Author |
Topic: Essay by Christian Freeling on inventing games (Read 540019 times) |
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #975 on: Oct 31st, 2012, 1:07pm » |
|
A new category of abstract strategy games Let's, for argument's sake, pretend that abstract strategy games matter, then I unwittingly seem to have followed a trend that would appear logical in the interconnected virtuality we live in: boardgames without physical boards. All classic board games can nowadays conveniently be played online, so the emergence of strategy games that can conveniently be played online, but that have features that make them difficult or even impossible to play in the physical world, was an accident waiting to happen. This trend is much wider than my contribution, but the three games that have been the most exposed here lately show different degrees of the idea. Symple can be played fairly conveniently over-the-board, provided one has Go stones with a small dot on one side. Placing several stones and turning them around after placement is more complex than placing a stone, but on the other hand, Symple knows only placement, so there's no removal of captured stones - it kind of evens out. Edit: Or say it would even out if it were not for keeping the score: here an applet is convenient to the point of being indispensable. Sygo is more problematic if one would consider over-the-board play, too problematic actually. Mu velox cannot be played over-the-board at all. The point being that a realm has been opened of strategy games that may be interesting or significant in some or all of their aspects, but that could never have been invented or discovered in a world where a strategy game is synonymous with a physical board game.
|
« Last Edit: Nov 2nd, 2012, 7:15am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
th of improvements)to where I live
« Reply #976 on: Nov 5th, 2012, 12:40pm » |
|
For the record: I've adapted the cooperative cycle in Ayu to the 11x11 board and put it in an applet. It's a clever discovery by iGGC's "PCM", but not a game position. Whether such a cycle can be found in a legally played game remains unknown . Our last Mu_velox game revealed the same as the one before that: two long starting combinations that build to a 5-column before the first explosion, beat three shorter ones that build to a 3 column. Crowded birth chambers pay off. Moreover, having three sections of territory always means one is the weakest. If the opponent manages to get on the wall of such a section, it is more than likely to fall. Keeping that in mind, here's the next game. Remco Bloemen, who's entry in the CodeCup was having the whole competition for lunch, happens to live close to where I live. Thus I learned that his program isn't MCTS based, but rather the heuristic intended to support a MCTS program he's working on. The next round this clever little 'stand-alone support program' won 29 out of 32 against a presumably stronger opposition (one month of improvements). I'm curious how strong the indended program might eventually become, if at all stronger than the heuristic. I'll play against the current version shortly. It's a bit of an irony that, while looking for games that humans can play at a high level, but that pose difficulties for programmers, I stumble upon a game that I don't understand at all (despite winning regularly), but that programs can play at a high level .
|
« Last Edit: Nov 6th, 2012, 8:38am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #977 on: Nov 9th, 2012, 7:44am » |
|
Game inventors like competitions and challenges and awards, so here's another one. The last one I entered was the "thousand year game design challenge". I wonder if anyone remembers the winner one year later - don't look! Judging by the choice of the winner, in a thousand years we'll be playing with pebbles . Made me consider how someone in 1912 might have envisioned the year 1920. And how someone in 2012 may envision 2020. Our government, or lack thereof, envisions up to 2050 or so. Good luck! Anyway, may the best game win and all that, and please enjoy these two thirds of Kobus, who posed so nicely this morning that I couldn't resist :
|
« Last Edit: Nov 9th, 2012, 7:45am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #978 on: Nov 11th, 2012, 6:33am » |
|
Update on the CodeCup 2013 The sixth test competition has been completed. There were 18 competitors. Abdessamad ELKASIMI won 20 out of 22. Here are two games he lost, one against Remco Bloemen (2nd) and one against Bertrand Lunderer (3rd). Remco won by 1 point. Note that he has eight groups, and his opponent only two, indicative of the programs' priorities being growth and connectivity respectively (at least in this game). For some reason a winner gets 100+ points added to the score. This may be part of the competition protocol, I don't know.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #979 on: Nov 12th, 2012, 11:29am » |
|
