MightyPork
/
sigmar
1
0
Fork 0
Code Issues 1 Pull Requests Releases Activity

update readme

Browse Source
recursive
Ondřej Hruška 6 years ago
parent 4240282b17
commit 7d8c1bfc4b
Signed by: MightyPork
GPG Key ID: 2C5FD5035250423D
1 changed files with 8 additions and 8 deletions
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Show Stats Download Patch File Download Diff File
  1. 16
      README.md

16
README.md
Unescape View File

@ -94,7 +94,6 @@ or changed without updating this reference. Search for `SettingsStorage`.
``` ```
{ {
log: 'info', // default log level log: 'info', // default log level
allowTemplateAugmenting: false, // sloppy mode
retryTemplate: 30, // retry attempts within one template retryTemplate: 30, // retry attempts within one template
attemptTemplates: 50, // number of templates to try on failure attemptTemplates: 50, // number of templates to try on failure
svgEffects: false, // fancy mode svgEffects: false, // fancy mode
@ -125,19 +124,20 @@ Algorithm Quirks
Every board must be solvable, otherwise it wouldn't be much of a game. Generating a valid board Every board must be solvable, otherwise it wouldn't be much of a game. Generating a valid board
turned out quite a bit more challenging than I thought. My algorithm is based in some heuristics turned out quite a bit more challenging than I thought. My algorithm is based in some heuristics
and good deal of luck. and luck.
To make things more fun, besides the marble matching rules, the board must be laid out in one of To make things more fun, besides the marble matching rules, the board must be laid out in one of
several pre-defined shapes of exactly 55 tiles. The algorithm can sometimes get itself into a dead several pre-defined shapes of exactly 55 tiles. The algorithm can sometimes get itself into a dead
end while placing marbles on the board. I experimented with backtracking, but the current version end while placing marbles on the board. After some experimentation I settled on a recursive algorithm
simply retries the whole thing with a different random seed. There are configurable retry limits with a heuristic and backtracking. There are configurable retry limits and max recursion / retry
as a failsafe. If the given template fails many consecutive times, the algorithm opts to switch to counter as a failsafe. If the given template fails many consecutive times, the algorithm opts to switch to
a different template. Some templates are harder than others, and some random seeds just seem to a different template. Some templates are harder than others, and some random seeds just seem to
have a problem. have a problem.
A workaround I added is called Sloppy Mode. When the algorithm can't place a marble, it may choose I added a few extra rules to the generation algorithm to make the boards harder: Vitae and Mors should not
to add a new tile to the template, trying to keep the overall shape as tidy as possible. This may appear as the first few marbles to remove, since they tend to eliminate any choice; and metals start a few
hurt game difficulty, but is generally much faster than retrying over and over. levels deeper still. This gives the player a lot more room to make a mistake in the beginning.
At least you can enjoy the Undo and Try Again buttons!
If you're curious about the inner workings, open dev tools and enable debug logging with the `debug=1` If you're curious about the inner workings, open dev tools and enable debug logging with the `debug=1`
GET parameter. GET parameter.

Write Preview
Loading…
Cancel
Save