In general your right, but with our case none of those 3 was the case. Due to massive number of transactions , log file was most of the times hammered by writes and also reads at the same time (log backup job) and that was causing the performance issue, not because the simple recovery is faster by it's nature.