At the time this putative rule was devised, Starfleet was not yet established as a century-old organization. To the contrary, many things about the Federation and its space navy appeared new - not so new that Kirk would have been field testing warp drives or phasers or transporters, but not so old that there would have been half a dozen generations of Kirks in the service yet. Quite possibly, then, all the cruiser types from 10 to 17 would have been introduced within a decade, perhaps a decade and a half before the TOS era and a decade and a half after the founding of the service, and the same manufacturer would have churned out these identical hulls so that another could equip them with slightly varying but contemporary innards.
It would only be later that Starfleet would gain in years, making it tempting to spread the 17 types evenly across a whole century (or, say, the first 70 years of it, with the last 30 reserved for the already evident type 18 and possible higher ones). But by that time the concept of the first two digits identifying the ship type would have been kicked out of the airlock, too.
"NC 17740" can't be coincidental.
Well, no. But "17740" certainly can. And Jeffries would have gotten the NC from generic aviation facts, not from any specific aircraft.