Now, that was some fun. I won't comment here on people who are banned from this site, only thank Omar and Whatever Divine Authority may be involved, for the ban. A few thinkers about abstract games ... Others, who are perfectly free to reply, may be mentioned though, like Benedikt Rosenau, reluctant co-author of Symple, who thinks it wise to ignore persistent attacks on the game, because the attacker is like terminator, not to be be reasoned with and absolutely unstoppable. So he may well be right to keep silent while being pissed on. I won't because I'm too pissed off. Then there's Luis Bolaños Mures, in my view the best game designer at the moment. Mark Steere has Oust, but basically (p)re-invented the game several times, and never better, in Tanbo, Flume, Redstone and Rive. Flume is a perfect little game, but it relates to Oust a bit as Othello does to Go: not quite that organic and not at all the same league. Luis' Ayu however is no less quintessential than Oust, and neither are his innovative twin games Xodd and Yodd. So I have high expectations of him. At the same time he values Mark's expertise to the point of accepting the inherent bullsh!t, especially if it is not directed at one of his games. But that might yet change. Actually I'm not sure about Nick Bentley who's best game, by his own admission, is Ketchup, but whose other games seem to fail to compete with that game's quality and appeal. Why would someone praise Corey Clark's Slither to heaven? Among square connection games it is an outstanding game. But 'square connection' isn't exactly a grand theme, like checkmate or elimination or territory. And Slither uses a pie for balance, where it doesn't work, unless the board is so big that it doesn't matter in the first place. So it's a great game, but nothing to get carried away about. I had been wondering about that. And about the fact that Ketchup went from version 1.0 to 5.0 and then back to 4.0 because 5.0 in retrospect "wasn't an improvement". Now I'm starting to see a pattern. Nick questioned the balancing mechanism for Redstone, but later called the game a viciously good invention. Redstone is not a Go variant but an elimination game using Go mechanics, and I can reassure Nick: it doesn't need a balancing mechanism. The rules suggest a pie. A pie presumes that you can judge the value of an initial move. Jugging the value of an initial move presumes a general idea of the relation between strategy and tactics. In Redstone this relation is evasive to the point of being invisible. The strategic goal is way off and the global impact of the outcome of local fights is unpredictabe. Strategy is hard to find and tactics serve themselves. And despite Redstone's finitude a game can seem to take no end. I can see that almost at a glance, but Nick apparently is blinded by simplicity of concept - a necessary condition for a good game, but not a sufficient one. It surprises me because Nick's criteria for what makes a good game are are very well thought through. Barring other reasons for his praise of Redstone, I'd say he values the right concepts, but lacks the intuition to see how they might pan out in an actual game. Redstone sucks, but it may take determined supporters a while to find that out. Why do I mention these guys? Barring Benedikt they're the only contributors to (and account for 50% of the visitors of) rec.games.abstract, a self declared 'nonsense-free forum' dedicated to cyclophobia and finitude. Read that again and see the irony . ... and what they have in common But including Benedikt they're all concerned with abstract strategy for the sake of it. They're not the only ones of course, yet most inventors are in the business of game inventing to market particular games, and that's a perfectly respectable object, but not in the least interesting for those who want a game to be a mental sports weapon that would last, if not for centuries, then at least for the foreseeable future. Exhaustible tactical funnies, mostly based on peculiar mechanics rather than concept, don't fall into that category. Games have a spirit There was a suggestion that my evaluation of Redstone was influenced by my low opinion of its inventor. I have no low opinion of its inventor, not in that quality. I have a low opinion of Redstone and was just surprised that the inventor's claims were left unchallenged by knowledgeable people like Nick and Luis. So let me be more precise, for those who cannot immediately see this monkey's disease. A contagious disconnect I mentioned the disconnect between tactics and strategy. Redstone is an 'organic' game, no doubt. The basic idea is simple (no pun intended) and not new (the Glass Bead Game has neutral stones that 'capture but cannot be captured') and its implementation doesn't leave much room to deviate from the rules as they suggest themselves. A self-explanatory organism, and in that case you get what you get, for better or worse. And no doubt Redstone is 'playable'. But the author put forward a number of claims in the comments on the game's entry in BGG's 'best combinatorial game of 2011/2012' award, that made Benedikt raise an eyebrow. I'm not going into them here, but in it the inventor presented the preliminary evaluation of the game by a knowledgeable Go player, and here the game's spirit showed its subtle powers of infiltration: Quote:"My initial impression of Redstone, from having played it twice now, is one of interest. Strategically, it's basically Go, but with a few twists. Tactically, it has an inherent system that I'm inclined to believe increases the depth of complexity quite a bit. This will have an impact on strategy, but I'm not entirely sure what yet." |
