In all seriousness, I feel we both paid too much attention to the structural design. The difference between, say, 90 and 110 ain't that big in practice, while redesigning eats up valuable time.
IDK the exact relationship but structural design impacts other design elements too - engine/weapon/armor are described in the design log as "based on base design modified for struct. design". My first and only attempt at Mechanized Artillery is a good example; it had an initial BD of 94 (80th percentile of initial BD) but its 76 SD left it with 54 ED, 63 WD, and 77 AD. Its 180mm gun had the same attack value as my initial dragged artillery's 105mm gun (plus, yes, -30% move modifier on a 555 weight vehicle w/an 800 power engine). I went back into one of the final saves and re-designed it - first changing nothing, then modernizing. The iterated version is not terribly impressive (bad rolls for design improvements are always possible) and at this point a 4th iteration is as expensive as a new base model (480pts in either case, but iterating off the new one will be 120pts):

Doing a redesign may not only give you a better high-end model, but it may give you a better model than the intermediate one you'd get by using and refining the crappier one. The main question becomes whether you'd be better off using something else to fill that tactical role while you do a redesign, or if there's no better alternative. Forex, my initial light tank was so bad (SD 86, BD 85, ED 60, WD 71, AD 86... all of which ended up with crappy combat stats and a -60% move modifier) that just to usefully field test that design I'd have to iterate on the initial unit b/c it wasn't combat-ready as out-of-the-box.
(Ofc, again, RNG can make or break you in either case.)
You forget how we started by playing the first 5 turns on the same day. At least the ones(?) that I was in.
We did 8 and 7 in the first and second 24h periods of this game. Though again, only 2 players.