I ran over a runlist provided by Jan:
/star/u/balewski/2009-march-bnl/4bernd/d76-89.runsOnly (attached below)

It comprises runs taken between day 76 and day 89 :4am, longer than 3 minutes, and containing L2W algo.

There are a total of 268 runs. All runs were marked as good by the Shift Leader.

The 5.1 pb-1 that Bernd quoted comes from the ZDC coincidence cross section calculated from the Vernier Scan calculated by Joe Seele of 2.554 mb (agrees with Bill's 2.4 mb).

If I look at the BBCMB events in the same runs with 32 mb cross section, I get 1.7 pb ^-1.

If I look at VPDMB events in the same runs with 32 mb cross section (stated during the Collaboration meeting by Bill), I get 1.3 pb ^-1.

BBCMB-Cat2 is only present in 147 of the 268 runs. This trigger did not care about BSMD live, so it should not be compared with the other triggers, which all did.

BBCMB-Cat3 and BBCMB-Cat0 are also both present in 147 runs (I did not check they were the same). Cat3 required all detectors except pp2pp and ftpc live. Cat0 didn't care about any. For the 147 runs, I calculate 0.9 pb-1 and 1.2 pb-1 respectively using 32 mb. I believe this tells us that the average live time over those runs was 75%. The ZDC coincidence luminosity for those runs is 2.9 pb-1. I think by comparing BBCMB-Cat0 to ZDC coincidence, we see that 32 mb is not correct (and it is more like 13 mb). If I go back and use 13 mb for BBCMB, I find we've integrated 4.26 pb ^-1.

There is also some question about whether the BBC trigger is saturating, so I made these plots for BBC and VPD: