Subject: efficiency table with wider pattern set Date: Tue, 04 Dec 2001 14:52:25 -0600 From: Alessandro Cerri To: svt@fnal.gov I measured once again the SVTsim efficiency using wider patterns than those used in our last week of running. I report both the old and the new numbers here as references... Comparing the two tables you'll see that what changes most is the probability that a pattern fires, which is increased by 4%. On the other hand, half of this 4% produces tracks that have weird parameters, and therefore the number of tracks with significantly different parameters from those of the offline mate is increased by 2%. The overall efficiency increases then only by 1%. Next step would probably be that of: a) implementing patterns without barrel crossers b) using the latest fit parameters with all the nonlinearity corrections... I'd like to do this in two steps... A "regular" pattern set: ^^^^^^^^^^^^^^^^^^^^^ total tracks 420 at least one hit didn't match 48 (48/420)= 11% (48/420)= 11% pattern didn't fire 52 (52/420)= 12% (52/(420-48))= 14% parameters are different (chi[svt-offl]<20) 49 (49/420)= 12% (49/(420-48-52))= 15% Total Tracks lost 149 35% Tracks with parameters matching but missing hit and/or pattern 20 5% tracks matched (hit, pattern and parameters) 271 65% "Wide" pattern set: ^^^^^^^^^^^^^^^^^^ total tracks 420 at least one hit didn't match 48 (48/420)= 11% (48/420)= 11% pattern didn't fire 38 (38/420)= 9% (38/(420-48))= 10% parameters are different (chi[svt-offl]<20) 56 (56/420)=13% (56/(420-48-38))= 17% Total Tracks lost 142 34% Tracks with parameters matching but missing hit and/or pattern 18 4% tracks matched (hit, pattern and parameters) 278 66%