Here are stats for ~35,000 SDSS fields in the r band.

These are from run 117, and are merged from several stages:

Stage	Objects	Code tolerance	Pixel error	Nverify	Scale		Solved This Stage	Solved Total	Remaining	Solved Total %
R1	0-50	0.005	1	25	0.39-0.41		38488	38488	457	98.8
R2	0-50	0.01	1	25	0.39-0.41		128	38616	329	99.2
R3	50-80	0.01	1	25	0.39-0.41		157	38773	172	99.6
R4	80-100	0.01	1	25	0.39-0.41		18	38791	154	99.6
R5	0-100	0.01	1	20	0.39-0.41		1	38792	153	99.6
R6	0-100	0.01	2	20	0.39-0.41		0	38792	153	99.6
R7	0-100	0.01	1	25	0.386-0.416		0	38792	153	99.6
R8	0-100	0.015	1	25	0.386-0.406		19	38811	134	99.7
R9	100-120	0.015	1	25	0.386-0.406		8	38819	126	99.7

This shows quite severe "diminishing returns" and shows that the settings we used in the first round are quite reasonable.

---

These plots were generated by source:trunk/scripts/runs/117/R1-8/plots.m

---

http://trac.astrometry.net/attachment/wiki/SloanStats/fig1.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig2.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig3.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig4.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig5.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig8.png?format=raw http://trac.astrometry.net/attachment/wiki/SloanStats/fig9.png?format=raw