Can any train ops people on here explain if this actually makes sense?
On the one hand, sure, you're making it so the trains with the longest trips don't have to travel an extra 2.25 miles. But do the trains really care? Does it make a difference to them? Going off of that logic, shouldn't the D reverse at GC?
And on the other hand, you're putting the people who work / live on the Medford/Tufts GLX branch on the shortest line with (perhaps) the worst connectivity. Isn't there downsides to that? What if a large share of Somerville GL riders are bound for Kenmore, or for BU. This makes travel less convenient for them. But at least the E still hits Longwood, and this change helps Somerville - NEU travelers.
It seems like in the perfect world you'd do an analysis of O-D pairings across the GL, and figure out which of the four lines is the best match to the GLX on a trips basis and assign that one as its pair. Actually try to determine if Somerville riders are more likely to travel to B destinations or C destinations or D destinations or E destinations (and vice versa) and make the decision based on that. But that's a much more sophisticated analysis than just saying "well, let's not have these trains going too far."