The trouble with the express is that only a fraction of the express station riders get to save that 4 minutes.
Let's assume for the sake of argument that both trains run on 8 min headways.
Say for the ake of argument that N's leave Ditmars at 8:00, 8:08, 8:16, etc.
When can the W's leave? They can't go at 8:04, 8:12, etc., because then they will exactly meet up with the N's just before Queensboro, and every time there will be a delay as one train waits for the other.
So instead the Ws go at 8:02, 8:10, 8:18 etc. so that they can get in to QP ahead of the N that left Ditmars 2 minutes earlier.
Now your trains out of Ditmars are at 8:00 (N), 8:02 (W), 8:08 (N), 8:10 (W), etc. If passengers arrive roughly at a uniform rate, only 1 in 4 gets to snag a W and save the 4 minutes. The others get to wait an average of 3 minutes and then get a local for their pains (assuming the average wait is half the headway).
Meanwhile, the passengers at the local stations all lose an average of 2 minutes because trains there come only every 8 minutes instead of every 4.
I realize I have oversimplified but this can explain how an express can actually have an overall net cost in time for the public. In actual fact, because not everything is uniform and punctual as I have assumed, the problem actually is worse.