| That a game of annihilation would 'basically be Go', strategically, is questionable in the first place, but notice how Redstone's disconnect is translating into a similar disconnect in the argument: if the increased depth of complexity has an as yet unknown impact on strategy, then calling the strategy 'basically like Go' boils down to presenting fog as clarity. Cancer Here's another example of the game's spirit showing its subtle powers of infiltration. It's original name was 'Cancer'. I'm not a big fan of descriptive names for games, but I must make an exception here because the name totally fits. In itself the gradual disappearance of a given playing area, the Atlantis effect, is neither new nor undesirable. Amazons is a well known example, as are the 'Atlantis Triplets', of which I mention Shakti. But in these games the blocks that arise, either by placement or removal, result from one player's strategy. In Redstone they result from capturing fights and escape attempts that may go this way or that, resulting in differences in placements of red stones that may have huge impact further on, without a clue as to the nature of the impact. It starts with a capture and a single red stone here, and a capture and a single red stone there, and no clue as to how the board will even remotely look in the middle game. Like cancer they are opportunistically growing and seeding, and players are fighting not only against one another, but against the game. Show a Go master a middle game position and he'll know in seconds what it is about, because of his experience, and because the game has great clarity. Cancer has no clarity. It's time to shift the focus. I mentioned the inventors above for their concern with abstract strategy for the sake of it. In so far as they're all game philosophers, I'd like to address the question What is the future of abstract games in a world full of virtual games? To this day, and for the foreseeable future, Go and Shogi seem all right, but what about Chess, Draughts, Checkers, to mention three stages of obvious decline? Do we need more 'modern' games, and if so, how might these involve more modern means. Is "playable on an actual bord with actual pieces" a significant criterion, only because two Chess players demonstrate this possibility at regular intervals at the center of an otherwise virtual worldwide display. Do we still start a car with a crank? Does the advance of Chess computers have an impact on the game? Are games between humans interesting if mistakes are immediately and continuously highlighted? Are games between humans and computers interesting if computers win? Are Go and Shogi sufficiently computerproof? Anyone for dead Checkers or terminal Draughts? If we want abstract games to play some kind of role in the future, do we have a problem?
|
« Last Edit: Nov 14th, 2012, 11:25am by christianF » |
IP Logged |
|
|
|
NickBentley
Forum Senior Member
Arimaa player #6704
Gender:
Posts: 44
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #980 on: Nov 14th, 2012, 4:21pm » |
|
Interesting! Some quick thoughts in response: I'm not sure what I think about my intuition either. I'm sure that I can make good games though. However limited my intuition may be, the larger bottleneck for me, without question, is time. I have a full life and in making tradeoffs, game design often gets left to the side. I assert that the pie rule in Slither works better than you propose. The key is that there's an opportunity cost to moving a stone. So you can move the pie stone inward, but at the expense of not moving other stones elsewhere. The cost is big enough that I'd guess you can finely balance the game on a board 10x10 or larger. Not that it matters at this point, because the tactics are so insanely hairy that an initial imbalance is washed away for everybody for now and probably will be for some time into the future. Another thought about Slither: when I evaluate a game, conceptual unity is a prerequisite for high marks, but if that bar is passed, my evaluation depends strongly on what it feels like to play. No game sets patterns dancing in my head like Slither does, and that's why I praise it so highly. Of course there's nothing universally objective about this and it could reflect nothing more than the idiosyncratic wiring of my brain. My opinion on Redstone: I assume that a) it's balanced for most of us who play it now; b) at higher levels of play, it will be imbalanced (can't prove it, but that's my thought); but that's ok because c) some kind of swap protocol will fix it. My statement about "vicious goodness" depends especially on c) being true. I emphasize that my opinions are provisional - my intuition doesn't work well enough for me to be confident (to corroborate your point about my intuition). My experience with Redstone comes from a bunch of games I've played against myself, which also complicates evaluation. Finally, about r.g.a.: yep, it's a wasteland. I would kill for a widely read/active discussion forum where nobody gets personal and the focus stays on productive discussion of game mechanics. One thing I will say: personal experience demonstrates that it's not hard to have a perfectly cordial relationship with Mark: just be respectful of his opinions even when he says things you don't like. You can even tell him he's wrong and he won't bite your head off, if you do so respectfully. And when mark does go into attack mode, just don't respond and there will be no problem. The proof is in the pudding: Mark has been very critical of my games in the past just as he has of yours, but you'll notice that my relationship with Mark remains warm. I don't know if I'm out of place offering advice about to interact with Mark (you certainly didn't ask for it), but I feel that some small changes in your online behavior would staunch the craziness that poisons so many of our discussions. Now you might tell me that really Mark is the one who should change his behavior, and I don't disagree, but I've made the same appeal to Mark before, to no avail. In the absence of any hope of change in Mark's conduct, I now appeal to you.
