Re: turn processing discussed in PBM Design group (yahoo groups)

From: lindahl@pbm.com (Greg Lindahl)

Date: Sat, 09 Mar 2002 02:45:54 +0000

In article <a6bksi02cga@enews1.newsguy.com>,
Rich Skrenta <skrenta@pbm.com> wrote:

>You can have both if the RNG is based on a good hash and if you seed the
>RNG with local data (say the turn #, the day #, the two character IDs,
>and the location ID).
>
>This way it would be random who went first, but reruns (even with
>altered events elsewhere that affected the global RNG seed) wouldn't
>affect this particular local outcome.

Yes, that's pretty good. And it doesn't depend on unnecessary stuff
such as the exact phase that you arrive, only the day. However, you do
have to watch out for things that *don't* involve random numbers in
the program, i.e. the before-mentioned "where am I in the location
stack" situation right now in Olympia. But it would be great for
battles, tradegoods, etc.

>Currently I only use this approach in Olympia for the trade cities,
>to avoid storing a full matrix of what city will accept what tradegood.
>The 50/50 chance city A will buy tradegood N is essentially MD5(A,N) % 2.

How interesting. We knew that this information was very persistent from our
testing, which seemed like a lot of work, but in fact it wasn't hard.

greg

References

Re: turn processing discussed in PBM Design group (yahoo groups)
lindahl@pbm.com (Greg Lindahl)
Fri, 08 Mar 2002 21:17:23 +0000
Re: turn processing discussed in PBM Design group (yahoo groups)
skrenta@pbm.com (Rich Skrenta)
Sat, 09 Mar 2002 00:26:58 +0000

Referenced By

Re: turn processing discussed in PBM Design group (yahoo groups)
lindahl@pbm.com (Greg Lindahl)
Fri, 08 Mar 2002 21:17:23 +0000
Re: turn processing discussed in PBM Design group (yahoo groups)
skrenta@pbm.com (Rich Skrenta)
Sat, 09 Mar 2002 00:26:58 +0000