Regarding breeding progression for stats in Project R, I do think that some of the mechanics behind it do need some work. The current system overall means that it's too easy to get absurdly high stats for single attributes, while also making the hard-capped stats like conception rate too easy to obtain. It also has the odd side-effect of making multi-attribute monsters much harder to manage, so it pushes players to just optimise single attributes. I mostly just have a strain of +hp melee goblins and a strain of +damage ranged goblins as there's little point breeding for much else; everything else I breed is just for completion's sake. Sure, getting some knockback resistance might help the melee goblins and getting the crit stats up on my ranged goblins might increase their DPS, but the added complexity of trying to get those stats means that I am often just as well off maximising their primary stat instead.
I don't really have a single solution to this, as there's plenty of potential ones that all provide their own advantages and disadvantages, particularly as there's several different potential issues that need addressing.
Making stats asymptotically approach a maximum value would lessen the exponential growth issue, but would effectively function as a hard cap once players get a bit of breeding underway as there's very little difference between, say a "pretty good" goblin at 950 hp and a "max stats" goblin at 1000 hp; this would effectively create a gameplay loop where players are encouraged to do a few rounds of breeding to get as close as is reasonable to the stat cap then to wait until they get a new type of monster. Making things like conception chance work like this too would have the exact same effect as there's almost no difference between a 98% conception rate and a 99% conception rate, despite the latter likely being much harder to obtain.
Another suggestion that hasn't been mentioned yet is to make the stats or circumstances somewhat antagonistic towards each other. For example, having increased hp also cause increased fetus growth time or having each birth reduce the conception rate of a slave by 5%. That way, sure you could create some kind of optimised supreme goblin, but it actually function like an inbred mess. An extreme form of this would be to do away with simple stat increases and instead embrace more thematically linked sets of attributes similar to MBM's Complex Breeding mod.
There's also the current mechanic of how stats basically are a 50/50 from the parents as to whether they are inherited, regardless of the actual values behind the traits. To me, it would make more sense if the breeding was less deterministic and required more deliberate efforts to obtain better stats, such as by making the child have their traits be randomised between their parents +-20% or so; so a +10 damage mother and a +30 damage father would create a potential range of +8 damage (+10 damage -20%) up to +36 damage (+30 damage +20%). This would avoid the weird inbred exponential growth, while also not punishing players to a large degree for trying to mix various stats as all stats would be inherited via the same formula.
There's also the point that providing simple flat values for traits is a pretty bad one in terms of flexibility. +damage basically only works on creatures that are cheap, attack quickly or do some kind of AoE, while a big monster with a heavy single-target attack would basically not care. Similarly, +hp unfairly helps cheap creatures that can be deployed en-masse as they allow the same +hp value to give a defence far more hp per deployment point. Moving across to percentage modifiers or at least giving different monsters some kind of multiplier on these stats would likely be necessary to make all stats at least somewhat relevant for all monsters.
