> I for one find this to be a very interesting topic. I would enjoy knowing how arcade > hardware generated random numbers (it would be different for every system, I imagine, > but there are probably enough similarities to be able to discuss how it works in > general).
To clarify: in most cases, this random number generation is internal to the *software* on a large majority of games. So the techniques used were likely similar to what was done on home computers in the early 80s.
A few later systems do have actual RNG hardware, specifically Sega System 32 and some of the later Neo Geo cartridges. The NG generator had a reversible pattern and is emulated correctly (I believe it's a LFSR, but don't hold me to that). We couldn't find the pattern for the Sega one, but nobody ever complains about those games on this topic > So if Don Hodges wants to share any juicy tidbits on this topic I would really love > to hear them!
Ditto.
The software-only RNGs likely are seeded by the frame count between boot and coin-up or other user inputs, so it's quite possible in that situation (especially if the RNG algorithm in the game software isn't very good) that MAME would behave differently from a cabinet that had been powered up for weeks or months at a time. This is one reason CAX is a nice sanity check on the emulation - the games are on only for the length of the show each day so they're in similar-to-MAME condition w.r.t. things like that, especially in the morning.
|