Quoted from barakandl:Bally games do everything with two PIAs, WMS needed four, and on top of that Bally games can have more complex rules. How Bally multiplexes the U10 PIA is genius, it does switches, lamps, and displays bouncing between different PIA modes.
I think this is because Bally invested in, and managed to hire, programmers that knew what they were doing. AFAIK williams' first programmer (system 3) was an electronic organ programmer, and quit when he got a job back in that industry. Until Williams hired Eugene Jarvis and Larry Demar, they didn't have a clear vision/person able to drive the development to what it became. Eugene Jarvis because he made stuff 'cool', and Larry Demar because he made stuff structured.
Bally's software sophistication was dominant until Stern mpu200's pigs interpreter, which blows away Bally's straight 6800 assembly in terms of what you can do with the game without writing a bunch of stuff to compensate. The Centaur/Flash Gordon era bally tried to graft things in to do lamp effects, etc. but it wasn't as easy to program as pigs.
As soon as Williams went system 7 with Larry Demar's version of a virtual machine, forget about the sophistication of the other systems. Far superior in every way (software wise). Much more flexible and designed with extensibility in mind. Of course, required way more resources than the earlier stuff.... the framework Larry set up though carries through to the end of williams pinball wpc95, there are elements of system 7 in every system since then (although they abandoned the interpreter at the end of system 11). The multitasking switching is the same system though. (extended in wpc). (all this not counting the 2 pinball 2000 released games, no idea what they did in terms of any of this...)