|
|
IP Logged |
|
|
|
NickBentley
Forum Senior Member
Arimaa player #6704
Gender:
Posts: 44
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #981 on: Nov 14th, 2012, 5:21pm » |
|
One thing I forgot to mention about effectiveness of pie rule in Slither: because the board fills up a lot more in Slither than in Hex, for example, individual stones are less powerful, which means the pie rule doesn't have to do as much work.
|
|
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #982 on: Nov 15th, 2012, 4:30am » |
|
on Nov 14th, 2012, 4:21pm, NickBentley wrote:Interesting! Some quick thoughts in response: I'm not sure what I think about my intuition either. I'm sure that I can make good games though. |
| No question about that, it was you characterization of Redstone that made me wonder. on Nov 14th, 2012, 4:21pm, NickBentley wrote: The cost is big enough that I'd guess you can finely balance the game on a board 10x10 or larger. Not that it matters at this point, because the tactics are so insanely hairy that an initial imbalance is washed away for everybody for now and probably will be for some time into the future. |
| Some considerable time indeed, and my point exactly. One thing I don't understand is the preoccupation with 'balance'. Slither doesn't need a pie and never will, because it is a great game in a neglected class full of competition. No-one cares, so the level of play will never get to the point where it matters. And if so, then it'll probably be MCTS/UCT programs that are less confused by hairy tactics. Slither is not the only game facing this predicament. on Nov 14th, 2012, 4:21pm, NickBentley wrote:My opinion on Redstone: I assume that a) it's balanced for most of us who play it now; b) at higher levels of play, it will be imbalanced (can't prove it, but that's my thought); but that's ok because c) some kind of swap protocol will fix it. My statement about "vicious goodness" depends especially on c) being true. |
| So far as I can see b) is true and c) is not true. I've tried to explain why. Consider the red stones as a hard to handle opponent for both players and their cancerous nature becomes visible. That's not to say that I consider the game unplayable or unbalanced. Redstone is inherently balanced, or as you put it "Not that it matters at this point, because the tactics are so insanely hairy that an initial imbalance is washed away for everybody for now and probably will be for some time into the future.". I couldn't have put it better regarding Redstone. on Nov 14th, 2012, 4:21pm, NickBentley wrote:Finally, about r.g.a.: yep, it's a wasteland. I would kill for a widely read/active discussion forum where nobody gets personal and the focus stays on productive discussion of game mechanics. |
| Not to put too fine apoint on it, but there's a tainted thread in forum called "The Arimaa Off Topic Forum". You're currently in it. It's tainted because the High Priest of Cyclophobia has issued a decree stating that Arimaa does not comply with the Church's Orders. Omar also issued a decree that makes this a place where "nobody gets personal and the focus stays on productive discussion of game mechanics." In fact I've just suggested a subject matter that would be boring if I were to be the only contributer. Not to mention predictable . P.S. Why do you think I published Mu only here? It's not finite, not drawless and, depending on the definition I might not even disagree with 'unbalanced' or even with the question whether it is a strategy game or a weird kind of pinball device. P.P.S. Why do you think Mark is so troubled by Symple? It's enough to know that in his view a game, ideally, should be finite, drawless and perfectly balanced. You kind of got away with it with Ketchup .
