My program "cheats"

[Mr Nabutovsky]

Hello,
Several monthes ago I developed a program (Agushka Backgammon)
for Palm which plays much better than any other PDA backgammon
program. However, user start to write in their reports that my
program cheats. One of them even compared it to Devil, another
one emailed me that my program should be banned.
My program allows to see or change the RNG seed at any time,
and I published the source of my RNG. How can I prove that
my program doesn't cheat?

[Don Hanlen]

There are two answers to your question. First, you can publish the
source code, and allow intelligent people to evaluate it -- that will
enable intelligent users to ascertain whether your program cheats or
not. Second, there is nothing you can do. There is a type of user
who will think your program cheats as long as it can beat them, or
perhaps if it can't -- you can't do anything to satisfy them.

--
don
dha...@GoCougs.WSU.edu

[Ric Gerace]

You don't have to prove it. If you write a strong program that plays well,
then good players will recognize that; most sensible players will recognize
that. IMHO that's where you should put your energies and skills. These
people who complain about programs cheating .... don't waste your time and
energy on them.

"Mr Nabutovsky" <ba...@pair.com> wrote in message
news:394BFF...@pair.com...

[André Nicoulin]

Hi,

You could implement a feature similar to the one we did into Snowie: put an
option in your program that will create a note (a standard Palm Note) which
contains the next coming 1000 rolls of both sides. After having played
against the machine, a user could go and verify that the rolls were the one
given in the note file. Of course, some people will look at it before
playing, to make sure they win, but who cares.

We do it in this way: whenever you change the seed, you can generate this
note. In your program you would
1. set the seed
2. generate the next dice
3. re-set the seed, such that the same sequence starts again

André

"Don Hanlen" <dha...@oneworld.owt.com> a écrit dans le message news:
8ihufv$foi$1...@news.owt.com...

> There are two answers to your question. First, you can publish the
> source code, and allow intelligent people to evaluate it -- that will
> enable intelligent users to ascertain whether your program cheats or
> not. Second, there is nothing you can do. There is a type of user
> who will think your program cheats as long as it can beat them, or
> perhaps if it can't -- you can't do anything to satisfy them.
>
> --
> don
> dha...@GoCougs.WSU.edu
>
> In article <394BFF...@pair.com>, Mr Nabutovsky <ba...@pair.com> wrote: