Timeline for Is it possible to design a two player game of skill with absolutely no luck?
Current License: CC BY-SA 3.0
24 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 25, 2017 at 1:37 | comment | added | DukeZhou | Thanks again for this answer, and please don't take my challenge as an attempt to diminish your correct and valid point. (Just had to work through the concept and understand the context;) It has been quite useful, particularly as I hadn't come across this idea previously, and have had to update my definition of deterministic games (still a work in progress) to discuss the inability to remove luck from unsolved games from the standpoint of strategy/optimal play. | |
| Aug 18, 2017 at 16:00 | comment | added | DukeZhou | I am upvoting because the language in the question was imprecise, and this answer is correct and valid from the standpoint of strategy (as opposed to game mechanics) | |
| Aug 18, 2017 at 15:59 | comment | added | DukeZhou | OK-I watched the Garfield lecture on this subject. There is some imprecision in that he is using plain language as opposed to formalize mathematical terms (per CGT). This definition of luck in games is in regard to "uncertainty of outcomes" and is purely a factor of intractability. (It is only a factor in non-trivial deterministic games, i.e. games where the solution is unknown to the participants.) From the standpoint of game mechanics, it is absolutely possible to create a game that involves no luck from a procedural perspective. | |
| Aug 17, 2017 at 21:37 | comment | added | DukeZhou | I think the division of opinions is partly philosophical and partly semantic: "of the forms of indeterminacy, which constitute luck?" For me, true randomness has to be involuntary, and luck is the result of factors beyond the player's control, where statistical analysis of probability cannot be applied, which is distinct from making random decisions and "hoping to get lucky in making the right decisions." Again, this may be a philosophical/semantic distinction. | |
| Aug 17, 2017 at 21:27 | comment | added | DukeZhou | It's a fair point, and Garfield seems to agree with you, but I still feel that it's a distinct form of indeterminacy, arising out of intractability, even if it represents a form of random number generation in this context. That said, worthwhile to bring up, and the questioner certainly does himself extend the definition of non-chance from the standard conception in relation to games. (The Garfield argument is similar to the million monkeys eventually typing Shakespeare;) | |
| Aug 17, 2017 at 21:06 | comment | added | Brilliand | @DukeZhou I'm talking about [1]: random number generation (imposed by player choice rather than game rules). As for a more optimal strategy winning more on average, that's standard for games involving both luck and skill; a game with no luck whatsoever would not require multiple games to determine the better player. | |
| Aug 17, 2017 at 20:28 | comment | added | DukeZhou | @GendoIkari AlphaGo heavily utilized Monte Carlo Search Trees, but if the decision-making process was purely stochastic, Lee Sedol would have won every game. | |
| Aug 17, 2017 at 20:21 | comment | added | DukeZhou | @GendoIkari It's technically true that if both players utilize random strategy, the outcome will be based on luck, per the Garfield categorization, but I think Game Theory has an answer. If one player consistently employs a more optimal strategy, they will, on aggregate, win more games. Take the example of classic, iterated Prisoner's Dilemma where one player chooses randomly, but the other player always betrays. The betraying player will always come out ahead of the random player. | |
| Aug 17, 2017 at 20:13 | comment | added | DukeZhou | @GendoIkari Great comment!!! What Garfield claims is theoretically possible and strikes me as true, but it seems to be more of a comment on intractability of non-trivial games. (I'll have to watch the lecture to get the full context, and I doubly thank you for mentioning it.) Garfield is unquestionably one of the most significant game designers in history, and studied combinatorics at University of Pennsylvania. Both factors add weight to his view. | |
| Aug 17, 2017 at 20:03 | comment | added | DukeZhou | Only the first two forms of indeterminacy constitute luck--the third form is a function of complexity, but does not render a game non-deterministic. | |
| Aug 17, 2017 at 20:02 | comment | added | DukeZhou | You're talking about indeterminacy as opposed to luck. Indeterminacy is introduced via: [1] random number generation (such as dice or even card shuffling, in the sense of random order); [2] hidden information (incomplete or imperfect information), and [3] intractability. The first is similar to quantum phenomena, where you can have a result not based on causality, the second is based on inability to confirm positions, opponent choices or decision-making process, and the third is based on the unsolvability of a game by the participants, whether human or algorithmic. | |
