Mitsu was constantly playing with the logic, equations, and curvefits (data tables) in StarQuest ECUs. For most years, the injector pulses occur once every "X" MAS pulses, where X is based on all the other sensors. The length of the pulse is also dependent on the other sensors. Without the MAS, the ECU can only assume certain things based on RPMs and TPS position, and the boost pressure sensor on the 86-earlier cars. In theory it has enough for a basic speed-density system but, until 86, Mitsu didn't really tune it much at all. Before 86, the limp-home stuff was basically enough to literally limp home... not really enough for the car to run right. In 86 Mitsu tried all sorts of stuff. In 86, Mitsu apparently played around with a speed-density type of setup, at least as the MAS-inop limp-home mode. That's why some 86s seem to run half-way decently without the MAS - it depends on what month a particular 86 ECU was built. Few 86s without the MAS however run perfectly... hesitation on mashing the throttle, excess fuel during part-throttle cruise, etc. In limp-home mode, Mitsu sacrifices fuel economy and emissions to some extent; the default limp-home mode is generally waayy rich for engine safety.
Even in the later years Mitsu kept monkeying with ECUs; that's partly why some 88/89 cars boost cut at different RPMs. In 87, Mitsu deleted the boost pressure sensor completely so all the speed-density limp-home stuff ended up in the trash. Note too that in the 85/86/87 timeframe Mitsu re-did several of their cars, including the Colts, to have a Starion-like throttle body injection system. Engineers concentrated on making regular mode (i.e. not the limp-home) equations/curves/tables for all the other cars... as opposed to spending time fine-tuning limp home modes.
Since the MAS is the main input to the ECU, a dead MAS pretty much keeps the car from starting/running. Except for some 86s.
mike c.