attending: Michele, Alberto, Neil, John, Ira, Sean, Ed, Stas, Antoine

evaluation completed for 3.2, 3.3, 3.4

MBH more multimodal than hoped/expected. Is that because of approximated waveforms? Lang/Hughes picture of shrinking error boxes is less rosy with multiple error boxes; spin modulations may mask orbital modulations

frequency annealing may suffer from more maxima, and may risk locking into wrong one; compare to real LISA localization

try to get nice maps for proceedings

discrepancy between 3.2 SNR evaluations explained by different Galaxy models

could try using the real Galaxy background put into 3.2; but even then the noise is not stationary because of Galaxy vs. LISA orientation

should keep official S_n to do SNRs; little sensitivity to integration limits

for EMRIs, better understanding of where degeneracies are helped

also SNR is better distributed along the year, helping localization, whereas final SMBH burst is quite degenerate

but other EMRI parameters have larger degeneracies

cosmic-string analysis was very effective for all groups; have analytical understanding of degeneracies

problem may be more challenging in the presence of MBH signals, and glitches

for 3.5, have two reports; together they give a consistent story

investigating the X, Y, Z level inconsistency—compare svn836 and a newer revision for http://www.tapir.caltech.edu/~mldc/results3/challenge3.5-key.xml

for 3.1, Neil’s MLDC2 codes are in the SVN, figure out who fits whom and does correlation; need to pass frequency derivative

LISA Simulator for Stas 3.2, MT 3.2, MT 3.3

Neil’s proposal seems good

MBHs need better taper at the end to avoid hiccup; increase mass ratio to 1-10

EMRIs could use taper, but is hidden in noise; have undisclosed number of EMRIs of each kind; consider lowering plunge eccentricity (0.05)

stay with eccentric LISA orbits although stochastic background may be strange (use new T channel)

could possibly do 3.5 with LISA Simulator (make barycentric files and delete as you go)

put strawman on wiki, then start generating soon