Having lower to-hit roll than your opponent is better because you hit more often than they do. The percentages are based on the curve of 2d6 dice. I'm not Demiurge and I don't know how I can express the idea more simply than that.
Sorry about the long response, but yes, you are explaining exactly what I had in mind.
You ignore some of the nuance of the game, but basically you can pretend that the average to-hit numbers that your mech encounters are a multiplicative damage modifier, and you can pretend that the average to-hit numbers that your mech generates for enemy gunners is a multiplicative armor modifier.
Assault mechs have enormous raw firepower and heavy armor (or at least they had *better*), but they generate poor to-hit modifiers relative to their faster-moving opponents, which means that they miss a lot more and get hit a lot more, which blunts the relative advantage of all those guns and all that armor. In the very worst cases, it means that assault mechs aren't giving you very good value.
With 3050 technology, the Timberwolf is at or very close to the optimal combination of speed, armor and firepower. But with 3145 tech, I contend that it's possible to do much better, and the result doesn't look much like a Mk IV.
The Mad Cat 1.0 has dramatically more interior room, so for carrying crit-intensive weapons it works better (not that it really does so in any canon configurations).
Ironically, it's all the post-3050 clan tech that tends to be crit-heavy. So the original Mad Cat is substantially more compatible with all the new goodies than its successor is.
Not that anyone cares, 3050 clan tech kills everything plenty dead.
This thread is not for discussing Demiurge's design, it's for Mad Cat MKIV versus Mad Cat.
I wasn't intending to derail, just explain average survivability/movement curves.