|
« Last Edit: Nov 15th, 2012, 5:31am by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #983 on: Nov 15th, 2012, 6:21am » |
|
What is the future of abstract games? (part 1) For context: we're living in an interconnected world where computers can display intelligent game boards and play intelligent games. Yet the pinacle of abstract gaming is two players engaged in a centuries old game, pushing wood of a somewhat medieval disposition on a wooden board, while computers in the background are displaying large electronic boards and giving their evaluation. An evaluation that may soon be better than the one of either contestant. Unless you're a Chess player it's a somewhat anachronistic picture. The world is changing, and if Chess has always been meant for eternity, eternity appears to have come to an end. Other games fare even worse. Checkers There was a time when Checkers was fairly highly regarded by all but the Chess community. No more. Checkers has been solved and humans play in the shadow of perfect play. They do so with balloted openings and the like, and it has become a sorry affair altogether. Checkers has once more become a recreational game, not something people would care about in terms of who's the world's best player. It's common knowledge Chinook is. International Draughts So take Draughts, still standing after all these years. Much wider, far richer in terms of combinations, and not yet fully conquered by the bots. But it's only a matter of time before the bots will dominate it. Draughts has accumulated a massive literature and theory, and smart programs will dig their way in, to support an ever improving evaluation. Draughts is a determined draw. It's not proven, but neither is life beyond earth. Only an idiot would doubt it. Sixtyfive years ago, Piet Roozenburg could score 17 wins and 3 draws in a tournament and become 'sportsman of the year' in the Netherlands. Nowadays matches and tournaments for the worldchampionship alternate but the top players are so evenly matched neither can win in a match. So there's tournaments, and draws are increasingly decided by tie-breaking blitz games. Did you know who the current world champion is? Here's the complete 100+ years list. Now let's take the last match, between Schwarzman and Georgiev. For starters 12 draws. The games were played at four different locations near Enschede, where I live. Hardly the center of the world, and despite local media, I didn't even notice the event. Then three rapids (20 min. pp) and three blitz (10 min. pp), and both had won 1 game. Then came micro matches (don't ask) and Schwarzman won. Such is the state of affairs, and Draughts remains popular despite it, but the circus gets increasingly localized and marginalized for lack of sponsoring. To even get some attention by national media in the Netherlands (never mind international), Ton Sijbrands must be persuaded to break his blindfold record once more. But there's an end to that too, eventually. Chess Here's an opening quote from wiki's human–computer chess matches: Quote:"After convincing victories in two matches in 2005 and 2006, it appears that chess programs can now defeat even the strongest chess players." |
| It seems inevitable that human Chess is condemned to the constant shadow of, if not perfect play, then in any case superior machine evaluation. Doesn't that suck just a little bit I wonder? Doesn't it take just a little bit away from the unconditional admiration one might have for human achievement on 64 squares? I think it does, though personally I'm not qualified to admire in the first place, for lack of talent and knowledge. I'm a spectator, I like to see heroes and idolize them. Will the heroes of the future be the mere butlers of programs? If Chess becomes a knowledge based dead end street where the main issue is whether a human player can still reach a draw against the strongest programs, and where humans are tested on their knowledge of the latest discoveries rather than on insight and style, I fear something will go wrong, eventually. And the writing is on the wall. There's no easy way out either. Draughts players hang on to a sinking ship and yet refuse to question the game. Checkers has already sunk back to the grass roots level it once was raised from, yet the circus goes on and nobody is wet, apparently. Chess is even 'holier' than these games. Any problem regarding matches ending in knowledge based draws will be denied till it slaps them in the face, and even then, draws will never be a problem for the vast majority of the Chess community, so what. Everything counts in large amounts, and Chess has the numbers. 2b continued ...
|
« Last Edit: Nov 18th, 2012, 8:33am by christianF » |
IP Logged |
|
|
|
clyring
Forum Guru
Arimaa player #6218
Gender:
Posts: 362
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #984 on: Nov 15th, 2012, 7:27am » |
|
I think your remarks on redstone are too harsh. Although the final goal of the game is elimination, this is achieved by gaining a territorial advantage. Since redstones can only be placed to capture a group and suicide via a regular stone is illegal, groups with two or more eyes are safe until the only legal moves are to fill these eyes. When this happens, the player who was able to create more extra eyes for his groups will be the last player compelled to sacrifice his own groups. Of course, the number of extra eyes is not the same as territory, but the two are nevertheless closely related. on Nov 12th, 2012, 11:29am, christianF wrote:Is "playable on an actual bord with actual pieces" a significant criterion, only because two Chess players demonstrate this possibility at regular intervals at the center of an otherwise virtual worldwide display. Do we still start a car with a crank? |
| There is something more to this than what you have described- Playing the game in public is basically free advertising.
|
« Last Edit: Nov 15th, 2012, 7:28am by clyring » |
IP Logged |
I administer the Endless Endgame Event (EEE). Players welcome!
