Timeline for Prisoner's Dilemma v.2 - Battle Royale
Current License: CC BY-SA 3.0
67 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Apr 13, 2017 at 12:39 | history | edited | CommunityBot |
replaced http://codegolf.stackexchange.com/ with https://codegolf.stackexchange.com/
|
|
| Apr 13, 2017 at 12:19 | history | edited | CommunityBot |
replaced http://math.stackexchange.com/ with https://math.stackexchange.com/
|
|
| Apr 11, 2015 at 22:02 | history | rollback | Martin Ender |
Rollback to Revision 12
|
|
| Apr 7, 2015 at 17:14 | history | edited | Joe Z. | CC BY-SA 3.0 |
deleted 26 characters in body
|
| May 14, 2014 at 22:09 | vote | accept | Joe Z. | ||
| May 14, 2014 at 22:09 | history | edited | Joe Z. | CC BY-SA 3.0 |
Posted score results.
|
| May 11, 2014 at 2:51 | comment | added | Joe Z. | That is the Nash equilibrium. The optimal strategy in practice is different. | |
| May 11, 2014 at 2:43 | comment | added | justhalf | " If the game is played exactly N times and both players know this, then it is always game theoretically optimal to defect in all rounds.". So it proofs my assertion? | |
| May 10, 2014 at 14:30 | comment | added | Joe Z. | @justhalf If you've actually read any resarch on strategies in the iterated prisoner's dilemma, you'd know what you're saying is false. The Wikipedia article is a good place to start. | |
| May 10, 2014 at 1:24 | comment | added | justhalf | Actually I don't understand this challenge as a challenge. It's more like a playground to me where people can submit answers to play. Because there is absolutely no dilemma here, according to the rules. Each time you "take 1", overall point will decrease by 1 and your point increase by 1. If you don't, overall point will stay the same. So if you want to maximize your points, always take 1. If you want to maximize overall points, never take 1. It's as simple as that. Unless you can come up with some interesting winning criteria, I don't see anything interesting in this challenge. | |
| May 9, 2014 at 23:50 | comment | added | PleaseStand | Will my "Sort the set of victims and remove duplicates" change to the scorer be used in the tournament? | |
| May 9, 2014 at 20:09 | comment | added | Geobits | @Trimsty When the challenge first went up, the code for the plants was not shown. Through the entire coding phase, we could not see other answers. Dupes could have happened purely by accident by choosing the very obvious greedy strategy. | |
| May 9, 2014 at 20:07 | comment | added | cjfaure | @Geobits I don't see how people submitting duplicates would be allowed, the goal is to make an ai, not choose one :P | |
| May 9, 2014 at 20:03 | comment | added | Geobits | @Trimsty It doesn't look like anyone did submit one, but I don't see that in the rules. I thought about doing it myself, but went with something else out of curiosity's sake. I agree that being Greedy is probably the best overall strategy if your goal is to look out for yourself. We don't care about the overall point total for this challenge. | |
| May 9, 2014 at 17:43 | comment | added | cjfaure | @justhalf A critical part of this competition is that you're not actually allowed to submit a Greedy. You have to find something else that works. | |
| May 9, 2014 at 15:27 | comment | added | justhalf | How can greedy not maximizing damage? If we take from other player, the other player can only get -1 or -2, while if we don't take, the other player can get 1 or 0. Obviously taking 1 from other player will maximize the damage. And so Greedy will never lose. And it will almost always win, unless all the opponent are also Greedy, like you said. | |
| May 9, 2014 at 14:57 | comment | added | cjfaure | @justhalf The Greedy doesn't maximize damage to the opponent, or try to mess with other AIs. For instance, a game with only Greedy bots can never gain points. | |
| May 9, 2014 at 6:48 | comment | added | justhalf | I don't understand. How can The Greedy lose under the game rules? | |
| May 9, 2014 at 0:00 | history | edited | Joe Z. | CC BY-SA 3.0 |
added 522 characters in body
|
| May 8, 2014 at 23:34 | answer | added | PleaseStand | timeline score: 3 | |
| May 3, 2014 at 17:09 | history | edited | Joe Z. | CC BY-SA 3.0 |
Added the judge program.
