16 trained men performed 2 testing protocols using 4 repetition maximum loads: TS (3 sets of bench pull followed by 3 sets of bench press performed in approximately 10 minutes) or PS (3 sets of bench pull and 3 sets of bench press performed in an alternating manner in approximately 10 minutes). Bench pull and bench press VL decreased significantly from set 1 to set 2 and from set 2 to set 3 under both the PS and TS conditions (p < 0.05). Bench pull and bench press VL per set were significantly less under TS as compared to PS over all sets, with the exception of the first set (bench pull set 1) (p < 0.05). Session totals for bench pull and bench press VL were significantly less under TS as compared to PS (p < 0.05). Paired set was determined to be more efficient (VL/time) as compared to TS.Unfortunately, Robbinson et al. missed to record the effects on strength or muscle gain. It may be speculated, however, that the subjects on the antagonistic "superset" (for a real superset rest periods were still too long) protocol may well have seen greater progress due to the increase in overall workload and the amount of weight used in the individual sets.
On a side-note: If you are a firm believer in complete muscular exhaustion, but still wan to benefit from the time effective antagonist approach, I suggest you copy my approach of antagonizing exercises and not sets; i.e. you do bench presses to exhaustion (even add a drop set at the end), then do the bench pull and afterwards you switch to another chest exercise, which, again is followed by a movement for your back.