Defenses, Elusivity and actual damage mitigation comparison

Ok, I know everyone is going to be like, "Defenses suck" or "Defenses were OP". Please hold your comments until you've read through the entire post. There is no tl;dr version since I think that most of the thread is relevant to the topic i'm discussing. If you're not interested in reading it or here to flame, please move along.

With I13 most defensive sets and defensive ancillary/patron shields had elusivity added to them. This was done to bring defense sets back into line with resistance sets in pvp. Many people believed elusivity was overpowered but few took the time to try to understand how it worked. Originally all elusivity was of the 0.3 value, currently it is 0.1.
The best way I could think of explaining how all this pvp hitting stuff works is through example. Let’s say that we’re trying to fight against a Fortunata with 45% defenses (pre-DR) and elusivity.
First we start with the attack accuracy multiplier. Pick the power you’d be interested in hitting with and see how much +accuracy% you get from slotting. For this example lets say you have 72% accuracy slotted into a power from enhancements. You take that as a decimal and add it to the base pvp accuracy (75%).

0.72 + 0.75 = 1.47

Next we take into account global accuracy from set bonuses. Let’s say for this guy we’ve got 40% accuracy from sets.

1.47 + 1.40 = 2.87

This is where the elusivity comes in. Elusivity is a .1 value or 10%, so the accuracy multiplier is reduced to 90% of its value.

2.87 x 0.9 = 2.58

Hold on to that number for a bit, we need to skip over to figure out the defense/to-hit numbers. As stated earlier the opponent had 45% defense prior to diminished returns. With diminished returns the defense is reduced to 31%. The base to-hit in pvp is 50%. First thing we do with defense is apply our character’s to-hit bonuses to the base to-hit. For this example let’s assume our character has slotted tactics, that adds 10% to-hit bonus(DR'd to 9%).

50 + 9 = 59

From there we subtract the defense values from the to-hit.

59 - 31 = 28

So we’ve accounted for defense and to-hit values all we have to do now is apply the accuracy multiplier we found above.

28 x 2.58 = 72.24

So our character running Tactics with 40% global accuracy and 72% accuracy from slotting has a 72.24% chance to hit a Fortunata with 31% (pve soft-cap) defenses and elusivity.

For the heck of it let’s say the character was SS and had rage. (rage slotted with 5-Gaussians)
50 + 19.5 = 69.5
69.5 – 31 = 38.5
38.5 x 2.58 = 99.33 = 95.00

With rage running the chance to hit is maxed. There’s always a 5% chance to miss or hit any attack so values will never be below 5.00 or above 95.00. No reason to wonder what the number would look like with aim since its to-hit is higher then rage, but build-up has the same +to-hit value as rage.

So from this we see that rage and pretty much any +to/+dmg click power completely negates defenses.

Let’s take a minute to compare the amount of damage taken by some average squishy and the Fortunata from the example above. Lets say an attack that does 200 dmg fired 10 times. For this example squishy has say 42% resistance (arbitrary number I pulled out of the air) and the Fortunata’s defense (74.83% chance to hit) protects them from 3 of the 10 attacks.

Squishy
(200)*(1-0.42) = 116
116*10 = 1160 damage taken total
Fortunata
(200)*(1-0.1) = 180
180*7 = 1260 damage taken total

That actually doesn’t look too bad. On average from the 10 attacks the Fortunata would only take 100 more damage then the squishy. This isn’t a truly valid example though. Most cases in pvp you’re facing someone with another to-hit buff (ss tanks, blasters, defenders, scrappers). So let’s do the same math with the numbers against a toon running rage.

Squishy
(200)*(1+0.38) = 276
276*10 = 2760 damage taken total
Fortunata
(200)*(1+0.7) = 340
340*10 = 3400 damage taken total

Since rage completely negates the Fortunata’s defense the damage difference here is more significant. The Fortunata takes 23% more damage from the attacks then the generic squishy. Also 200 damage is rather generous, a slotted KoB would be doing significantly more damage.

Allow me to rewind for a second and take it back to the good old days when elusivity was 30%. The chance to hit of the original build (40% global accuracy and 72% accuracy from slotting +tactics) would be 56.28%. Using that math let’s reexamine the damage taken.

Squishy
(200)*(1-0.42) = 116
116*100 = 11600 damage taken total
Fortunata
(200)*(1-0.1) = 180
180*56 = 10080 damage taken total

Here the Fortunata takes 1520 less damage then the squishy over 100 hits. This tips the balance in defense's favor by 15%. Creating an example of what would happen with rage would be pointless, as the chance to hit would still be 95%. With the same slotting rage still negates all defenses even with elusivity.

From these examples I’d dare to say that Castle got it right the first time. This is debatable, the Fortunata has 15% more damage mitigation, but is running several toggles/set bonuses while the squishy only runs one. The squishy also has other forms of damage mitigation; blasters do significantly more damage and have higher hit points then fortunatas do, corruptors/defenders/controllers get buffs, debuffs, and/or heals.