|
| May 3, 2014 at 13:06 | comment | added | cjfaure | @JoeZ. Cool. Thanks. | |
| May 3, 2014 at 13:01 | comment | added | Joe Z. | It's alright. You can post all the alternate entries you want after the tournament is over, by the way. | |
| May 3, 2014 at 13:01 | comment | added | cjfaure | @JoeZ. Oh, okay, sorry, didn't see that comment for some reason. | |
| May 3, 2014 at 12:55 | comment | added | Joe Z. | @Trimsty No. | |
| May 3, 2014 at 11:08 | comment | added | cjfaure | @JoeZ. Is it okay to have multiple entries here? | |
| May 3, 2014 at 3:46 | history | edited | Joe Z. | CC BY-SA 3.0 |
added 126 characters in body
|
| May 2, 2014 at 21:19 | answer | added | cjfaure | timeline score: 1 | |
| May 2, 2014 at 20:14 | history | edited | Joe Z. | CC BY-SA 3.0 |
added 142 characters in body
|
| May 2, 2014 at 17:48 | answer | added | kitcar2000 | timeline score: 1 | |
| May 2, 2014 at 16:52 | answer | added | Joe Z. | timeline score: 1 | |
| May 2, 2014 at 16:37 | comment | added | Joe Z. | Also, remember that all edits are logged, and they need to be in order to count. If a person posts 30 edits within 5 minutes of each other in order to have a choice in IDs, 1) they won't all show up in the edit history because of the grace period, and 2) it would definitely be very suspicious and I'd probably ask them about what those 30 edits in a row were all about. | |
| May 2, 2014 at 16:35 | comment | added | Joe Z. | I did the "any hash" thing so that people who discover errors in their latest revision could revert to an earlier one. | |
| May 2, 2014 at 16:30 | comment | added | PleaseStand | My concern is then with "If the SHA256 hash of your posted source code does not match any hash that you posted before the deadline [...]". The hash should match the last one posted before the deadline so the player who posts his or her source code last cannot choose which program to post based on the player ID assignments (e.g. one that assumes Tit-for-tat has ID 1). | |
| May 2, 2014 at 16:18 | comment | added | Joe Z. | @PleaseStand Yes. If I do it before then, the number of entrants might change. (And if I do it after then, I might accidentally run the tests more than once and bias myself toward a certain result.) | |
| May 2, 2014 at 15:19 | comment | added | PleaseStand | When will player ID assignments (which could potentially affect program behavior) be posted? At 2014-05-09 00:00 UTC (the beginning of the source code posting window)? | |
| May 2, 2014 at 13:48 | comment | added | cjfaure | @JoeZ. Kk. Working on my answer now. | |
| May 2, 2014 at 13:17 | comment | added | Joe Z. | I'm writing a judge program for that right now. Sit tight. | |
| May 2, 2014 at 10:07 | comment | added | cjfaure | How would we test our programs (say, against the plants)? | |
| May 2, 2014 at 10:02 | comment | added | cjfaure | @ZoveGames It would be, but then everybody would use that strategy and no points would ever be gained. | |
| May 2, 2014 at 3:42 | comment | added | DankMemes | I don't understand. Is there some sort of penalty for everyone taking 1 all the time? Wouldn't it be the most advantageous to always take 1? | |
| May 2, 2014 at 2:21 | history | edited | Joe Z. | CC BY-SA 3.0 |
Added tournament scores for the four plants.
