World of Tanks used a gaussian curve to determine weapon accuracy. Supposedly, two standard deviation of the shots were supposed to land inside the aiming circle.
In practice, there's always that 3-or-more-sigma shot that gets thrown multiple times outside the circle because the gaussian curve tapers off at infinity. IIRC this was fixed by re-rolling the shot if it missed hard enough.
After trying to optimize a cannon's mass for "mass of slugs thrown per ton per second" i've realized low-power autoloaders are much more efficient... for unmanned designs. This led to another question: "What's the mass overhead of a crew member?"
A crew member and his station add about 2.3 tons assuming an impossible 1 crew member per deck and a massless pressure vessel. In practice you're limited to 2.4~2.5 tons per crew member as those extremely long designs get prohibitively expensive length-wise.
Using a memory editor to remove the aluminum from the equation and finding out how much a crew member and his supplies weigh, you get this:
Apparently crew members are teenage girls who drink water while consuming aluminum, while letting gases they're supposed to breathe leak slowly for no apparent reason.
As of now, you don't have FOOF as an oxidizer ingame. Theoretically, you can add its chemical properties and reactions by editing the game's data through text files, assuming you've somehow managed to actually get your measuring equipment to resist the damn thing in real life long enough to get data on them.
In any case, monopropellants are used due to density and simplicity. They're denser than most bipropellants(not that big of a concern in space as you don't have to deal with drag, though hydrogen still isn't very viable for combat craft because it needs so much space for tankage it's nigh impossible to armor), and don't need complicated injection/ignition equipment to start them(hypergolics still need proper mixing despite spontaneously reacting when in contact with eachother, while more stable bipropellants tend to require an ignition source). The latter isn't represented in COADE as the game assumes you've got proper mixing and ignition systems for your thrusters already. On the other side, they're unstable and tend to have their tanks explode if provoked(say, by a high velocity projectile which managed to penetrate the tank) and are not very energetic(which limits the engine's exhaust velocity).
The problem i see with controlled homing is you're going to need more delta-v if you try to do corrections late into the flight path.
You'd get better performance with "relative velocity neutral homing" where the missile performs corrections as soon as errors appear but doesn't try to continuously accelerate towards its target. Controlled homing could be basically full homing switching to that once it gets within "target is likely to spend about as much delta-v throughout the intercept window as what i have left" distance.
That's kind of a mixed bag - you also have to take propellant cost into account. Mercury is the obvious choice for thrust-to-power ratios, but it's expensive.
For example, take this thruster. 100MW power in, 15 kN thrust out at 12.6km/s.
Lawson's laser drone uses 68.5kc worth of xenon to get 10.1km/s of delta-v. Getting 10.1km/s of delta-v with that MPD thruster requires 54 tons of mercury for a total cost of 918kc. That's nearly enough to double the drone's cost. The drone would weigh 98 tons. getting the same initial TMR would need 115.2kN as opposed to 78.7kN.