Some people have proposed moving elusivity to a middle ground of 20%. Using that example I’m fond of that would put chance to hit at 64.29. For sake of validity let’s take damage out of 100 hits.

Squishy
(200)*(1-0.42) = 116
116*100 = 11600 damage taken total
Fortunata
(200)*(1-0.1) = 180
180*64 = 11520 damage taken total

The difference from 100 hits would only be 80 damage. That is less then a 1% difference. IMO that is about as close to balance as can be asked.

I feel that when elusivity first came around, people were not objective on their complaints and didn’t look at actual damage mitigation provided and simply saw that they were missing a lot (duh the point of defenses). Did anyone truly look at the amount of damage that defenses were mitigating compared to resistances? Please discuss, I’m interested on seeing how people take this.

edit: adjusted numbers to reflect DR applied to to-hit buffs

The one thing that the nerf to Elusivity did right was change the way players slot their toons in zone. Most players no longer worry about having a decent base acc in their powers now since almost no defense toons remain. I've had excellent success with my Shield tank so far because of this. That toon clocks in at 50% melee and 58% range and rarely saw anyone other than SS brutes hitting me on greater than 60%. Being that you have posted such great numbers, perhaps I'll try and find someone to run some tests with me tonight and I'll report back with specific examples.

People don't like to miss. Like really don't like consistently wiffing. It drives them up the wall.

Missing means no status effects/debuffs are applied, for a very large number of toons that pvp that is paramount to their survival.

Very few toons have more than a 10 second window of high tohit value. The ones that did wreaked elusivity based toons in i13. ie SS tanks w/ tactics and FA.

Additionally attacks with higher base acc trumped elusivity, while attacks with base acc or lower were at a significant disadvantage. Maybe base acc attacks were WAI, but then the long list of attacks with a 1.2 modifier were OP'd.

I didn't look closely at your calculations to verify their accuracy, but I noticed you DR'd defense, but never applied DR to any of your tohit buffs or acc values (ie rage is not a 31% tohit buff in pvp). tohit and acc trump defense, but I think your results are skewing the issue.

when someone misses their procs miss too. pvpio and purple procs firing off every 3rd attack is something resistance based toons have to worry about alot more than defensive builds. to be fair you should probably add something in for that too.

Interesting. A rather nice, very thorough, analysis.

Yes, the effects of Rage versus Defense builds have been known about for a while. By being able to stack significant amounts of To-hit, a character can largely ignore Defense. This is something which Resist and Regen builds are less susceptible to.

This has lead to issues getting reliable information on how Elusivity is working. It's difficult to track how Defense is balanced when there is a very big divide between normal damage powerset performance versus the very common Superstrength builds.

The problem with defense as a survivability increaser, is that it is extremely susceptible to bad luck. The probability that a character will be defeated "on Time" is actually rather small. It's more likely the player will be defeated early or late.

In PVP, Since many attacks can deal extremely high damage, a bad streak of hits can cause a player do be defeated long before they should be. On the flip side, a lucky streak of misses can make a Defense build seem nearly invincible.

Ok, ran some tests in RV today. Without DR my Shield tank has 50% def to melee and 58% def to range. With DR running that drops to 33% melee and 37% ranged. I tested with a dom who had +17 to hit from tactics and kismet before DR. (forgot to ask him what that DR'd down to) With a melee ranged power that had 78% Acc and was running the tactics kismet combo he had a 47% chance to hit me. (That's vs my 33% melee defense) Turning off Tactics, his chances dropped to a staggering 34%. We then tested with Drain Psi as that was the only power he had with 95% slotted Accuracy. With Tactics running he needed 53%, without it, he dropped to 39%.

I then had him slap Waterspout on me to see how bad the debuff was. I dropped to 27% ranged defense, then it hopped right back up to 32% right away due to Grant Cover's debuff resistance provided and stayed there. It was difficult to test attacks on me since the debuff from spout appeared to be very short duration, so we were really unable to get hard numbers. I'll try and test with poison ray next time if possible.

I'd say that defense is alive and kicking, at least for shield tanks.

The only time I notice *Anything* missing my ninjitsu stalker was when I was running retsu or shadow meld, which brought my defence to around 39% to positionals, 1/2% variance. In which case I still got hit ALOT. Comparing it with my current build with sharks, I litterally get hit everytime practically. Tested it today, then sifted my combat logs for these numbers.


2 2 and a half hour(ish) sessions, 79, 83, 48 Kills.
1367 Hits.
389 Misses.
1756 Total.

For an approx 1 : 3.5 Miss:Hit ratio. Of course this is considering EVERY attack or power with a ToHit check thrown against and does quite include the general variance that is the "RNG" it does seem good on paper, but the truth is, this is quite horrible considering how squishy stalkers are and how defense is supposed to work, I think this is kinda bad.

question - since I cannot seem to monitor it - is elusivity enhancable? For one At it seems that it is.

