It seems that I can do better then her.
I downloaded comet data from Minor Planet Center and for each one, calculated the Sun-comet angle on 2010/2/17 and 2011/3/11. I mean, if it is an earthquake predictor, then the angle should be the same on both dates. However, this is magic, so we can be a bit more flexible, and require the angles to be within +/- 6 degrees of each other. (The woman gets 5.71 degree error with Elenin). Okay, I run the code...
Out of 209 comets in database, 33 (15.79%) pass the test. P/1997 C1 (Gehrels) has the best fit, with only 0.43 degrees difference. 237P/LINEAR is a close second, forming the same angle with the Sun on both dates with 0.46 degree error. The bronze medal goes to 29P/Schwassmann-Wachmann, which has it right down to 0.58 degrees. Another recently famous contestant, C/2006 Q1 (McNaught) erred 1.06 degrees, landing sixth place, just 0.12 degrees worse than CL94 (Lemmon)... Everyone's favorite, C/2010 X1 (Elenin) is a huge disappointment at #33 with 5.71 degree error.
The list is below. Format: error, magic angle, comet name.
Code:
0.43 13.0 P/1997 C1 (Gehrels)
0.46 25.1 237P/LINEAR
0.58 173.0 29P/Schwassmann-Wachmann
0.77 83.5 P/2003 HT15 (LINEAR)
0.94 104.7 P/2008 CL94 (Lemmon)
1.06 64.6 C/2006 Q1 (McNaught)
1.30 24.9 C/2007 B2 (Skiff)
1.50 89.1 C/2010 D3 (WISE)
1.53 14.8 C/2005 EL173 (LONEOS)
1.66 12.7 187P/LINEAR
1.90 163.0 37P/Forbes
1.98 7.3 111P/Helin-Roman-Crockett
2.00 34.9 C/2009 K2 (Catalina)
2.60 160.3 C/2009 U5 (Grauer)
2.70 170.8 P/2007 H1 (McNaught)
2.85 121.2 128P/Shoemaker-Holt
3.15 127.1 C/2007 S2 (Lemmon)
3.18 123.8 195P/Hill
3.20 52.4 P/2004 FY140 (LINEAR)
3.75 137.6 P/2005 S2 (Skiff)
4.11 93.5 C/2008 A1 (McNaught)
4.45 66.3 119P/Parker-Hartley
4.68 19.3 97P/Metcalf-Brewington
4.77 114.0 P/2008 Y3 (McNaught)
4.77 11.4 P/1996 A1 (Jedicke)
4.78 165.8 179P/Jedicke
4.83 111.1 P/2007 R1 (Larson)
4.83 37.9 C/2008 T2 (Cardinal)
4.98 111.7 C/2009 F2 (McNaught)
5.43 124.2 P/2010 H5 (Scotti)
5.57 7.7 P/1998 U3 (Jager)
5.64 66.5 P/1997 T3 (Lagerkvist-Carsenty)
5.71 169.4 C/2010 X1 (Elenin)
The code used for calculations is here: http://pastebin.com/mJmrZN70