It should be substantially easier to create a dijet filter for the 2009 simulation than an inclusive jet filter. I extended the filter I wrote to look for inclusive jets to look for dijets. As expected the situation is substantially improved.

The filter, with the current settings, accepted 12 events in 20,000 thrown in the partonic pT bin 2-3. Only 1 of those had a dijet reconstructed at the BFC. No events not accepted by the filter had dijets reconstructed. There is certainly room to improve the rejection of the filter.