Stars, then, seem to matter in calculating a jump, but they are not essential. We'll move on with the assumption that an empty hex jump is always more difficult than a star jump and sometimes impossible.
That said, what's the deal with stars, and what about the missing stars ?
See, thoughts about empty jumps came about as a result of too much thinking about the missing stars in the empty hexes.
We know that some kind of complex, time limited and location dependent calculations need to be made for jump. I have to say that I've never been overfond of the argument that these are mostly due to differing velocities. Multi vector calculations are doable now, abiet not by hand, and can reliably solve for any number of vectors (in fact, they can be rotated thru any number of dimensions relative or not to any one or set of the vectors- it's the basics of factor analysis). If you want a more in depth explanation of why I think its unlikely, read the * below. Otherwise we can move on.
My theory (cough,cough) is based on this: all points in jumpspace, except for the 100d situation, map to points in realspace, this is explicit. However, this mapping need not be one to one, or exclusive.
In fact, I'd suggest that the key nature of jumpspace is that its spaces routinely map to potentially many other internal jumpspace points. The key then is finding out what point in JS is the one you want -that maps to the point corresponding to the jumpspace point of exit; essentially you are coesisting in both points. The entry takes lots of power, and the exit works much like particle decay -your jump field (which is holding you in the correct location) collapses somewhat randomly, and dumps you out at your exit point (presumably something about the enntry process also fixes which realspace mapped point you return to). One cannot choose an exit point within 100d of a major since those areas do not exist for jumpspace mapping. Precipitation is simply the result of the exit map being displaced.
If one likes jump interruption (ie a mass intersects your vector , you pop out there) it's a bit trickier to explain, but still works based on the idea that internal jumpspace points cannot map to each thru an appropriately sized real world object. One essentially gets the first available displacement point.
For the jumpspace calculation then, one is looking for a series of points that solve the desired travel plan, if one posits that the interface of jump and normalspace is informative to some degree about both sides, the difficulty is that the exit point must be chosen based on indirect observation of the entry point.
My suggestion is that the initial jump calculations at each level use the fact that the stars 100d limit as a beacon - a point connected to a displaced point is easier to find - possibly because a large number of points all would tend to be displaced and give similar solutions. Using this, one can say, aim for the point were everything has a (techbabble) rating of .06, since that is clear rating of a displaced point -too, possibly, one could also identify the star's 100d displacement zone based on known characteristics.
The thing is, none of those advantages exist in an empty hex. So it is far, far harder.
Heres the deal. Jump space is effected by diameter, and not mass. While very small (cosmologically speaking) bodies can mess up a jump, and force precipitation, they may not be big enough to provide a useful target. Using inddirect calculations to find the exit point, there ca be posited a critical point where the number of displaced solutions cannot be distinguished from an incorrect or unwanted solution. In short, one has a statistical problem, one that ones life is bet upon.
However, we still have the issue of why those hexes act as empty even if they have all kinds of junk and possibly stars in them ? Traveller has too few stars, and potentially too many waypoints....so here's my gloss on the situation.
The absolute size of a star determines its useability as a target for early jump calculations.
Since at least 50% of all stars are missing from the map, we can set the limit there -bodies smaller than a particular size of M type stars (which are about 50% of the stellar population) are too small to be useful (and luckily, are also some of the least likely to be useful for anything). Gaps exist for low TL jump because, simply, they don't have a big enough beacon.
Now we throw in the differential tech limits on empty jumps, and Volla ! A number of seeming contradictions clear up. The early vilani were unable to make empty jumps early in their history, because they needed beacons, but developed beconless jumps later at tech 10. Beacons allow jump tapes as a wide variety of solutions work for star to star jumps; Terrans may or may not have had it from the start, but they did not have to have a massive waypoint to enable jump.
Once jump 2 had been developed, it was the initial star to star kind, precluding serious jumps to empty hexes; the ability to jump to empty hexes would be developed at the next tech level - where jump 3 tech would make it unecessary. So, with this scheme, there only is an issue for a brief window -the two tech period of jump 1; at any time empty jumps are more hazardous, but only then was the inability to calculate them not mooted by an advanced jump drive.
Waypoints matter with regard to fuel and supplies; and some kind of calibration may perhaps reduce the hazard...but do not make the jump possible if TL will not allow it. In fact, space can be full of undiscovered or discovered substellar objects...it doesn't matter if one can only jump from star to star, and it doesn't matter (unless you hit one, I suppose) once you can do empty jumps.
Yes, I'm aware I havent adressed jumping from inside a 100d limit. That's for later. The above is relevant to exit points.
Wow. That was a lot of babble. Time for bed. Dig in !
No, NOT to my bed. Bad Traveller !
*Yes, everything's moving, but one can measure and record any needed movement of a given system and or planet, and store it for each system as a set of vectors relative to a common arbitrary origin ...say, Reference (in whichever sector that is in). Compare the data for the target and the jump exit system, fire up the drive, and set the resultant vector -which will be preserved at exit. Sure, you'll want to be sure that you miss any planets, but that's what a navigator is for. And in unexplored systems, much of the movement will be known beforehand -and as to those pesky smaller but still massive bodies ? Well, that's what the scout survival roll is all about.