One thought we had was to make defense a slightly different animal all together. Let it retain the ability to cause a total miss if the accuracy of the attack is too low versus the defense but allow for glancing blows and such.

The idea would be the more accurate the attack versus the defense based armor, the more damage can be delivered. This same idea can even extend towards secondary effects of the attack. So a power that slows or drains endurance or whatever would have an effect if it hit but would be diminished if it were not an accurate enough attack.

In this way we would not have such a binary result to the battles like we seen with Fortunatas when they were kings. Some of the calculations and such I seen in here were quite nice but I recall many times playing against Forts that were set built prior to DR and getting streaks of 30 plus misses with a character with 99% accuracy. Even missing all attacks on this same toon with build up active. It would have been much more realistic to have at least scored glancing blows that resulted in diminished damage.

This would make it more in line with resistance but dependent upon the accuracy of the attack as to how effective the result would be. Then of course the nemesis of defense would be the accurate opponent but the average accuracy types could still have a chance to play against defense.

I know this thread is old, but I saw it after doing a search for something else. I am glad this issue was brought up again.

I had argued back when i13 was on that Defense was not broken and elusivity was not broken. I was not alone, many others were saying the same thing, but it was not nearly as much as the other players crying or pleading that defense was broken. As Spiritchaser had put it, back then, too many players wanted the Devs to hand them on a platter the solution to deal with defense based players.

Yep, too many toons had builds for +rech, hit points, and maybe even +dmg. I had decided to post reviews based on duels I had. I racked up over 100 duels/team events after i13 but before i14 launched. After posting examples of Broad sword /regen scrappers, katana /regen scrappers, spines/ scrappers, blasters (fire, psy, even sonic), corrupters, and lets not even mention MMs who had beaten solid defense toons, it seemed that they were still in denial about exagerating Defense/elusivity.

I also wrote about specific examples of SS/ defense vs SS/ resist toons. I used an extreme example since Rage is a great tool for +to hit. The point was and still is that a defense based toon is facing a difficult battle vs a resist toon. The resist toon enjoys the advantage in that fight. Even when facing toons with Rage and I noticed there were not hitting as much as I expected, I would ask them how much global acc they had. Many did not even know till I asked them, and that's a clear sign that they did not plan their toon with global acc / elusivity in mind. Some did know but had given numbers of +16- +30, few were above +40 global acc.

Now I feel the majority of the people got what they wanted, a solution to defense on a platter, and given to them by the Devs instead of them readjusting their builds. I know Arcanaville had posted arguments of why elusivity was OP'd. At the same time I asked myself how much is Arcana PvPíng. By the sheer quantity of duels I had, and PvP events I had been in, I could tell it was not luck that many players were reliably hitting defense based toons and enough to produce solid results. I remember just predicting some fights between Forts and Spines/ scrappers who I knew had great global acc numbers. Not too hard to see Scrappers were hitting the Fort at full or near full dmg since there was no dmg mitigation when the Scrapper was getting through and Phase had already been used by the Forts.

Many practical results from different builds and sets motivated me to feel it was the builds from previous issues or the inclination to favor certain bonuses in builds or perhaps underestimating how important global accuracy now is as the culprit is a lot cases, if not all related to defense/elusivity. I remember some matches players telling me I was impossible to hit... yet, I had to hit hibernate two times in match. Strange but true that some players dont hit as often as they wish and feel they are not hitting. Defense is a gamble and wether you win or lose should be the result of the gamble. Nerf elusivity or defense and you tend to tilt the odds a bit too much vs defense.

Now, after seeing a lot on the forums and in practice, it is clear defense needs to be revisited and changes need to reflect current practical results, and not reflect posts by peeps who just beg for Devs to hand them the solutions to their PvP problems at times when they could readjust their builds.

This is a great thread and a great analysis by LuxunS, and Arcanaville's discussion of just why Elusivity failed in PvP 2.0 met with my own suspicions of what was going on, and increased my understanding all at the same time.

Here's hoping Elusivity can be raised to 20%, as that seems like the best solution in the current PvP design (just don't give that 20% to Forts).

WARNING: I bold names.

It seems the problem with defense (as alluded to by earlier post) is +tohit powers that don't have a window- (ie. rage and to a lesser extent FA and Tactics).

When people build they need to build for specifics and in many common PvP builds they build for +dam/HP/rec. Therefore in most instances they do not have +acc. So I think defense is working fine besides those that can stack tohit.

The fix I would try is increasing elusivity after the defense character is hit. More consecutive hits=more elusivity. I understand this may bring up other problems (such as tanks with mass hitpoints having mass elusiviy) so it might be like that Simpson's episode when they brought in the lizards to solve the pigeon problem.