Defenses, Elusivity and actual damage mitigation comparison

Specifically what killed forts was that nerf in both Defense AND their Damage, one or the other mighti've been fine, but both....

Its impossible to "erase" Elusivity. If you have two characters, one with 60% resistance to damage and one with 60% elusivity, then any increase in tohit or accuracy will affect both the same way - to a limit. Technically speaking higher accuracy means you get hit more often and you could consider that "erasing" Elusivity but then the same thing is true for Resistance: higher accuracy causes you to take more damage, which is "erasing" resistance as well.

This works to the limit of about +90% accuracy. At that point a zero defense/zero elusivity character is being hit 95% of the time (50% * 1.9 = 95%) which is the tohit ceiling. More accuracy can't hurt the resistance character, but it can continue to hurt the avoidance character. Under that limit, though, Elusivity reacts "smoothly" to increased accuracy and tohit.**

The real question is how to value debuff avoidance relative to damage stability and burst damage resilience against each other. If they are worth exactly the same qualitatively, then X% resistance is balanced against X% elusivity. If they are not, then one of the two would need to be adjusted. For example, if foe debuff avoidance is considered highly valuable in PvP (which it is) then 60% Elusivity is probably "more valuable" than 60% Resistance, even though the numerical damage mitigation is the same. In that case, you could decide that 60% Resistance should be balanced against 50% Elusivity, but that's less than ideal for a lot of reasons. A better option is some combination of Elusivity and Defense that has the same mitigation as 60% Resistance, but is more vulnerable to tohit buffs could be generated that balanced against 60% resistance (i.e. 20% defense/33% Elusivity has about the same 60% damage mitigation as 60% resistance). That way, the resistance character is vulnerable to all debuffs, while the avoidance character is less vulnerable to all debuffs, but more vulnerable specifically to defense debuffs and tohit buffs.

So if you compare to an attacker without very high tohit or accuracy, they will deal about the same amount of effective damage to both the 60% resistance character and the 20%def/33%elu character. If they have stacking debuffs, those would be more effective against the 60% resistance character. If they have tohit those would be a little more effective against the 20%/33% character. But usually, those advantages would be proportional, not runaway. Most PvPers would have at least a little of both, which would offset to a high degree. But avoidance characters would be stronger against some things and weaker against others, without being impossible to hit or trivially easy to tag in most situations.

Overall that's very possible to balance correctly, given a specific valuation target. Its not generally possible to balance this in a way that everyone will agree with because everyone has their own ideas of what different things are worth in PvP. But that's not a numerical problem: that's a "people can't agree where to go for lunch, much less what's important in PvP" problem.



** Above the +90% accuracy limit its a bit trickier to balance things. DR could be engineered to specifically moderate those extreme possibilities, or a hyperaccuracy critical system could be implemented in PvP that would eliminate the disadvantage completely.

You haven't countered anything I said. Rather you have decided to pick a specific nit and go off on a tangent thinking I meant "make disappear" when I simply meant "make not matter".

But rather than go around in circles here's what I'm requesting of you. Pick a value for it that you believe works in pvp that doesn't just create a new minimum barrier to entry.

Since we are specifically concerned with acc lets ignore the difference in tohit that would also be present.
Pick a value that would allow for a toon with FA + 80% global acc+ 80% acc slotting = 180% acc to not be significantly better than a toon with just 80% acc slotting.

Specifically assign a value that you think would be balanced, if you think it should vary set by set then feel free to give an example for say SR and w/e else you like. I don't need an explanation of how it stacks up against resistance, I already know who it should work.

Extra points if you don't slap a cement shoes DR curve on acc like was done to tohit buffing or else it is just falling victim to the last thing I said in my previous post. Or is the 60% example you used your answer?

Neither tohit buffs nor accuracy make Elusivity "not matter" either, short of supersaturation. Its not nit-picking: Elusivity is specifically designed to be, at least in terms of damage mitigation, as immune to tohit buffs and accuracy as resistance is, again up to the saturation limit of +90% accuracy or equivalant in accuracy and tohit. From that point forward Elusivity is still far better than nothing and still better vs tohit buffs in any case, up to very high saturation limits.

To make Elusivity really not matter would technically require numbers approaching 400% accuracy.


There is no such value. But that request does not have to be satisfied to generate a balanced game. For one thing, there's no requirement that someone with more than twice the accuracy must be no more effective than someone with half the accuracy. The only requirement for good game design is that the advantage isn't magnified to absurd degrees, like tohit buffs were originally. For another thing, you're suggesting that the game should be balanced on the assumption that the only thing available to players in the invention system is accuracy. It also has defense. If you're going to assume that players can get +80% accuracy from inventions, that should come with a comparable assumption about the amount of +DEF they can also get, because +DEF will benefit Defense sets more than non-Defense sets. Suppose I were to guestimate that a Defensive player - lets pick SR for simplicity - can get at least +10% defense from the invention system for about the same effort as +80% accuracy.

We start with an SR scrapper with about 30% defense. I'm going to round off just to make the numbers simpler. That's about 60% damage mitigation, equal to about 60% resistance in terms of damage mitigation only (this does not count foe debuff protection). For SR to be roughly balanced against its own intrinsic damage mitigation (or if you prefer, a hypothetical 60% resistance character) it should be true that the SR scrapper with +10% defense should take about the same damage as a 60% resistance target when attacked by an attacker with +180% (2.8) accuracy.

