So I think part of the thing here is that there are a lot of checks
for strength, and having +1 strength affects a lot of things in
positive ways. They aren't, as a whole, overly large or noticeable
things though, or storm giant strength build would look and feel
noticeably worse than fire/cloud/frost.

So back to this particular case, we'd have to look at Lug/Mamlauk/Orc
specifically and see if there are checks in those specific to orc
skills that *really* care about +1str. Which, given that the adapts
were written before edges, would be a weird thing to have included.
It would also (probably) mean you'd really care about -1str as well,
so losing 1str would be roughly as bad as +1str would be good, and
that doesn't seem overly noticeable either.

If it was up to me to assess how good/bad such and edge is, and
apply a cost that was in line with that value, I'd try and stick an
objective measure on it in some way. My first thought is some sort
of HP equivalent value. I probably won't be able to explain this as
well as it needs, but I'd look at how much HP equivalent value does
+1str yield in a "test" fight. So we'd set some base values of like
100hp per hit, 10 rounds of fighting, attempting to maximize the
effect, and then run the math to see how much that extra 1str yields
in extra damage and damage mitigation over the 10 round fight when
we're trying to maximize the difference made. So in this case you'd
parry at a slightly higher effectiveness, and maybe a little bit of
dodge based on % carry weight, and then chose skills that care a lot
about strength, so maybe something like whomp/brute force/hack.

I suspect, the overal value here would be fairly low (<100hp diff),
but obviously someone with code access would have to take the time
to actually establish this measure. You could then compare it to,
for instance, the throw edge, which if throw is usually 80% chance,
but boosted to 100% with edge, you'd expect to hit 1 more throw per
10 round fight, which would have a clear HP equivalent value. If
Lug is less than the throw edge, then the cost should be relatively
cheaper.

No measure would be perfect in a game as complex as CF, but you
could probably establish a few of them along with some code to run
the testing and have a pretty solid basis for balancing things,
cost wise, for edges and other things.