6/12:
Using the minimal Stetson data currently at my disposal, I combined Stetson and Clem for M13 to make a preliminary fit of V-I vs. g'-V. This is preliminary because not only do we not have all the data yet, but I also ignored the offset correction to get a reasonable sample size and only used Stetson data from a particular part of the sky, because there appears to be two seperate offsets to correct for. Got the full Stetson dataset, looked at CMDs of annuli centered on the cluster to find a good region to look at, decided on a ring from .02 to .12 degrees from the center of the cluster. Used a sharp cut of .15, and chi cuts of .8 on the low end and 1.5 (or 3 if the I magnitude was brighter than 12) on the high end. We're not sure why the chi's are so bad.


6/13:
Using the ring from yesterday, along with chi and sharp cuts, matched the Stetson and Kyle data, then matched the Stetson/Kyle to the Clem/Kyle to make Kyle/Clem/Stetson. plotted up the resulting points, and after manually clearing out a lot of mess was left with this. Discovered that a lot of the bad points were from stars with low separation values, made a good sep cut (sep > 6) and the plot looked much better. Spent way too long fighting with SM trying to figure out how it handles strings in vectors, gave up after ~2 hours, used what of the program I actually had working to make another fit with some extraneous points removed from the actual fit.


6/14:
Went back to square one, edited my matchers to find the best match within 1 arcsecond instead of the first match, which is what I had been doing. Serendipitously found that somehow my Kyle/Clem matcher had an error in it, which meant that several different stars were being given the same ugriz magnitudes. Don't know what was causing that, but the sorting matcher fixed it on its own. Redid M13 matches to get a new Kyle/Clem/Stetson file, made a new V-I vs. g'-V plot. Much cleaner. Next, made M13's g'-i' vs. V-g' and g'-r' vs. V-g' plots.


6/15: Redid the V-I vs. g'-V for several different error cuts, with no success at getting a good enough fit. Moved to more difficult fits, starting with fitting to g' = a + b*V + c*(V-I). Plotted V and V-I vs. the residuals (points in blue have the smallest errors, followed by red, then green). Next fit to g' = a + b*V + c*(V-I) + d*(V-I)^2, got a somewhat better residual plot (better mostly because it mostly contains only the blue points from the previous plot). Turned my attention to M92. Adapted the new matching sort to work for M92 and ran it. plotted V-I vs. g'-V. It can be clearly seen from this plot that we're not dealing with a linear fit here. Overlaid that plot with the same plot for M13 (shown in red).


6/16:
Made the prettiest plot EVER.

Back to main