Greg Costikyan (2013)
Uncertainty in Games First MIT Press paperback edition, 2015
I saw this title online and immediately thought of weaponizing "uncertainty" in over-the-board chess.
The Art of War states "all warfare is based on deception", but outright deception is difficult in a game of perfect information. Could uncertainty be a reasonable proxy for deception?
The author's focus is actually how a game designer can make a game more appealing by inserting different forms of uncertainty. Still, I think the topic can be developed for us chess players.
Costikyan identifies eleven sources of uncertainty in games:
- Performative Uncertainty : "your ability to master the skills of hand-eye coordination" (pg.20), "uncertainty of physical performance" (pg.71)
- Solver's Uncertainty : "uncertainty lies in the challenge of puzzle solving" (pg.24) (by deduction, or trial and error)
- Player Unpredictability : "unpredictable actions of the target" (pg.28)
- Randomness : (dice or roulette)
- Analytic Complexity : (decision tree with high branching factor)
- Hidden Information : (cards)
- Narrative Anticipation : "a desire to see what comes next ... (in) games with strong story elements" (pg.94-95)
- Development Anticipation : "postrelease content" (pg.99)
- Schedule Uncertainty : "short play sessions but engagement over months or years" (pg.100)
- Uncertainty of Perception : "difficulty of perceiving what's going on in the game space" (pg.101)
- Malaby's Semiotic Contingency : "the unpredictability of meaning that always (sic) accompanies attempts to interpret the game's outcomes" (pg.102)
Solver's Uncertainty is somewhat related to
Analytic Complexity and I suppose the real difference is one of degree, "puzzles" being relatively simple and solvable by logic. Costikyan alludes to this.
Quote:... if you look for solver's uncertainty, you'll find that it turns up as an embedded element in many games that are not primarily concerned with puzzles. For instance, in a turn-based strategy game, a player must, each turn, try to solve the puzzle of how to make the most effective attack with the limited resources available to him (sic) at that point in time, given the opposing disposition of forces and the terrain on which combat occurs.
(pg.77)
About
Chess, Costikyan says there are two sources of uncertainty:
Player Unpredictability and
Analytic Complexity.
Quote:... no Chess player -- human or computer -- can truly cope with the strategic depth of Chess, at least until the endgame when the reduction in the number of pieces prunes the decision tree drastically. (p/) The uncertainty in Chess lies not solely in the unpredictability of the opponent's move, but also in the player's uncertain ability to cope with the sheer strategic complexity of the game. The difficulty of analyzing the strategic options is so high that the player is rarely certain of the consequences of his next move.
(pg.39)
Player Unpredictability Costikyan's examples of player unpredictability in other games are actually not interesting for chess. Mostly those games allow far greater freedom to the players than a turn-based rule-based game like chess. Even "the unpredictability of the opponent's move" mentioned above is not really "unpredictability" for a decent chess player -- done right, our move was supposed to take into account our opponent's best possible reply even before we made it.
Unpredictability of the strong player
Every weak player knows the feeling when, having made the best move of which they are capable, they are not quite sure if their strong opponent will "find a hole" in their plan. It's critical to trust your own thinking, because you literally have nothing else. When your strong opponent plays an unpredicted move, it might be a mistake! An especially bad habit is to panic and trust your opponent more than you trust yourself. If you play like that, all your opponent's bad moves will go unpunished ... and strong players don't make many bad moves, so you have to seize on each one. Shake it off, pretend your opponent is a patzer, and just try to answer the unpredicted move with a good move of your own. Easier said than done, I know.
Unpredictability of the weak player
Contrariwise, the strong player is always expecting the weak player to "show their true colors" at any moment. They are just not sure when this might happen. As long as the strong player continues to make good moves this is not a big problem. Difficulties arise when the strong player tries to "provoke" an error. As a friend at the club was fond of saying after he made a good move, "Even a blind squirrel can find a nut." Even worse difficulties arise from suspicions of cheating, which is a real hazard these days.
Unpredictability as a strategy
I think the "universal" player can be unpredictable in a good way. It would be a lot of work to get there, but we can begin by taking a first step. If you excel at one aspect of chess (e.g. attack), try to pick up some unrelated skill (e.g. endgames). Not only will you become a better player, but you will also become more unpredictable. This is bound to have an effect on your opponents.
Analytic Complexity Quote:... they (Knizia and Meier, two famous game designers) want games in which players need to think about what to do, have to parse a complicated decision tree, and perhaps are uncertain, even as they make a decision, that it is necessarily the correct decision to make.
(pg.86)
Tal's quote about the deep dark forest comes to mind. But Tal was a genius and mostly we are not, so can we still use analytic complexity to our advantage? Well, the first thing to realize is, we are not going to solve all the problems over the board, it's just too hard. So at the very least, don't get into time pressure over it. Costikyan mentions this, although he is mostly concerned with how the poor time allocation affects the enjoyment of the other players. (The second sentence shows Costikyan doesn't know much about blitz chess.)
Quote:... analytic uncertainty often leads to analysis paralysis, the phenomenon whereby one player agonizes over his (sic) choices and delays the game for others. And of course, action games can support only modest levels of analytical complexity, because, by nature, a game that requires quick responses by players cannot also pose difficult mental challenges for them.
(pg.91)
The second thing about analytic complexity is that if the weak player is guessing and the strong player knows, the game can only end one way. The weak player should strive for the type of game where
both players have to guess, which means aiming for the type of complications only an engine could solve. Of course it mostly won't work, but it's worth a shot!
There were some other points made in the book that might be relevant for chess.
Hidden Information Quote:everything is a surprise when first encountered
(pg.93)
This is closely related to player unpredictability. Did our opponent play something new out of ignorance or out of knowledge?
Randomness Quote:Another common use of randomness is to break symmetry. That is, many games begin symmetrically, with all players in equal and equivalent positions, in order to ensure game balance; the problem with this is that unless symmetry is broken, all players are likely to value resources equivalently, adopt very similar strategies, and be able to judge quite easily where they stand relative to one another as the game progresses. It is desirable to break this symmetry, in order to provide uncertainty among the players about the other players' objectives, goals, and standing...
(pg.84)
I thought this one was interesting from an openings perspective. Certainly the asymmetrical Sicilian is a much better winning try than the Classical 1...e5; the problem is the Sicilian has too much theory and if white knows it then it might end in the dreaded "0.00". So playing
random openings might be a way to get asymmetry without all the theory. I think it's what Carlsen does with his openings experiments in rapid and blitz, but it would take strong nerves and a good sense of how far is too far.
Narrative Anticipation Quote:And it's even true of games as abstract as Chess; playing, we want to see how our opponent will react, how forces will ebb and flow over the course of play. There's a sort of narrative arc at work here, even if there is no direct connection to story. (p/) In terms of narrative, anticipation is the key; this means keeping the player uncertain as to how the story, or play arc, of the game will evolve. Many games fail on this score, at least after a time; Chess is an example. At some point, it is fairly obvious who is going to win a Chess game; the endgame is dull. The game does not conform to the classic narrative arc of increasing tension followed by release. Rather, tension builds to a point, and then slowly declines, as the board is cleared and the stronger player emerges, with a whimper of a coda as, often, the king is chased about the board until the inevitable checkmate. (p/) In general, games that have positive reinforcement cycles, in which success begets greater strength, suffer from endgames lacking narrative tension.
(pg.95-96)
This is pure psychology! Chess games don't have a narrative until they are over. I won a state championship once when the guy who was "winning easily" on board two started daydreaming about the playoff against me and fell into a stalemate trap, handing me clear first.
Pages: 1
Uncertainty in chess (Read 792 times)
