I have been doing some SMR2 testing as a capped max SMR2 skilled human rogue, and the more I test, the more and more disturbing it's becoming to me.

I already discussed this to a great extent in Discord, but figured I should post my findings here. The testing was done with a human rogue in defensive stance with no armor, against a 100 Wizard with 202 wizard spells ranks, casting 917. I also tested it with a cleric casting 309, and the SMR2 defense numbers were identical between them.

Essentially, what I have discovered, is that a fully trained semi (in SMR2 skills), will have roughly 7 less SMR2 defense than a fully trained square (The square having 200 more SMR2 skill ranks.). Since your SMR2 rolls fall somewhere in the 1-95% chance to avoid range, with those skills, then that essentially means we can consider that to be a 7% difference in chance to be hit (Unless you are at 95%+ chance to avoid, then the difference is even less. Much much less.).

To do my testing, I started by dropping my dodge from 303 ranks to 203 ranks, which dropped my SMR2 defense by a mere -4, or 4%. Taking it further, and dropping my dodge to 2x and my perception to 2x (same training as a semi), it resulted in a fluctuation between 7 and 8% (mostly 7). Adding some +40 Agi +5 Dex enhancives, Blurs, and 215 to the mix, saw only a 6-7% difference between 2x and 3x in those 2 skills. Also, those enhancives only increased my SMR2 defense by 2.

From what I found, there seems to be a very heavy soft cap, which seems to include most bonuses. Once you reach a certain amount of SMR2 defense, no matter where the bonuses are coming from - any +skills/stats/levels/spells will have a diminished bonus.

In my opinion, these diminishing returns are so severe, that it makes the advantage of being a square and being able to train in more SMR2 skills extremely insignificant.

Also, the most disturbing thing to me, was what I found from testing armor Action Penalty. When I put on MBP with no overtraining or armored evasion (So -20 AP), it did not even make enough difference to reduce my SMR2 defense by even 1. So essentially, as far as SMR2 is concerned, armor overtraining is pointless, and armored evasion is pointless. As long as you are trained to wear your armor, you are fine.

So what really is the point of armor overtraining? And Armored Evasion? For a few miniscule extra DS from evade? In my opinion, this makes both overtraining and Armored Evasion a complete joke. And here, it was my understanding that maneuver skills were the MAIN reason to want to reduce your Action Penalty - but with SMR2 it doesn't even increase my avoidance chance by 1%, going from -20 AP to -5 (202 Armor ranks + 18 Armor Use enhancives + Rank 5 armored evasion in MBP.).

Konacon stated in Discord that it is the same in every current SMR2 equation. His exact words: "At the moment, which SMR shouldn't matter," when I stated that I was using 917 in my testing.

So whether or not the SMR2 is physical or magical, shouldn't matter. At least currently with any existing maneuvers.

So here are my thoughts on what I think should be done with SMR2.

I feel that action penalty should have a much more severe penalty to it. The more severe the better, if you ask me. Now, it's possible the penalty IS more severe than it seems, but that the way the current equation works, the result is being almost entirely diminished in the case of a fully trained square. Either way, it's the real world results that matter the most.

But so that it's not a flat out nerf (Especially to squares, who generally have the worst action penalty!), the diminishing returns should be less harsh. I'd like to see the advantage of those 200 ranks over a semi be a good 2 or 3 times as much as it is!

I think it's only fair that a square should have a significant advantage when it comes to SMR2 defense, as far as my understanding of how class weaknesses are supposed to be.

Well, if changed as I suggested, it should have little or even no impact on how easily a pure defends. I mean, if you wear double or reinforced leather there could be some drop, but hopefully not a lot - and it could be partially mitigated by armor overtraining, or completely mitigated by armored evasion (Oh if only people would actually WANT Armored Evasion.).

This could perhaps have a larger impact on lower level characters, since they would not be affected as much (or at all) by any changes to the diminishing returns.

But whatever the case, I think it's at the very LEAST, fair to expect Action Penalty to make more difference than it currently does, at least for a max trained square. Just looking at the wiki under Armor Action Penalty, the first thing it says is "The action penalty is a modifier subtracted from your chance to perform a standard maneuver roll." The fact that my -AP had no noticeable impact on my SMR2, in real world application, shows that it's main purpose is completely invalidated for me in the case of SMR2 (And who knows how it works with SMR, but that will eventually be phased out, so this will become even more of a problem.).

Sure, some characters might see a nerf from this. But if it has to be done to make Action Penalty matter, then I feel it's something that needs to happen.