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #985 on: Nov 15th, 2012, 8:09am » |
|
on Nov 15th, 2012, 7:27am, clyring wrote:I think your remarks on redstone are too harsh. Although the final goal of the game is elimination, this is achieved by gaining a territorial advantage. Since redstones can only be placed to capture a group and suicide via a regular stone is illegal, groups with two or more eyes are safe until the only legal moves are to fill these eyes. When this happens, the player who was able to create more extra eyes for his groups will be the last player compelled to sacrifice his own groups. |
| Since we're engaged in a game of Symple you know I can but agree. It's the same mechanism only Symple is co-existential and Redstone isn't - I'd almost add 'how could it be?' Yet the additional red stones that Redstone needs to meet that end, do not add to clarity. That's not meant for fact, but it's a considered opinion. This is what Graham wrote in his evaluation: Quote:"My game against illluck was an eye-opener, as I fabricated a living group out of thin air inside my opponent's territory. There are a number of hane / crosscut situations which simply don't work in Go, but do in redstone, as you can end up with eyes appearing out of nowhere when red stones hit the board. It's also clear that two point eyes are frequently not eyes at all, and single stone self atari sacrifice play has gained a lot of leverage." |
| In Redstone you'll have to consider the position after a capture much more carefully than in Go. Some may consider that a good quality, I consider it a bad effect on clarity. It's more difficult to see where a game goes, strategically speaking, unless the position is uneven enough see that anyway. on Nov 15th, 2012, 7:27am, clyring wrote:There is something more to this than what you have described - Playing the game in public is basically free advertising. |
| Well, yes, but it's a bit anachronistic. Often I see one of the contestants studying the electronic display rather than the board. It's not the public display that I find strange, the more the better, but the means by which the game is played. Boards are convenient, but more convenient boards are being invented on a regular basis. At some point tablets will probably be able to cover most games. This is bound to change the nature of playing board games, in ways we can't yet predict. For one, if I cheat, using a strong program instead of my brains, and I got an equally human opponent (in the sense that using a bot is very human), then what's happening if it becomes a trend? A continuous worldwide tournament of programs posing as players? Wouldn't we need games that at least temporarily were devoid of strong bots, to preserve the fun?
|
« Last Edit: Nov 15th, 2012, 2:24pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #986 on: Nov 15th, 2012, 4:10pm » |
|
How about that. I got a question by Mark Steere: "Christian, what does it mean for a game to be perfectly balanced?" Four posts up I said this: "Why do you think I published Mu only here? It's not finite, not drawless and, depending on the definition I might not even disagree with 'unbalanced' or even with the question whether it is a strategy game or a weird kind of pinball device." This is a question that can be answered from several points of view, and I won't rush into one answer because I think there is no one answer that would cover all viewpoints. It's no secret that a discussion over say the pie rule can be pretty extensive, given the basic simplicity of the rule. And the pie rule is only one aspect of Mark's question. But given a true interest in the subject I'll try to give whatever insights I have and hope that I get response in the same spirit. One point of order, I don't post at RGA and Mark doesn't post here. I don't expect that to be a problem if everyone has the same agenda: "what does it mean for a game to be perfectly balanced?" I'll sleep over it first because we're approaching midnight here. However, I can argue one valid point of view from which no finite drawless game is perfectly balanced. But that point of view is not concerned with the human condition and human players cannot profit from it. I also feel that the question "what do you mean when you say Symple or Ketchup are perfectly balanced" is easier to answer than the generalized question, so maybe we can approach the issue along specific lines.
