10 Seconds Random Game Generator

[pyro...@yahoo.fr]

Hello guys !
I created the first version of my 10 Seconds Random Game Generator !

Using the awesome tool Puzzlescript, you can now create a playable experiment just by entering a title !

I hope you'll enjoy. :D

The link :
https://dl.dropboxusercontent.com/u/69344892/puzzlescriptgen/index.html

[sftr...@gmail.com]

Interesting! Here's my game, Giant Pig: http://www.puzzlescript.net/play.html?p=6941681

[Stephen Lavelle]

ahaha, cool - if you want it to work seamlessly with puzzlescript, you could have it submit a gist to github, and send that link to puzzlescript in the url (check out the sourcecode to puzzlescript on github if you want to see how - it's all client-side javascript)

On Fri, Oct 11, 2013 at 9:40 PM, <sftr...@gmail.com> wrote:

Interesting! Here's my game, Giant Pig: http://www.puzzlescript.net/play.html?p=6941681

--
You received this message because you are subscribed to the Google Groups "PuzzleScript" group.
To unsubscribe from this group and stop receiving emails from it, send an email to puzzlescript...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Stephen Lavelle]

One thing - most of the levels seem only one move away from completion?  I would rather it generate 100x levels, if that's the case, rather than just 1 or two.

You allow for keys for generation, but they don't generate unique levels. I guess it doesn't matter.

[xavi...@gmail.com]

It's got some issues, notably the fact that it generated an impossible game.
http://www.puzzlescript.net/editor.html?hack=6946318
I did go through the rules though and it was moderately interesting trying to figure out what they all did. I commented them and it pretty much comes down to having complete stupidity in the rules in a variety of ways. My favourite part though was the last two which included the sole win condition.

[< Chlorine] -> [] sfx2
[< Player|> Player|right Chlorine] -> WIN

That is a serious problem deleting all moving chlorine and then saying you need one to win straight after.

[pyro...@yahoo.fr]

Hello

The levels aren't all the same, they are also random-generated.

The fact that rules can create impossible games is normal. It's random-generated rules. Not someone with an accurate idea of a real puzzle game.
I would say it generates much playable experiments than games.

Actually, if you have the right seed, you can generate any Puzzlescript game.

[xavi...@gmail.com]

It could provide similar level of analysis as I did at the very least. The most notable and important things being 2 rules that should apply 99% of the time anyway.

1. Have some kind of match on the left hand side, empty matches and matches only containing no don't count.
2. Don't require something moving if it can't have moved in that direction due to a previous rule.

Those two rules would likely make it much less likely to make mistakes, at the same time rule number 2 should always apply and there isn't very many times to break number 1 since it results in filling everywhere with something which is very rarely needed.

[Michael Cook]

I think even these would be a bit general at first. I'd recommend going the other way - start with rules that you know function on their own as sensible fragments of a PuzzleScript game. So you have a bag of 100 rules like:

[> #T1 | #T2 ] -> [> #T1 | > #T2]

And replace the #T markers with types you've randomly declared above. You grab 5 of these from the bag of 100 rules and use that instead of randomly generating them. Then you can slowly start to make the process a bit more intelligent - breaking up the fragments on either side of the ( -> ), say, or only supplying the LHS of a rule and getting the system to produce mutations of it that become the RHS?

Definitely interesting though! If I get the time I'd like to tinker with something like this too.

[rikki...@gmail.com]

Hi,

Do you still have a copy of this I could look at? I'm rather intrigued!

Cheers,
Rikki

[21hol...@blan.org]

[Want - Diverse Content]

I want to try this, but the file no longer exists.

[sohyb.k...@gmail.com]