> V = f - 0.095(f) + 0.6(J-Ks) + 0.92 standard deviation 0.14

f-0.095(f) is of course 0.905f...


Anyway, for UCAC3 "model fit" magnitude range (denoted f here) 10 to 14.5, and J-Ks 0 to 1 using 16,2000 V magnitudes covering the whole sky taken from Brian Skiff's loneos.phot critically assessed compilation the result is :-

V = 0.91f + 0.499(J-Ks) + 0.92

the J-Ks quotient shifted a fair bit with more data. That's five times the amount of data I used earlier.

When comparing the 16,200 actual V magnitudes against the V magnitudes calculated for the same stars from the above and taking the difference of the two (observerd minus calculated residuals, so to speak), the Standard Deviation on this is now 0.16, slightly higher, than for the smaller sample, but the Mean has improved now being at 0.000. The Median is also improved at 0.003.

How this looks for those 16,200 objects is shown in the Files section in a file called piccy.gif. Outliers extend beyond the graph both positive numbers and negative numbers, increasingly sparsely so. One or two are magnitudes out, but could well be mismatched linkages, or variables (2MASS J and K are contemporaneous within themselves, but Loneos V and UCAC3 f are random epoch).

The graph has measured V via loneos.phot minus calculated V from the above formula being applied to UCAC3 f magnitudes in the range 10 to 14.5 and J-Ks colours in the range 0 to 1 as y axis and UCAC3 f magnitude as x axis.

That'll be that for this. Folk'll likely take ucac3 mags and do what they want with them, regardless, anyway, irrespective of logic, it's what's happened afore with all other recentish astrometric catalogues.

Cheers

John