new rule may lengthen the avarage game by maybe ten moves each. That's still less than 50. If length be the measure, how 'tedious' is Go in comparison? mparison?

The tedium this references is what might happen if both players know there is no space left of sufficient size to profitably invade but have many small spaces on the board... The optimal strategy would seem to be for each player to fill in these holes one stone at a time until one player runs out of holes to fill and is forced to lose points by invading. If there are enough such small holes this may drag on for quite awhile... Since the game is 'hot' until near the end anyway, perhaps a remedy for this potential problem would be to change
  • Grow any or all (but at least one) of his groups by one stone, or ...

to  
  • Grow all of his groups that can be grown by one stone (In the sense that, under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends)

...such that this dull phase of the game would really only add 5-10 turns and not many more.
« Last Edit: Dec 26th, 2011, 12:45pm by clyring » IP Logged

I administer the Endless Endgame Event (EEE). Players welcome!
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
Re: Essay by Christian Freeling on inventing games
« Reply #815 on: Dec 26th, 2011, 1:30pm »

on Dec 26th, 2011, 12:38pm, clyring wrote:
Since the game is 'hot' until near the end anyway, perhaps a remedy for this potential problem would be to change
  • Grow any or all (but at least one) of his groups by one stone ...

to  
  • Grow all of his groups that can be grown by one stone (In the sense that, under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends)

...such that this dull phase of the game would really only add 5-10 turns and not many more.

I'm not wholly sure what to make of "under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends", but forcedly growing all groups (groups that have no vacant adjacent points are excluded implicitly) might throw out the baby with the bathwater. There would no longer be any strategic judgement involved, on whether or not to grow a particular group. 'Scaling' the amount of growth between the implied 'advantage versus risk' is a strategical consideration. Moreover, I feel restrictions and obligations should always be kept to a bare minimum, and under the current rules you are obliged to put at least one stone - if more are allowed, that's optional. Also, I don't quite see how a game with a one stone pace in one particular stage would suddenly be dull. What about games that have the same pace during the whole game? Like most games?
 Huh
 
The current game between Luis and me is played under the new rule and might serve as a tentative reality check.
Luis has used black's prerogative on his 4th move, taking three grown stones as the compensation for lagging one behind in the number of groups. If he had declined at that point, then I would have grown four stones next, and I would have lagged behind instead. That's the nature of the balancing rule. Now we're in the placement stage for a while, but since it is a P-8 game, keep in mind that groups must at least grow to four stones to be 'neutral'.  
 
« Last Edit: Dec 26th, 2011, 2:23pm by christianF » IP Logged
MarkSteere
Forum Guru
*****




Finite games rule

   
WWW

Gender: male
Posts: 289
Re: Essay by Christian Freeling on inventing games
« Reply #816 on: Dec 26th, 2011, 3:34pm »

on Dec 26th, 2011, 12:38pm, clyring wrote:

Since the game is 'hot' until near the end anyway,

"Game heat" was precisely defined by brilliant combinatorial game theorists who know everything about game analysis and [fbleep] squat about game design, which latter category seems to include you.
IP Logged
clyring
Forum Guru
*****



Arimaa player #6218

   


Gender: female
Posts: 359
Re: Essay by Christian Freeling on inventing games
« Reply #817 on: Dec 26th, 2011, 7:48pm »

on Dec 26th, 2011, 1:30pm, christianF wrote:

I'm not wholly sure what to make of "under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends", but forcedly growing all groups (groups that have no vacant adjacent points are excluded implicitly) might throw out the baby with the bathwater. There would no longer be any strategic judgement involved, on whether or not to grow a particular group. 'Scaling' the amount of growth between the implied 'advantage versus risk' is a strategical consideration. Moreover, I feel restrictions and obligations should always be kept to a bare minimum, and under the current rules you are obliged to put at least one stone - if more are allowed, that's optional. Also, I don't quite see how a game with a one stone pace in one particular stage would suddenly be dull. What about games that have the same pace during the whole game? Like most games?
 Huh

To give you an idea of what I mean, I'll provide an example:
(EDIT: Assume that P>4)
However artificial this position may be, I shudder at the thought that positions with similar properties may eventually crop up regularly in high-level match play. A slow game pace in and of itself, as in go, is not bad, but having players mindlessly lay down 1 (or 2) stones in their own territory for dozens of turns is uninteresting, I'm certain you will agree.
 
