Timeline for Sandbox for Proposed Challenges
Current License: CC BY-SA 4.0
30 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 13, 2019 at 8:45 | comment | added | TwilightSparkle | Is this posted? I like it. | |
| Sep 24, 2019 at 23:10 | comment | added | trichoplax is on Codidact now | This allows leaving the tournament running long term and regularly posting the leaderboard with some joint places, which gradually become distinct places the longer it's left running | |
| Sep 24, 2019 at 23:08 | comment | added | trichoplax is on Codidact now | For deciding how many games to run in a tournament, if you don't need to fix that number up front you could use the same approach as Formic Functions: I settled on running 6 tournaments in parallel (well, sequentially and keeping track of totals, with each game being counted towards one of 6 buckets, in rotation). This means that if player A is above player B in all 6 tournaments, they have a >95% chance of belonging above player B. So you can display a leaderboard which is the sum of all 6 tournaments, and display players in joint place if the ordering isn't yet unanimous. | |
| Sep 24, 2019 at 23:02 | comment | added | trichoplax is on Codidact now | Another possibility is to place the player at a random position so they don't always have the same alignment with their opponent. So not just a random quadrant, but make each quadrant bigger than the player and position them at a random point within that quadrant. For example, 256x256 playing field, split into four 128x128 quadrants. Each player is a 64x64 region, positioned randomly in its 128x128 quadrant | |
| Sep 24, 2019 at 22:56 | comment | added | trichoplax is on Codidact now | If the playing field wraps, the player can be approached from any angle. If the boundary is closed, the player is shielded on 2 edges. Will the player be able to design knowing which 2 edges those will be? This would require rotating each player according to which quadrant they start in each game. Alternatively the player can have a consistent orientation but which 2 edges are shielded will vary from game to game. You could even vary both if you wanted to encourage highly robust solutions. (You will be placed in a random quadrant with a random orientation.) | |
| Sep 24, 2019 at 22:49 | comment | added | trichoplax is on Codidact now | I've been giving some thought to the middle ground between a language specific KotH and a language agnostic KotH. Agnostic makes it open to more challengers, but makes much more work for the challenge author. Specific is easiest for the challenge author, but excludes people who don't know that language. I like this idea of having the entry be a configuration that could have been discovered using any language, but can still be run on a single language controller. | |
| Sep 24, 2019 at 22:07 | comment | added | Alion | 1) The game should last a fair bit, so that there's enough time for stuff to happen. I'd say somewhere in the range of 10-50x side length. 2) As for the side length itself, I don't have an issue with 64. 3) I think the optimal way to find the overall winner would be to repeatedly choose 4 random entries per game until either a set number of games or some statistical threshold is reached. | |
| Sep 24, 2019 at 4:52 | history | edited | Jo KingMod | CC BY-SA 4.0 |
added 142 characters in body
|
| Nov 7, 2018 at 3:17 | comment | added | Veskah | You might want to specify the tourney structure, which I'm guessing will be round-robin style. Will standings be based on total points over the whole event or number of wins (i.e. 1st is 3 pts, 2nd is 2, 3rd is 1, or something like that)? Might shape the design choices. For the other points, 128x128 seems a good size, and you could try running both wrapping and non-wrapping as separate tourneys (but twice the upkeep unfortunately). | |
| Nov 3, 2018 at 1:35 | history | edited | Jo KingMod | CC BY-SA 4.0 |
changed to neutral colour
|
| Nov 2, 2018 at 23:12 | comment | added | user45941 | You should probably explicitly mention that GoL uses a Moore neighborhood (which includes diagonals), as opposed to a von Neumann neighborhood (which does not). | |
| Nov 2, 2018 at 21:35 | comment | added | Laikoni | I like the idea of introducing a neutral color. | |
| Nov 2, 2018 at 10:29 | comment | added | Alion | @Laikoni For clarity: we're talking about cell colors being chosen randomly, right? While I agree that this could generate large bias within a single game, I don't think it matters in the grand scheme of things. As for the solution (if this is to be considered a problem) - breaking the GoL rules seems odd. For example, a 3-color blinker would just disappear. How about adding an additional, neutral color in that case? It would behave the same way that all the other colors behave, but wouldn't give any points to anyone. A 3-color blinker would lose its outer colors that way, but keep existing. | |
| Nov 2, 2018 at 8:49 | comment | added | Laikoni | If they are all different colours, then the cell will be randomly picked from those colours. Given the nature of complex GoL patterns, I would suspect that one random cell choice can make a huge difference, so it might be nicer to keep the outcome deterministic and just kill such a cell. | |
| Nov 2, 2018 at 6:04 | comment | added | Bubbler | Oops, I didn't notice that. | |
| Nov 2, 2018 at 5:44 | comment | added | Jo King Mod | @Bubbler Cells outside the playing field will always be considered as dead, and cannot be resurrected. Though wrapping cells might be interesting... | |
| Nov 2, 2018 at 5:32 | comment | added | Bubbler | We need to clarify the behavior on the borders outside of the whole playing field. Does the grid extend indefinitely? Or does it wrap around? Or maybe simply hard dead cells (cannot turn to alive whatsoever) outside the given area? | |
| Nov 1, 2018 at 23:52 | comment | added | Jo King Mod | @user202729 It might not be intuitive, but it is possible. Very possible. | |
| Nov 1, 2018 at 23:48 | comment | added | Jo King Mod | @Alion I've fixed the image and changed the field size to 128, thanks. | |
| Nov 1, 2018 at 23:41 | history | edited | Jo KingMod | CC BY-SA 4.0 |
fixed stuff
|
| Nov 1, 2018 at 15:16 | comment | added | user202729 | @Alion "best" static pattern? I think how a pattern behaves also depends on its opponents. | |
| Nov 1, 2018 at 15:15 | comment | added | user202729 | I don't know if the challenge is interesting or not, mainly because it's very unintuitive to "program" in GoL. | |
| Nov 1, 2018 at 14:36 | comment | added | Alion | 1) Is it just me, or is the example picture screwed up? The distances to the dotted lines don't match up. 2) My computer would greatly appreciate it if the side length of the board was a power of 2 - the closest size to that is 128x128. 3) You could make the game last until the pattern settles down, i.e. starts being periodic. | |
| Nov 1, 2018 at 13:04 | comment | added | Alion | Unless you want to see the best static pattern, full stop. I can see that being fun as well - my comment was just a suggestion. | |
| Nov 1, 2018 at 13:03 | comment | added | Alion |
Fairy enough. I just think it's a missed opportunity. You could easily make it into a JavaScript competition, where programs just run multiple rounds of the currently planned game, having access to what happened last round, and thus being able to iterate on their patterns. You don't exclude anything - at the start, you're just going to get a lot of static entries (and I don't think anyone's going to be discouraged by the fact the they have to write return before their pattern).
|
|
| Nov 1, 2018 at 12:52 | comment | added | Jo King Mod | @Alion Your submission is a GOL pattern. I guess you could say you're programming in GOL | |
| Nov 1, 2018 at 12:36 | comment | added | Alion | So... where's the programming part? | |
| Nov 1, 2018 at 10:38 | history | edited | Jo KingMod | CC BY-SA 4.0 |
edited body
|
| Nov 1, 2018 at 9:07 | history | edited | Jo KingMod | CC BY-SA 4.0 |
changed to 50 by 50
|
| Nov 1, 2018 at 6:56 | history | answered | Jo KingMod | CC BY-SA 4.0 |