Hard Cap Removal on Resists
TL/DR: Resistance gems are being buffed for people who have high base resistance, nerfed for people who have crappy skills and no other resist gear. We are quartering their resistance rate, which sounds bad, but they were almost useless before with no benefit of going beyond r1 in most cases. A person with plenty of resistance gear and skills was barely ever getting any bonus from using resistance gems due to a hard cap that existed.
It was recently observed that a hard cap on resist rates of a base 50% (prior to CNC/CNT taking effect  so actual % could still have been much higher if your CNT was higher than foe's CNC) was causing gems, items, and skills to have zero effectiveness after a relatively low threshold. This was contrary to what we had stated in the past that all increases to resists would have some returns, albeit diminishing.
For example, if all your skills and gear caused you to have a base resistance rate of say 35%, adding a resistance aura gem in play would only ever bring you to 50%  a meager gain considering with pally, repel, and shared bliss bonuses, an r5 gem alone was supposed to give you an additional 86.8% chance to resist.
This was a design flaw, since in Nodiatis we prefer to have every bonus stack, other than like auras. Hard caps are not fun for people who really want to test the limits of a particular build by min/maxing.
We cannot simply remove the hard cap since that would make resistance rates soar. To address the problem in a balanced manner we are doing the following:
Removing the 50% hard cap completely and using a new formula to scale final % based on CNT/CNC with smooth, uncapped bonuses
Quartering the base % of all resistance gems so that they can no longer get to absurd levels with multiplicative bonuses
Note, when understanding % chance effects, it is important to distinguish between effects that enhance an existing effect and an effect that adds a separate chance of the occurrence. The pally bonus, repel skill, and shared bliss skill are the former, and they enhance the % effect of the gem multiplicatively. A gem with a 10% chance of resist will turn to 14% (10 * 1.4) with a max repel skill (40%) for instance. Add shared bliss (25%) to that and you have 14 * 1.25 = 17.5%. Add pally (24%) and you have 17.5 * 1.24 = 21.7%. This method of stacking could theoretically go over 100%, but shouldn't with the new changes to base gem resists.
Other sources such as impedance, grim resolution, diabolism, resist runes, armor resists, etc will each add a separate chance to occur. This is more difficult to evaluate (see signature) but their combined values can never go over 100%. Lets say you are starting with the above 21.7% chance to resist from having a gem in play from the above example, and you want to calculate what your resist would be if you added 10% impedance, 15% grim resolution, an 11% rune, 14% helm, and 5% diabolism. It would be 1  the chance of all them failing, which is 1  (.783 * .9 * .85 * .89 * .86 * .95) = roughly 56.4% Whereas before it would have been capped to 50%, you'll make use of all your effects now. Your ultimate resist rate will still be much lower if your CNT is lower than your foe's CNC, or higher in the opposite case.
In summary, people who are stacking as much resist as they can and using the highest rank gems will have a tiny bit higher resist rate. People not using the gems won't notice a difference. People who were only using resist gems for resist (which should be nobody, the skills are free to train) will have lower resists.
There was another small problem found with Block, Parry, and Kite, where having absurdly high stats compared to your foe's stat which prevented such things did not continue to increase your success %. That hard cap (though it would be rare for anyone to have hit it) has been removed and replaced with a logarithmic curve. If you happen to be in this category, you will only see a 5 to 10 percent multiplicative gain, i.e. a parry chance of 20% vs a really crappy foe might increase to 21% or 22% in an extreme case.
__________________
Computing the probability that at least one of the following events will occur:
P(a or b ... or z) = 1  P(!a and !b ... and !z)
Probability
