I and others have reported what seems to be an increase in the number of complete misses on attacks, with even deniers with multiple modes of denial missing far too often. @Carlos and others said nothing has changed in the code and it will take more than anecdotal reports to want to look into it more. So I decided to do the experiments.
Hypothesis: player attacks are missing far too often, in particular with multimode deny attacks like freeze+possess by Kihaku.
Approach: collect data on adventure map node 377 where we have 3 hardened defenders: 20% chance of a status attack failing to land when they are hit.
Expt 1 - run single mode AOE deny attacks and see how often we get 0/1/2/3 misses; used Rabbish and Total Blind.
Expected % hit vs miss: 80% hit, 20% miss on any one attack,
Actual hit % was 76% after over 100 attacks. No difference.
Expected distribution of landing TB on 3/2/1/0 hardened monsters:
51% | 38% | 10% | 1%
Actual distribution after 80 rounds:
50% | 43% | 6% | 0%. No statistical difference.
Conclusion: single effect spells are landing as expected.
Expt 2 - ok now let's see if things change at all when using attacks with two independent status effects. Used Baba Yaga's Kid Eater with negate healing and dmg reduction, and Kihaku's Kuzushi with freeze and possess.
I measured two different results: distribution of how many total effects landed vs 3 enemies (could be 6/5/4/3/2/1/0), and the count per defender of how often both effects landed, one landed, or none landed.
Results 1 - (flawed!! I'll explain in a minute!)
First test I got results matching my hypothesis of missing too often. The % of attacks which landed 2/1/0 effects were expected to be 64%/32%/4%. Actual results after 100 runs were 52%, 26%, 22%, which correctly had dual hits twice as often as single effect landings, but way too many complete whiffs. At this point I realized, sheepishly, I had not looked to see on a whiff whether the attack itself actually landed or not! (Doh!!!) Since the AOE's used have only 90% accuracy, the chance of making the attack check for all three is not 100% but 72%. And you'll miss the attack 28% of the time. When the attack misses, no status effect check is made: it misses too - or with two potential effects, both miss.
Repeat of experiment, counting results only when the attack landed!
Data collected: two/one/zero effects landing per monster, expecting 64%/32%/4% if everything is in order. Actual results after 160 attacks...
65% both effects landed (104/160)
30% one effect landed (48/160)
5% no effects landed (8/160)
Follow-up: I then ran a small number of tests on node 411 on the map with three normal monsters, none hardened. Results were as expected, no oddities there.
--> Conclusion: the frequency that deny or negative effects land is as expected according to the accuracy of the spell and hardening effects. The number of misses may be frustrating, but it's not bugged.
Discussion: I'm kind of glad I messed up the experiment the first time. I was convinced from the frequency of missing both denies, or missing all denies on multiple monsters, was considerably too high. However, once you account for the fact that AOE's miss, everything fell into place. Almost 30% of the time an AOE is going to miss at least one attack completely which means no effects will land. On top of that, if a foe is hardended, once the attack hit is made, you have only a 2 in 3 chance of landing both status effects, and 4% chance both will miss. Make that a tough defender and you should expect seeing both/one/no effects land change to 42%, 46%, 12%. Bulwark? 25%/50%/25%.
Finally, if you're curious, what is the chance of landing those "two stun attacks" or "two nanovirus" attacks? IF the attack itself hits, you should see:
- 100% effect lands with no immunities
- 96% chance of landing vs hardened
- 88% chance of landing vs tough-35%
- 75% chance of landing it vs Bulwark.
But keep in mind you'll miss 10% of your attacks with a normal AOE.
I hope this analysis is helpful, let me know if you have questions or comments.