|
| May 2, 2014 at 1:25 | comment | added | Joe Z. | @Kevin: Yeah, most likely. | |
| May 2, 2014 at 1:23 | comment | added | Joe Z. | @dmckee: Yes, that is valid, as it will still always return the same output for the same input. The person who suggested that rule suggested it for exactly the concern you mentioned, so it is definitely allowed. | |
| May 2, 2014 at 1:21 | comment | added | dmckee --- ex-moderator kitten | Devil's advocate question: suppose I write a solution which seeds it's PRNG from R during the first round but passes the state of the PRNG forward from round to round (rather than re-seeding with the new R). Still "deterministic" under the rules? | |
| May 2, 2014 at 1:17 | comment | added | Kevin | There are going to be a lot of upvotes for answers once the deadline passes, aren't there? | |
| May 2, 2014 at 0:49 | answer | added | Geobits | timeline score: 2 | |
| May 1, 2014 at 23:56 | answer | added | Ypnypn | timeline score: 1 | |
| May 1, 2014 at 23:49 | comment | added | Joe Z. | I've edited in something that allows you to pass one line of data onto the next round. Does that solve your problem? | |
| May 1, 2014 at 23:48 | history | edited | Joe Z. | CC BY-SA 3.0 |
Added in an option to pass internal state in a recordable format for internal use between rounds.
|
| May 1, 2014 at 22:57 | comment | added | PleaseStand | So because R changes each round, if my program were to choose one of several strategies at random, it would have to "forget" which one it picked, instead having to infer that from the history? | |
| May 1, 2014 at 22:45 | comment | added | Joe Z. | 4) If you take 1 from yourself, the only result is that your own score decreases by 1, so I'm not sure why you'd want to do that. But the answer is no, you cannot take 1 from yourself. | |
| May 1, 2014 at 22:44 | comment | added | Joe Z. |
@PleaseStand 1) To prevent collusion between solutions by the same user, each contestant may only enter one program for the actual tournament. However, in the after-tournament stage, feel free to post as many alternate solutions as you want, as long as they are distinct strategies. 2) Your program will get a different random value of R each round, but each program will get the same value of R for that round. 3) I haven't written the official scorer's code yet, so you'll have to write your own until I get that done.
|
|
| May 1, 2014 at 22:37 | history | edited | Joe Z. | CC BY-SA 3.0 |
Removed the description thing. It doesn't seem right with the whole "hide everything until the tournament's over" thing.
|
| May 1, 2014 at 22:32 | comment | added | PleaseStand | 1) May I only enter a single program? 2) Will my program get the same value of R in each round? 3) Is the official scorer's code available, or do I have to write my own for testing? 4) Does "some other player B" mean I cannot take 1 from myself? | |
| May 1, 2014 at 21:58 | history | edited | Joe Z. | CC BY-SA 3.0 |
Replaced all the spaces with tabs, and fixed a debug output problem with The Envious.
|
| May 1, 2014 at 21:11 | comment | added | Joe Z. | @Geobits: The terminated output doesn't matter, as long as it's a list of numbers separated by spaces. | |
| May 1, 2014 at 21:09 | comment | added | Joe Z. | @ypnypn: When you submit your code, email your source code to me and I'll edit it into your answer for you during the code-submission period. | |
| May 1, 2014 at 20:55 | history | tweeted | twitter.com/#!/StackCodeGolf/status/461972256227606528 | ||
| May 1, 2014 at 19:53 | comment | added | Geobits |
Should the output be terminated with \n, \0, something else, nothing, or does it matter?
|
|
| May 1, 2014 at 19:35 | comment | added | Ypnypn | Can we have a little more time to submit the actual code? If I'll be busy for those twenty-four hours, am I disqualified? | |
| May 1, 2014 at 19:17 | history | edited | Joe Z. | CC BY-SA 3.0 |
Included the source code for the four plants.
|
| May 1, 2014 at 19:13 | comment | added | Joe Z. | I suppose not. Lemme edit them in right now. | |
| May 1, 2014 at 19:12 | comment | added | Geobits | Would posting the source to the control program and/or plants hurt anything? We know what they do anyway, and I'd prefer to be able to test against something without writing five extra programs. | |
| May 1, 2014 at 18:59 | comment | added | Joe Z. | I've named all the plants after vices, since I expect that this competition will be more vicious than the previous one, and I thought I'd theme the plants' names to reflect that. | |
| May 1, 2014 at 18:56 | history | edited | Joe Z. | CC BY-SA 3.0 |
Strategy descriptions are optional.
|
| May 1, 2014 at 18:46 | history | asked | Joe Z. | CC BY-SA 3.0 |