Suffice to say that if no other new frigate was designed between the time Soyuz was laid down and the time Miranda was laid down, then by the late 2270s Starfleet must have been really hurting for frigates. Likewise for cruisers unless the Rangers filled that role.
It's not so obvious that this makes a difference if we use 2300-era ship stats (by which standard there's not a lot of room for improvement between a Soyuz and a Miranda). But if we use statlines normalized to, say, 2250 (by which standards a stock Connie is a big impressive ship that was to them what the Excelsiors were to us in 2300)... there's likely to be a bigger gap.
This is something I explored when trying to rough out statlines for the "2235 game" idea I've been playing with. Basically, I started by roughly doubling everything's stats, except for Defense which is... complicated. I rounded up or down to fit some mathematical formulae I set up in an attempt for balance. The idea being that people were easier to impress back then; a "minimal science suite" corresponding to S1 in those days would be a veritable S0 now, whereas something like a Centaur-A would be an impressive ship with S5 to S7 by their standards- they couldn't duplicate its capabilities without building a dedicated science vessel.
So for them, the Soyuz-class is a modern escort with a statline of...
[checks]
C3 S2 H2 L2 P1 D3.
By contrast, a Ranger has C4 S4 H3 L3 P4 D4
Of course, Starfleet of that era has its own obsolete 'dog' ships, including classes that are refits of what the member worlds were using at the founding of the Federation. But things are starting to look up.