|
« Last Edit: Nov 15th, 2012, 4:54pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #987 on: Nov 16th, 2012, 6:04am » |
|
The game theoretical perspective This is not new and it may be boring. Zermelo's theorem is well known and empirically supported by a host of games that now are solved, including Checkers and Awari, both determined draws. Neither game is finite. Pit Chinook or the Awari Oracle against itself, and you get endless strings of draws. Perfectly balanced! By the same token, if applied to finite drawless games, you would get decisions without end, like Hex, in so far as it is solved for smaller boards. In finite drawless games any initial move is either winning or losing, so in that sense such a game cannot be 'perfectly balanced'. The AI perspective Solved games are no longer games from an AI perspective. In solved Hex the truth of any legal position can be quickly determined. Truth kills the meaning of concepts like strategy, tactics, style, attack, defense and fun. Truth is perfectly boring because it kills the game. And where it is not perfectly boring, like knowing that your first move in a game of 19x19 Hex is either winning or losing, it's perfectly useless. Even the knowledge that opening in the center is a win, may be perfectly useless against a significantly stronger player. The human perspective From a game theoretical perspective Hex cannot be called 'balanced' because Hex is decisive. From an AI perspective, solving the next boardsize is thrilling, but the result from then on is boring. Not yet from a human perspective, because knowing which opening placements are winning doesn't imply the knowledge how to do it. The basis for my idea of a 'perfectly balanced' game is- a) not the game itself, but the way in which the 'turn-order advantage' can be addressed, and
- b) human imperfection
Regarding the first, the most important property must be that it divides responsibility so that neither player can find reason to complain about it. Regarding the second: it's a necessary condition, though not sufficient, to play games at all. Hex-19 and the pie rule Going by example, imagine two seasoned Hex players playing a 19x19 game. Both are familiar with the division of winning and losing moves on smaller boards, even though they may not know the corresponding move sequences, that is: how to win such a 'won' position. They use the pie rule, extrapolating from their knowledge of the win/lose division on smaller boards.- The first player tries to find a placement that he can equally well play with as play against. It's a judgement call. He only knows the move is winning or losing, but since he doesn't know which he must maximize his foggy notion.
- The second player must evaluate the first one's judgement. He has no choice of placement, so he tries to minimize the foggyness presented to him, to decide to play either with the stone or against it.
So they play and one player wins. He won't complain in the first place. The loser is responsible for his own mistakes, but what if he didn't make mistakes? Of course there may and usually will be no way to tell, but say we invoke a divine entity to tell us the truth and see, he made no mistakes! In that case he made the wrong choice in the first place. He has no-one to blame but himself and has no reason to complain. That is what I mean. In 19x19 Hex, the pie rule is a protocol to 'perfectly balance the game', that is to divide responsibilities so as to give each player the full resposibility for the outcome, whether win or loss. 1-2-2-2 ... This is a move protocol most everyone knows from games like Connect6 and Ketchup. It is a protocol that alternately puts players 'in the lead', instead of putting one in front and having the other catch up. Nick gave it a brilliant twist in Ketchup. It's a bit like a bicycle race where, if you overtake your opponent, he gets a burst of wind from behind (that is, he may move 3 instead of 2). If he uses it in turn to overtake you, then you get an extra burst. You don't have to 'overtake' (create a group, larger than any one that previously existed of either color), but in order to win, you cannot decline all the time either. It carries the feeling of balance all through the game, but this is not related to turn-order balance: that is provided by the 1-2-2 protocol as such. I would have no trouble calling Ketchup 'perfectly balanced' like I do Hex-19 with the pie rule, but not on the same grounds. The 1-2-2 protocol is not perfect in the same sense because there is no division of resposibility. But it works to most everyone's satisfaction, and I'm human and a player, not a game theorist. Many games don't even have a turn-order balancing protocol and yet don't show any problems, because any impact the advantage might have is dwarfed under the impact of human imperfection. Luis' Ayu is not proven finite but it looks finite enough, and if anyone has a clue as to its turn-order advantage, please enlighten us. Intuition would put it at the first player's side but proving anything looks like a long shot. Many of thousands of games might empirically provide a statistical answer. For the moment the