Maneuver planning of any kind is basing its start of burn time on the total TWR of every engine onboard the spacecraft, causing late starts 100% of the time for crafts pushing or pulling other docked vessels that have their own engines which are disabled. Does not seem to count engines that are physically separated by decouplers from the current stage, only every engine that is attached to the current stage wether active or not.

Example: interplanetary ship with 4 LVN engines, TWR of 0.2 executing a node of 2000m/s with a burn time of 30 minutes.
Payload is several small landers with the orange radial engines, landers connected via docking ports to the interplanetary ship. Engines have never been activated and are in a separate stage than the LVN engines.
Combined TWR of all the engines is 0.6
MechJeb will warp to the node assuming all of the engines are active and begin trying to execute it from 6 minutes rather than 15, resulting in over 20% (400m/s) deltaV being lost to steering loss.

(Numbers are approximate just used to illustrate an example, but they are not far from being correct)

I know the numbers are approximate, but a 2000 m/s burn with a Kerbin TWR of 0.2 will take 17 minutes, not 30 minutes: [check it out](http://www.wolframalpha.com/input/?i=(2000+(m+%2F+s\)\)+%2F+(0.2+*+9.8+(m+%2F+(s%5E2\)\)\)+in+minutes). So with these numbers the 6 minute lead time sounds pretty reasonable, especially considering the TWR will increase over the course of the burn (which MJ tries to account for).

But maybe that calculation doesn't mean anything, since we don't have the exact numbers. Was the burn in fact highly asymmetrical across the node? Did it really start around T-6 minutes and end around T+24 minutes? If so I'd be grateful if you could provide a .craft or .sfs file demonstrating the problem.

Note that the "estimated burn time" KSP shows next to the navball can't really be trusted until you actually start burning, because until then it often doesn't have a good sense of how much acceleration your craft has.

Hmm, doesn't the gravity used in the TWR calculation vary with altitude? I'm sure I've seen staging stats in the VAB showing 0.4 start TWR for my final stage, but by the time I get up to 300-400 km and am about to activate that stage, it shows something in the 0.8-1.0 range instead.

These days the gravity used in the TWR calculation shown in the dV stats and in the "Surface TWR" info item is the TWR calculated using the surface gravity.

Okay, I hadn't noticed that change. :)

Speaking of which, is it currently possible to show TWR based on a different planet's gravity, like in Engineer Redux?