I have DFCO turned off at the moment for tuning.

I just went to COS5 2bar OLSD and I'm having the same issue. It's going rich on decel so my low MAP (20-35kPa) columns in the VE table are getting the shaft when I apply the correction factor from my wideband.

 

I ended up making B9012, B4003, and B4004 all zeros to try to achieve a lower minimum pulsewidth. It helped a little but it's still rich on decel ~1200rpm. The next thing would be to knock down the min pulse adjust which adds pulse width if under 2.5ms is commanded.