Best fix systems - Are not valid for very long.
As distances increase to the fix systems (as our livingspace expands) the accuracy goes down.
For a given set - Sure it might be worth to find them.
But we don't have a fixed set (in the future).
That ought to be plenty of accuracy. Only 1/32th LY accuracy is needed (and even 0.0112492 is well below that)
Try running it again, but this time round your candidates to the 1/32th grid.
And then run the Match error again.
Pretty good chance you'll nail all of them exactly I think.
Yes if players go deeper into space somewhen 4 better fixes should be evaluated. However that requires that you have precise coordinates form systems out there. You can calulate them with those 4 but you don't have it garantet acurrate. So as farther out you go and as more second generation fix points you take the more uncertain everything will become. The grid lock to 1/32 and compareing the distances to a lot of systems for a new fix out ther will help.
~
Byway I haven't checked those 4 fix systems position within ED if they really still at the exact positon like we have from the file. My invalide filtering was only done with one fix reference. That's not 100% sufficient and a check in ED (alldistances of the 4 fix systems to each other) would have to take place first before useing.
Also this check has to be repeated for every new ED release.
~
With:
CalulatedX = Math.Round(32.0 * CalulatedX) / 32;
CalulatedY = Math.Round(32.0 * CalulatedY) / 32;
CalulatedZ = Math.Round(32.0 * CalulatedZ) / 32;
~
I get this:
Average Distance Error 0.0
MinDistanceError 0.0
MaxDistanceError 0.0
~
Average Distance Match Error: 0.0024294
MinDistanceMatchError : 0.00000049071
MaxDistanceMatchError : 0.00499461
~
This means I got indead the exact position for ever of the 230 Systems if I grid lock them to 1/32 useing the 4 fix references.
The Match Error might look confusing at first but is correct.
The Match Error is to expect to be +/-0.005 since the entered Distances are rounded to 2 digi after the decimal point
(I used the distance Error (= absolut error value) for the Match which emans -0.005 abecomes +0.005 and the average will be the half of that
~
I quickly did an additional simulation by adding +1 for the entered Distances to the RefSystem1 and got this:
Average Distance Error 1.08
MinDistanceError 0.211
MaxDistanceError 2.58
~
Average Distance Match Error: 0.4775
MinDistanceMatchError : 0.00008375
MaxDistanceMatchError : 2.44034
~
That says it is possible to detect faulty distance inputs with the match error. However it's not possible to say which of the 4 distance is the wrong one.
Forexample I saw when I debugging that this min error is from system 117 and infact to RefSystem1 ! the one we manipulated. However the distances to the other refsystem then were way of. So you just have to check every of the 4 distances and if one is off the whole input is invalid.
~
Edit to add: might have to clarify one point but you can read it from the code too: What I call distance error is in truth the position error. I called it distance error because I calculate the distance of expected position (from file) to calculated position.
And the Distance Match Error is the difference between the entered distance (the point 2 digi) and the distance we can calculate with the calculated position to the ref system.