game does not seem to have any such problem. Draughts is not finite, but it has never had any such problem either. Not in the black/white division of decisions. These games may not be 'perfecly balanced', but they're perfectly playable. Ah, we still got Symple. We also got a nice evening with a couple of games and movies. Let's see ... movies it is Symple I woke up to an almost provocatively glorious day and decided to go to the weekly downtown market. By bike of course, it being an almost provocatively glorious day and all. I landed in the middle of the Entry of Sinterklaas, engulfed by hundreds of exited children and their parents, 'black Peters' all over the place, even on the roofs, and a very relaxed atmosphere despite all the excitement. I bought grapes and kiwis and bananas, and french fries with mayonnaise (Vincent in Pulp Fiction: "I've seen 'em do it, man. They fuggin' drown 'em in that sh!t."). On the way back I encountered the GoodHolyMan himself, riding a medium sized white horse. We even exchanged greetings. I must say the horse used to be bigger when I was a child. And all flat roads seemed downhill for excess energy. I had the same feeling on the way back, but it was backwind. Regarding Symple and why I feel it is totally correct to say that it is 'perfectly balanced', I must say that anyone who- understands the move protocol and its inherent strategic dilemma, and
- understands the working of the embedded balancing rule
would agree that it perfectly balances the game on the same grounds as I mentioned in the case of Hex-19: it divides responsibilities so as to give each player the full resposibility for the outcome, whether win or loss. And like the pie it is almost the same judgement call for both players. The difference is that Symple's balancing rule is embedded in the move protocol. Luis' offbeat generalization is not a generalization of it, as he tried to suggest. Symple's balancing mechanism is a unique trade-off between territory and growth potential and poses the same judgement call to both players over a trajectory of moves. Possible positions involving the call run into billions (no, not in one game). I had a very pragmatic girlfriend once, who defined "perfect" as "without any known imperfections". I've never heard a more practical definition.
|
« Last Edit: Nov 17th, 2012, 1:23pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #988 on: Nov 17th, 2012, 9:36am » |
|
Just saw Tom Ford make a 147 break in the European Tour in Sofia. Great, I felt the tension as if I were there. Later it turned out it was a replay during the midsession interval. Never mind . We're still fully engaged in Mu-velox, and strategic thinking as well as the tactics employed - dirty tactics Ed! - are still developing. Only thing is, they seem to be developing faster in Ed's case.
|
« Last Edit: Nov 17th, 2012, 12:03pm by christianF » |
IP Logged |
|
|
|
christianF
Forum Moderator Forum Guru
Arimaa player #4019
Gender:
Posts: 804
|
|
Re: Essay by Christian Freeling on inventing games
« Reply #989 on: Nov 18th, 2012, 10:14am » |
|
What is the future of abstract games? (part 2) In terms of going under, Checkers has drowned, Draughts is splashing and fighting for air and calling that everything is all right, and Chess players will soon serve computerprograms rather than the other way around. Keep in mind that Chess is most likely a determined draw too, and a bleak perspective unfolds. Deniable of course, but not forever deniable. Unfortunately, and maybe surprisingly, Shogi doesn't seem to fare much better when it comes to the strength of programs. Nor does Xiangqi, a game more associated with the past of Chess than with its future. Othello has been solved up to 6x6 (win +4 by the second player) and is on the brink of being solved on 8x8. Not that it matters, 8x8 computers are beating humans consistently nonetheless, and for what it's worth: Othello cannot captivate the human mind like Chess or Go, or even Draughts. The one that keeps up something of a struggle is Go. You can read the state of affairs in wiki. Combining knowledge-based evaluation based on pattern recognition with MCTS/UCT evaluation has led to "impressive results", and till now impressive results have usually led to programs that are stronger than the best human opponents. Yet it's noteworthy that the game keeping the castle for now, is a game of placement and capture. It would be hard to imagine that new games will fare much better. If Grand Chess were to develop a significant player base, computers would follow suit. Both humans and computers would initially be deprived of "knowledge" but in the end I can well imagine that the situation would not develop differently from the current one in Chess. So I think the future of abstract board games will not depend on the strongest players: they might invariably be bots. What it would depend on is fun, simplicity, availability on smart devices, speed and hype. Intelligent recreation maybe, but no longer interesting as a "sport", because in that capacity the players would end up serving the computers.
|
« Last Edit: Nov 18th, 2012, 10:17am by christianF » |
IP Logged |
|
|
|
|