I included the Endcap towers in only the away side veto sum in Jan's W-algo ver4.2.2 (near side will come later).  Below are some plots before and after the changes.

Figure 1:  Away ET sum before and after adding ETOW.  The cut is left at awayTotEt < 8 GeV.

Figure 2: Final W selection plot before (black histogram) and after (red histogram) adding ETOW.  The 2nd panel is the difference between the two histograms.

Table 1: Counts above and below final cluster ET threshold with and with out ETOW veto.

Just for fun, I fit the distribution above with an exponential to approximate the background rejected by adding ETOW to the away side veto.  Assuming this same distribution would be true for a "second endcap" at -2<eta<-1, this fit apporoximates the background that should be subtracted by including it in the away side veto as well.  Below is the Final selection plot after subtracting the "second endcap" background.

Figure 3: Final W selection plot after subracting exponential background from '2nd Endcap Veto'    (ver4.2.2)

Other ideas not implemented yet:

Include ETOW in near side ET sum.  This could improve efficience of 2x2/nearJet cut at the edge of the barrel (eta~1).

Look for Z's with one arm in endcap.