As for "In the sense that, under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends," I refer to this situation:

White can grow any of their groups in particular, but not all of them on the same turn. However, on further thought, perhaps a simpler phrasing would work as well and perhaps even add some tactical trickery involving 'almost-surrounded groups...'
  • "Grow as many as possible of his groups by one stone, or..."

Black to play and cut
« Last Edit: Dec 26th, 2011, 7:51pm by clyring » IP Logged

I administer the Endless Endgame Event (EEE). Players welcome!
SpeedRazor
Forum Guru
*****




Skepticism is the Immune System of Science

   
Email

Gender: male
Posts: 72
Re: Essay by Christian Freeling on inventing games
« Reply #818 on: Dec 26th, 2011, 8:25pm »

Mark is going to get a woody over that, Clyring.  Your pattern is in the same sequence as his opus, his swansong, Tanbo - "The greatest abstract ever", or some such.
 
(Actually, 16-stone Tanbo plays better, not on an 19x19, or 13x13, but on an 16x16 torus.  No edges:  stones on the boundary aren't partially "bounded" to start.)
 
Put that in your pipe and smoke it, "architect".
« Last Edit: Dec 26th, 2011, 8:44pm by SpeedRazor » IP Logged
MarkSteere
Forum Guru
*****




Finite games rule

   
WWW

Gender: male
Posts: 289
Re: Essay by Christian Freeling on inventing games
« Reply #819 on: Dec 26th, 2011, 9:30pm »

The disjoint border connections of the torus are an aesthetic issue, as well as a clarity issue.
IP Logged
SpeedRazor
Forum Guru
*****




Skepticism is the Immune System of Science

   
Email

Gender: male
Posts: 72
Re: Essay by Christian Freeling on inventing games
« Reply #820 on: Dec 27th, 2011, 1:43am »

on Dec 26th, 2011, 9:30pm, MarkSteere wrote:
The disjoint border connections of the torus are an aesthetic issue, as well as a clarity issue.

Toroidal Go Board:  It would seem that if you took a Go board and let the lines wrap around to each other in 3D so that there is no edge, and each line is endless, that there would be a problem in the corners, that they should be triangular, or something - a "disjoint border" as you mentioned.  But this is not the case with a torus:  every node is equal to every other node.  There are no borders.
 
 
 
19 x 19 Toroidal Go Board
 
 
But you are right that clarity can be a problem; especially with poorly made boards.  With good boards, looking close, this is not a problem.  Actually, it's easier to see the 3D model the more stones there are.
« Last Edit: Dec 27th, 2011, 2:16am by SpeedRazor » IP Logged
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
Re: Essay by Christian Freeling on inventing games
« Reply #821 on: Dec 27th, 2011, 5:34am »

Wow, did you actually make that?
IP Logged
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
rogramming Symple
« Reply #822 on: Dec 27th, 2011, 6:19am »

on Dec 26th, 2011, 7:48pm, clyring wrote:

To give you an idea of what I mean, I'll provide an example:
...
However artificial this position may be, I shudder at the thought that positions with similar properties may eventually crop up regularly in high-level match play. A slow game pace in and of itself, as in go, is not bad, but having players mindlessly lay down 1 (or 2) stones in their own territory for dozens of turns is uninteresting, I'm certain you will agree.

Yes I do and you may be right. On the other hand the position is indeed artificial.
 
Whether or not you are right might soon be established. Symple games aren't that long, not even now, and if this becomes a repetitive and annoying phenomenon, well, then I was wrong and we can return to the optional movement version. There is nothing wrong with it in terms of rules, except maybe the legal draw after a successive pass in an even score position that started this discussion in he first place. But I would regret the impied lack of drama.
 
on Dec 26th, 2011, 7:48pm, clyring wrote:

As for "In the sense that, under the current rules, there are no legal placements for a growing stone on any ungrown group after the turn ends," I refer to this situation:

White can grow any of their groups in particular, but not all of them on the same turn. However, on further thought, perhaps a simpler phrasing would work as well and perhaps even add some tactical trickery involving 'almost-surrounded groups...'
  • "Grow as many as possible of his groups by one stone, or..."

Yes, here white can grow at most one, otherwise the original 'centergroup' will grow at more than one point, which is illegal. I still don't quite see what is problematic about that. As to growing "as many as possible of one's groups", that would be all groups that have vacant adjacencies. But since growing one may imply not being able to grow another, unclear situations might arise.
 
on Dec 26th, 2011, 7:48pm, clyring wrote:

Black to play and cut

B3 and B4 (legal since neither black group gets more than one stone at its original adjacent vacant points) would separate the four whites. Add F2 and F5 and you prepare a second cut, with white's potential growth at the three rightmost groups reduced to one stone (if he wants to connect the two rightmost ones). Yes?


What makes me trust the current ruleset is its minimalism and rigid logic. Quote:
A group consists of one or more orthogonally connected like colored stones. Players move in turn, white moves first. On his turn a player must either:
    Grow any or all (but at least one) of his groups by one stone, or ...
    ... put a stone on a vacant cell, not connected to a like colored group, thereby creating a new group.
    A player may only grow at groups as they exist at the beginning of his turn, and no such group may grow more than one stone in that particular turn.
    Turn order balance: If, and only if, neither player has grown yet, then black may grow any or all of his groups followed by a single stone placement, in the same turn.
    The game ends when the board is full. A player's score is the number of his stones minus P times the number of his groups, where P is a beforehand agreed upon even number.

That's pretty concise. And there's one more thing about it: handling the 'cooling' stage must be an absolute crime in terms of MCTS (never mind traditional evaluation and alpha-bèta pruning). Programming Symple and simple programming must be very different worlds.
« Last Edit: Dec 27th, 2011, 8:15am by christianF » IP Logged
clyring
Forum Guru
*****



Arimaa player #6218

   


Gender: female
Posts: 359
Re: Essay by Christian Freeling on inventing games
« Reply #823 on: Dec 27th, 2011, 8:13am »

on Dec 26th, 2011, 7:48pm, clyring wrote:
...and perhaps even add some tactical trickery involving 'almost-surrounded groups...'
  • "Grow as many as possible of his groups by one stone, or..."

Black to play and cut
on Dec 27th, 2011, 6:19am, christianF wrote:
B3 and B4 (legal since neither black group gets more than one stone at its original adjacent vacant points) would separate the four whites. Add F2 and F5 and you prepare a second cut, with white's potential growth at the three rightmost groups reduced to one stone (if he wants to connect the two rightmost ones). Yes?

You have the right move, but not the full answer... If you assume the growing rule to be my proposed alternative, then white cannot legally connect at F3 or F4 on their move because doing so would allow for only 3 of the onscreen groups to be grown while growing all 4 of them is possible, thus allowing black to completely cut off the middle group(s) on his next move.
IP Logged

I administer the Endless Endgame Event (EEE). Players welcome!
christianF
Forum Moderator
Forum Guru
*****



Arimaa player #4019

   


Gender: male
Posts: 804
Re: Essay by Christian Freeling on inventing games
« Reply #824 on: Dec 27th, 2011, 8:52am »

on Dec 27th, 2011, 8:13am, clyring wrote:
You have the right move, but not the full answer... If you assume the growing rule to be my proposed alternative, then white cannot legally connect at F3 or F4 on their move because doing so would allow for only 3 of the onscreen groups to be grown while growing all 4 of them is possible, thus allowing black to completely cut off the middle group(s) on his next move.

That's why I fear for disagreement over the interpretation of a player "growing as many as possible of his groups by one stone". Figuring out the maximum doesn't seem to add to clarity.
 
As for the example, you consider white's turn after black has moved B3-B4-F2-F5 I presume.
The A-colomn is now isolated. The rightmost three groups could grow two stones max.: G1(G6) combined with D3(D4). Connecting at the F-column would allow only one and therefore be illegal under your suggested rule.
I don't think this adds to clarity - very complex situations would have to be 'sorted out' on maximizing growth.
 
Usually, if at all. I want obligations to be as unintrusive as possible. As far as compulsory movement is concerned, having to place at least one stone does comply with that.
« Last Edit: Dec 27th, 2011, 8:55am by christianF » IP Logged
Pages: 1 ... 53 54 55 56 57  ...  78 Notify of replies Notify of replies Send Topic Send Topic Print Print

« Previous topic | Next topic »

Arimaa Forum » Powered by YaBB 1 Gold - SP 1.3.1!
YaBB © 2000-2003. All Rights Reserved.