Why wouldn't it make sense to have the W leave Ditmars a minute or 2 after the N, but arrive at QP a mknute or 2 earlier...
If you compare the transit times for the N and W between Astoria and Queensboro Plaza, you will notice a 1 minute difference in favor of the W. If the W left Ditmars 2 minutes after the N and arrived at QBP 2 minutes before that same N, it would imply that the W were 4 minutes faster than the W. It isn't and is not likely to be, thanks to the slow express speeds now in vogue.
There are other factors, even assuming your 4 minute differential. Passengers entering at either Ditmars or Astoria Blv have a good shot at getting a seat. If the N came first, what would be his reward for taking the W? He would have to wait 2 minutes (standing) for the next W, which would be 4 minutes quicker than the N. His net gain would be 2 minutes but his standing time would be 2 minutes longer. Suppose the W came first. Some passengers might be destined to an N-only stop. They may wait for the next N or take the W and transfer later on. My analysis indictates that the passenger takes the first train and worries about the destination later. The additional reward for this behavior is to avoid any bad weather on the exposed El platforms.
Now, given that passenger behavior is to take the first available train, what scheduling strategy should be used to harness this behavior to the route planners' desires? The only mechanism is to play with the headways. The W-1 min-N-8 min-W... spacing is the most extreme one can have without merging conflicts at QBP. If the express were slightly faster, one might be able to make the headway imbalance to ½ : 8½ This would raise the W loading to 41%, instead of 38% of the Astoria line's traffic.