tom wrote:

19/11/2009    Improved randomness of game board initialisation

 I noticed the seemingly unrandom placement in Private games prior to your "improvement", but I'm not so sure it is improved.

See the game links below, no territory has been taken but all 3 have the exact same starting locations for the players.

Total Fog and Heavy Fog (Okay you can't see these one, but they're the same)

No Fog

Light Fog

No Fog (Seat position switched, so territories switched colors, but same exact territories).

I haven't tested this with Public games and haven't tried it with more players or anything, so it could just be a fluke with me and myself and quick invites to the same map or something.  Just thought you should be on the lookout.

(Yurog is me, we should never be in a Public game together, I only use it to check stuff like this .)

Good spot thanks Yertle. It turned out there is a bug in one of the RNG functions in the specific version of PHP that the site is using. I've changed the territory selection RNG to use a different function, it now appears to be working correctly.

Let me know if you spot any other discrepancies :)

Better reset all the rankings back to 0.

asm wrote: Better reset all the rankings back to 0.

I glanced at a couple Public 1vs1 games yesterday when looking at this issue and didn't notice it to the degree that I had in my examples, but it could be there, didn't see anything that gave a huge advantage though.  (Even if you are joking, still important to note IMO.)

ps this bug only affected boards with no starting scenario and no neutral territories which is a very small subset of games.