[NeonSamurai]

Maybe it has been improved then. I'm still pretty out of date with KSP as I'm still trying to finish my interstellar quest game from 23.5, so I wouldn't know how well MJ performs in 1.04.

In 23.5 launch guidance was pretty horrible using a realistic atmosphere (FAR) as its mid flight course corrections were very sharp and would frequently cause my rockets to flip out of control (even with fins and gimbaling engines). Turning off corrective steering didn't help a whole lot either as MJ's control imputs were still mostly all or nothing. The only thing that did help was carefully editing the ascent path by hand, which I had to do anyways as the default profile was too aggressive and inefficient in FAR anyhow. Limiting TWR didn't work in FAR either (and probably still doesn't) as MJ couldn't properly read it when using FAR. The reason MJ got away with the sloppy aggressive controls, was because stock KSP's atmosphere was like flying through sludge.

As for thrust limiting, personaly I don't do it using throttle control (MJ or by hand), I do it in the VAB (kind of in the same manner as real rockets, as most rocket engines do not have much if any in the way of throttle control). I intentionally set the engines (or add enough fuel/payload) so that the rocket's starting TWR is about 1.1-1.4 (including solid fuel boosters) depending on how aerodynamic it is. It works really well in a more realistic atmosphere as it keeps the rocket slower in the lower thicker atmosphere, but the TWR picks up as the rocket expends fuel and stages, so that in the upper atmosphere (35k+) when I am accelerating near horizontally, to orbital speeds, my TWR is typically 2.0 or higher.

This is also in line with most real space craft designs. Most real life rockets typically have a launch thrust to weight ratio of 1.05 to 1.3 (1.2 is the average) with a full payload, and an upper atmosphere (~60km+) TWR of around 2-3 (or higher if unmanned as going beyond 2.5 gets very uncomfortable because of the g forces).

I've often seen on the KSP forums people claiming that the ideal thrust to weight ratio for terminal velocity is 2 in ksp. They are wrong. The reason why, is that the value for terminal velocity is not a static number and depends highly on the aerodynamic profile of the rocket. A rocket with high aerodynamic resistance would have a much lower terminal velocity value compared to a rocket with low aerodynamic resistance.