This the same set up and thus DPS is very good indicator and in addition it simpler to use.
If you want TTK you need to put accuracy into play and compare probabilities. For instance how much is that TTK that unit A will kill unit with in this time frame with 90% probability.
For economy in time and space. I am not going to do a full statistical analysis when I am not really interested how much better the new SU-85 bad the simple fact that it is better.
Let me use another method thou that will bring us to the same result:
Lets arbitrary assume that half the missed shots score collision hit.
The old SU-85 firing on OKW PzIV has a 50% chance to score a hit we increases to 75% since half of that shot collide.
It also has a 77% to penetrate and that bring the total chance of doing damage to 0.58. So the probability of killing a PzIV with only four shots is around 11%.
The new SU-85 firing on OKW PzIV has a 80% chance to score a hit we increases to 90% since half of that shot collide.
It also has a 94% to penetrate and that bring the total chance of doing damage to 85%. So the probability of killing a PzIV with only four shots is around 51%.
One can factor in ROF advantage but it will not reverse odds like these and it is simply too much work since one would have to calculate for a similar probability.
The fact is that SU-85 might had great ROF but had terrible accuracy and since it is case mate and had less rotation I doubt that it could even take advantage of that ROF unless on certain map.
So the result is the same the new SU-85 is simply superior vs mediums unless one have the RNG GOD on your side. The question of how much better is hard to establish and not worth the effort.
If your "economy of time" leads to potentially wrong results, you should not publish. Especially not without any note why they are flawed and how much it could effect the outcome. If you are "not interested how much better the new SU85" is, then don't state any numbers that imply it would be exactly X% better. Easy as that.
Second: You completely switched your math now to support your view. Previously you calculated DPS as a function of "how long do I need to kill X". Now you calculate the chance to kill X with the least shots possible.
PLUS, you solely take the OKW P4 because you know that the lower pen will affect the old version the most. And last point (which is okayish but also should be noted): you calc everything for range 60 which also favors the new version.
But let's do it your way then. Old SU vs OKW P4 damage chance is as you said with my scatter preassumption about 58%. This brings us to 6,9 (or ~7) shots to kill, meaning 5,9(6)reload cycles or (with actual ROF) 25,8 seconds (26,25 rounding; unless I forgot to add 0,125 somewhere).
New SU, same thing: 85% chance to damage meaning 4,7 shots (5)-> 3,7(4) reloads -> 21,37 (23,1) seconds. In this case, the new SU is expected to be 21% better. Due to the penetration and accuracy profile, this advantage should diminish at lower ranges since we currently calculated everything for range 60.
Now the thing OST also has a P4.
Pen rates go up to 100%.
TTK old: 18,96 (without rounding the expected shots)
TTK new: 19,89. (without rounding the expected shots)
Basically identical.
range 30 (mid range for comparison)
TTK old: 15,63/20,27 sec vs OST/OKW P4
TTK new: 18,54/19,95
old SU is 19% better/identical to the new one.
So overall the new SU is better at long ranges vs OKW P4, while this advantage diminishes the closer the P4 gets. Against OST there is either no difference or the old SU performs better.
So conclusion of all this? New SU might have a slight edge at least against OKW since the unit gets used at rather longer ranges and now gets a bit more rotation (however it is also more costly).
Once vetted, the new SU has no chance though. It does not gain much from penetration increase whereas the old version gets much higher ROF. But even if we'd leave out the vet for no good reason, the "winner" is
1. not as clear as you try to convey here
2. depending on the set up and
3. the numbers you posted to support your hypothesis (19-47% better) are wrong for the most part.