I got into tabletop Battletech ages ago. I have a bunch of rulebooks, tech readouts, scenario books, universe background books, and other bits and pieces in a box. When I was young I played games with friends or at hobby shops, or just read the books for fun. Just saying that to make clear I am (was - just haven't done anything with it in ages) a big and long-time Battletech fan.
There is one simple problem with all real-time Battletech games that attempt to translate the tabletop stats directly to real-time gameplay: the fact that players can aim, whereas in the tabletop game hit location was a probability based on die rolls. (2D6 so there was a bell curve distribution from 2-12; some die rolls were pure random 1D6 but not hit location.) This is exactly what rythex described.
This problem goes all the way back to the original Mechwarrior game. My favorite mech in that game was the Phoenix Hawk - I just headshotted enemies with the large laser. Now this is realistic from a Mechwarrior roleplaying point of view, but it simply does not work when weapons and armor are direct translations from the tabletop game. There are other things that don't translate well, like weapon cycle timing and heat management but those can usually be pretty well adjusted. Players aiming cannot be adjusted without making a game very few gamers would want to play. In MPBT 3025 one common tactics was to have a group of Blackjacks, a very crappy mech in the tabletop game but the only one with AC/2 - the longest range weapon - in MPBT, line up across an open field. When an enemy was nearing range, begin to backpedal, and everyone concentrate AC/2 fire on center torso or an ammo-carrying area. A lance of crappy Blackjacks could take out half a lance before it was in range to fire back but that wouldn't happen in the tabletop game because hit location is randomized. Legging was also a common tactic. Those are just a few examples of how this problem makes real-time Battletech a mess.
Until a Battletech game randomizes hit location there will always be this problem, and the amount of kludging weapon damage or armor to overcome it is not practical and wouldn't be balanced either. MPBT was fun, and probably the best first person Battletech/Mechwarrior game I played, but I don't see there ever being a good real-time Battletech game that has a chance of being successful.