Thanks for posting those splits. Great information to chew on.
Your question though, "How much faster would they have been running steady 2:54s ?" raises one of the key issues that I don't think we've seen discussed much: is even pace always going to produce the fastest results? On paper I guess the answer is Yes, but we all know that a lot can happen in a marathon. People experience bad patches where they just feel rough for a few miles or more--stomach issues, muscle cramps, exhaustion. And often we can get through these and get back on pace.
Nike is planning for these guys to run like metronomes but what happens if at 15k one of the three isn't feeling it and drops back for a bit? Will some pacers stay with him and try to work him back up to the lead group like a domestique pulling a team leader back to the peleton in cycling? To do that they'd have to be running even faster than sub-2 pace, so it seems like anyone who has an off day or a bad patch is done for.
And do we really know that running exactly on pace for this distance is the best plan? When the course was announced and we talked about elevation some people argued--frequently with anecdotal information--that a slightly hilly course was probably preferable to a dead flat one because of the opportunity to offer a slight variation in the muscles you're using or how you're using them. Might a variation in pace over the course of two hours be preferable over 2:54s x 42.2?