Dan suggested trying a model where armour is what you might call "hard" - that is, it doesn't get modified by weapons, but can be negated by weapons over a binary threshold. I'd expect this to make armour a lot more effective, since even a weapon slightly underpowered for the armour has to deal with the full value of it. As much of the time Monitors will have the better equipment, this would tend to make Monitors more effective in combat. It would also streamline combat because armour will either be ignored, or unmodified - no need for arithmetic.
As there's now no use for a modifier in this model, critical hits (half skill) would instead allow you to roll an additional die for injury and discard the worst result. Instead of a wounding roll, let's call this an Armour save - partly to differentiate them, and partly to highlight that the weapons are generally powerful enough to cause injury if not blocked by armour.
Heavy weapons, which I haven't really talked about yet, could roll additional dice, making them valuable against high-Wound enemies.
Greenclaw fires a volley of shots at some hostile gun-runners, hitting three times. As they've been loading crates, they're wearing exosuits that give Armour 12. Greenclaw's rifle is only Strength 7, so the suits protect against it. The gun-runners can roll a d20, and on a 12 or less their armour protects them.
One of the shots rolled a critical hit (half Greenclaw's skill). In this case, the gun-runner must roll two saves and discard the best result.
Hearing the racket, Toa pops up over a ridge and looses off a precious armour-piercing stun round. The Strength 14 shot blasts straight through the gun-runner's armour and releases powerful chemicals into her bloodstream, leaving her in a limp heap on the floor.
This model would transform some of the results from last time, because now rather than one factor (discrepancy between weapon and armour) there are two (is the armour better; if so, what is the armour value?). Note that in the examples below, I've tweaked Armour values to reflect the fact that it's rolled under on a 1d20, keeping it in line with Skill for simplicity.
Professor Rayner fights two thugs with armour no better than her pistol, and the same for theirs. With no armour save available, both are on 45%. It takes her five shots to take both down, and seven shots for them. Her odds are poor unless she goes first, but if so, she's got an edge.
Xerxes (skill 8) will wound a thug (skill 4) on 75%, while they're still on a 45%. He can take two and even three down in two rounds, but three will beat him before he gets the chance. Thugs with armour would be a much tougher proposition; they'd need at least 3 points of armour to withstand a pistol, but this would drop his chances to 66% and lose him the fight (this is, of course, why Xerxes carries a flare pistol).
Greenclaw (armour 7, rifle 7, skill 12) gets an armour save and ignores armour when fighting bandits (skill 8, armour 3-4, rifle 6). She should cause a wound 71% of the time, but it'll take 7-9 shots to defeat 4-5 bandits respectively (a little faster than the last model). She'd take 38 attacks from five, or 24 from four. They have only a 35% chance to wound her, so she can expect to survive 9 shots, or about one round.
This one has worked out less satisfactory than last time, perhaps. But as I said, I'm not sure it's actually appropriate for her to take on four or five bandits at once. This is sort of the burning question at the moment. If I do want that sort of thing, then I suspect I need to seriously consider upping the number of wounds Monitors have, because the undamaged/unconscious threshold is very sharp right now. This isn't a maths question, but a genre one, and right now I don't have a definite answer.
My inclination is that no, I don't. In that case, this isn't a problem - from the maths above, two Monitors of Greenclaw's competence, or her and a couple of less accomplished fighters, should be able to handle the bandits. The kind of rabble that Greenclaw should be able to handle on her own would be less accomplished, and probably better modelled as a mob. I will need to consider how to make those kinds of DMing distinctions intuitive.
Ukala (armour 10, skill 16, rifle 8) would like to fight several mercenaries (armour 6, skill 12, rifle 7) at once. She gets a hit with 86% probability, while they're on a 39%. Essentially, she can take nearly one merc per shot, but can still only withstand 7 shots, which means she can't afford to take fire from more than two mercs simultaneously - she could just about take four if she gets the drop on them and rolls well early on, but could equally well be doomed.
The Kargbeast has four Wounds, armour 10, str 9 and probably melée of about 15. The Kargbeast has a substantial advantage once it gets into combat (there's no way Ukala has a Strength 10 melée weapon, and she's probably got a lower melée skill), which means Ukala needs to get some shots off before combat starts. She has a 50% chance of wounding here, and so would need eight shots to bring it down - she won't get that. However, if she can inflict a wound or two she's got roughly even odds of taking it in melée.
Toa, providing he has a rifle capable of penetrating robot armour, can effectively take out a robot with every shot. He might well be much less effective against other targets, though. If he has military armour at about a 10 (better than most weapons can handle) he can survive about 10 shots from Skill 10 robots - less than last time, actually, and I'm increasingly thinking that I've misjudged armour values. This means he can realistically fight maybe three such robots if he gets lucky. Of course, if he's getting to fire first, he can take five on.
A high-grade military bot might have armour 16 in this system - some of the heaviest around, and more than Toa's rifle can handle. Toa can cause a wound around 26% of the time, calling for 12 shots. This means it will take him about six rounds to defeat the bot, and it's likely to beat him first. This is, of course, ignoring any side-effects of using a shock rifle on a robot - something I'm considering implementing.