| Aug 17, 2017 at 19:54 | comment | added | Brilliand | @DukeZhou My argument is about a player (the opponent of the person who chose the "no luck" game) choosing to move randomly, thus introducing an element of chance (he may randomly move perfectly, or throw the game, or anything in between). Against a player of non-perfect skill, this results in the winner being decided by chance (albeit heavily weighted toward the player moving intelligently). | |
| Aug 17, 2017 at 19:18 | comment | added | DukeZhou | This strategy is really about introducing indeterminacy by attempting to confound the rational, non-random opponent. However, intractability of a game does not constitute luck being involved, merely the uncertainty as to whether a given move is truly optimal. (Winning play in unsolved games may only be deemed "more optimal". Objectively optimal play only exists in solved games, such as Tic-tac-toe.) Unsolved, intractable games such as Chess are still deterministic, regardless of whether the outcome can be predicted. | |
| Jun 5, 2017 at 20:46 | comment | added | Brilliand | @JustinTime Otherwise known as "play the opponent, not the game". I strongly suspect that that would still be too random for the asker's tastes. | |
| Jun 4, 2017 at 16:28 | comment | added | Justin Time - Reinstate Monica | @Brilliand My point is that even when people try to be random, it's usually (if not always) still possible to predict what they're going to do, if you know them well enough to be able to read them. To offset this inherent readability, they would need some exterior source of randomness that isn't influenced by their personality. | |
| Jun 1, 2017 at 2:54 | comment | added | GendoIkari | I agree with this. Richard Garfield (creator of MTG) has a great lecture on chance in games where he says basically the same thing, using chess as an example. He shows that chess does in fact have luck because a player making random moves could theoretically make the right set of moves to beat the best players in the world. | |
| May 31, 2017 at 21:59 | comment | added | Brilliand | @JustinTime I would argue that "humans are not truly random" is not a strict theorem, and even to the degree that it is true, it does not mean "humans are never random at all". Indeed, I would argue the opposite - that human behavior ALWAYS includes a random factor, simply due to the random environment acting on our bodies. A large part of skill is the ability to mitigate this randomness in yourself. | |
| May 31, 2017 at 20:35 | comment | added | Justin Time - Reinstate Monica | Assuming that exterior sources of randomness are against the rules (if the intent of the game is to eliminate all chance, then using an outside source to add chance is akin to cheating), then random actions will initially introduce chance (as you initially lose the ability to read your opponent), but won't reduce the game as a whole to chance (since your opponent can't just flip a coin, they'll have to think about what the "random" thing to do is, which makes them predictable; you just need to learn how to read their attempts at randomness). | |
| May 31, 2017 at 20:35 | comment | added | Justin Time - Reinstate Monica | My argument is also that even a "random" player will become readable over time, due to humans not being truly random unless they use an exterior source of randomness. | |
| May 31, 2017 at 19:47 | comment | added | Brilliand | @JustinTime My point is that it is possible for acting randomly to cause an unskilled player to play perfectly, hence "getting lucky". You argue that games with a random player can still involve strategy, and I agree; but the question is whether a game can be found with no luck whatsoever, and in that context, your argument is irrelevant. | |
| May 31, 2017 at 19:16 | review | First posts | |||
| May 31, 2017 at 20:35 | |||||
| May 31, 2017 at 19:14 | comment | added | Justin Time - Reinstate Monica | and a valid tactical choice as long as it falls within the game's rules. It may introduce chance into one side, but this doesn't necessarily cause the game as a whole to fall back on chance (no human can be truly random, so this strategy can and will backfire as soon as your opponent learns to read your "randomness"). | |
| May 31, 2017 at 19:14 | comment | added | Justin Time - Reinstate Monica | Not necessarily. The ability to read your opponent, and the ability to outsmart your opponent, can both be seen as components of skill and/or strategy. In this case, the player decides to make unreadable moves which confuse their opponent through sheer unpredictability, at the expense of their own ability to strategise and plan. Knowing when to discard strategy and act solely at random is a skill-based decision, which makes such confusion-oriented gameplay a valid tactic for outwitting an opponent (it's a mind game, basically, which causes them to doubt their ability to read you), | |
| May 31, 2017 at 19:08 | history | answered | Brilliand | CC BY-SA 3.0 |