Patching Conics

[Starstrider42]

I've got a situation where I generate a hyperbolic orbit around a planet and would like to obtain the Sun-centric orbit that leads into it. If I simply create the orbit with an appropriate epoch, KSP will automatically patch the orbit, but the results will be wildly inaccurate -- unless the original orbit goes deep into the planet's SoI, the "corrected" orbit will miss it entirely.

Does anybody know how to solve for patched conics manually? Both the Orbit and PatchedConics classes have methods that look relevant, but I can't figure out what most of the arguments mean... (and they're not documented in Anatid's API).

Edited December 27, 2015 by Starstrider42
Incorrect statement; correction in later post

[linuxgurugamer]

I wonder if you are running into the floating point inaccuracies.  It sounds like you can create the orbit, but the floating point messes it up.

[sarbian]

The orbit code uses double so I doubt it. (not at home, so I can't help...) 

[Guest]

Is it possible that this is only visual?  Are you also taking into account that the planet needs to orbit around, meaning it will eventually be right?

[Starstrider42]

Actually, based on the image below my statement that KSP does even approximate conic patching seems to be wrong. My question still stands, though -- there MUST be an API method for doing this cleanly.