The resistance character is supersaturated, so a solution that will work for all attackers can't be formulated that also extends to this example, but we'll assume that this example *must* be balanced for (that's a premise I don't agree with completely, but I will accept for the purposes of discussion). We'll solve for this situation first, then see where the offsets are.

The resistance set is basically being hit by 38% of the damage output from the attacker (0.95 * (1 - 0.6) = 0.38). That's our target for SR

If we converted the entire SR set into 60% elusivity, then the net tohit would be 44.8% ((1-0.6) * (0.5 - 0.1)) * 2.8 = 0.448. That's slightly too weak. Better would be 43% Elusivity and 15% defense. That generates a net tohit of ( (1-0.43) * (0.5 - 0.1 - 0.15) * 2.8 = 0.399. That's close enough, and 43% elusivity and 15% defense vs an attacker with *no* enhanced accuracy generates a net tohit of (1-0.43) * (0.5-0.15) = 0.1995 or 19.95% - almost exactly the number we want, which is 20% net to hit for something with 30% defense vs zero tohit or accuracy.

So, specifically for your example, 43% Elusivity and 15% defense. However, I should point out that were I to balance the sets, I would not focus on that one extreme case only. I would average out a number of them. But actually, the numbers above do come very close to the numbers I originally proposed for SR long ago (or rather, one alternative version of them), which were 12.5% defense, 45% elusivity (I deliberately picked round numbers to make life simple, and they were reasonably close to the values my calculations required).

This makes Avoidance *fair* relative to Resistance, for even the extreme case of someone walking around with 280% accuracy. That's all the math I'm prepared to do. It does not make someone with 280% accuracy no better than someone with 80% accuracy. I have no idea how to do that, or why you would want to do that, so I'm not prepared to even try to figure out the math for that. But that has no bearing at all on whether its possible to balance avoidance with Elusivity in PvP. There is no "new minimum barrier to entry" with these numbers. Someone with no accuracy at all would hit this Elusivity-enhanced SR scrapper at exactly the same rate (20%) that they would hit an SR scrapper with just SOs, no power pools, and no invention bonuses. That is how often they are *supposed* to hit SR when they have no accuracy.


The math to make this work actually specifically only works when DR (at least, the current DR with the current values) *isn't* in force, at least to the degree it is now. Elusivity and DR are actually incompatible on a very deep level unless the DR parameters are set in a very specific way, which currently they are not.


And I have very much countered something you said; specifically this:

It does not. It is specifically engineered to prevent such things. In the case of tohit and defense, with the original mechanics even small amounts of tohit and defense could have huge consequences, so it was imperative to ensure that someone couldn't stack just a little amount of either and significantly negate a large percentage of the value of the other. In other words, if you build up to 45% defense, someone packing just 15% tohit can negate the vast majority of the value of the entire pile of 45% defense: 15% tohit causes them to hit you four times as often. From there, there is only another factor of 2.5 before you reach zero defense again.

*That* was the primary motivator for the tohit/defense arms race: that and people just didn't like to miss. Elusivity defuses most of that in two separate ways. First, you can't buy more to stack with it. There *is* no fuel for that end of the arms race. You can still buy more defense, but with no one anywhere near 30 or 40% defense anymore, its not as attractive to buy "that last 5% defense" to reach the soft cap. The soft cap becomes very far away, even for avoidance sets.

Conversely, tohit isn't as binary as it used to be. Back when people "deep floored" their builds, if your target was at 55% defense, the first +10% tohit did *nothing*. But every point after that did huge proportional damage to the soft-capped defense. So it was to your advantage to use as much of it as possible, because anything less might be worthless. But everything else was devastating. Under high Elusivity/low Defense situations, you don't get that effect anymore. When you take deep-flooring away, tohit always works. But it never works very much better than accuracy does, and neither ever gets any stronger the more you buy. There is no special incentive to target a specific amount of accuracy, like there is targeting a specific value of defense (45). More is always better, but a little more is always only a little better.

People can still go crazy building for every high accuracy if they want to. But the arms race has no accelerator: it has no prize at the end of the race. Building for high accuracy (or high tohit) has not much better return than building for more damage. And that was the point: the arms race would be moot.

Why does the arms race still exist? Mostly because PvP has Elusivity in the same sense that ACE Hardware has a fashion department. Its too weak, too untargeted, and too unfocused to have the benefits it was designed to have. But that's a problem with the numbers, not with Elusivity itself.


Your turn:

Now you've seen values. Find something that, by any reasonable definition you care to choose, can "erase the elusivity value." Clearly, 280% accuracy does not make Elusivity "not matter" in this case. So how many builds are there that can basically make these numbers all but worthless? I can imagine some that would be a little stronger against these numbers than a non-defense set, but to get anywhere near "worthless" you'd have to do a lot better than that.

lol, what?

@macskull, @Not Mac | XBL: macskull | Steam: macskull | Skype: macskull
"One day we all may see each other elsewhere. In Tyria, in Azeroth. We may pass each other and never know it. And that's sad. But if nothing else, we'll still have Rhode Island."

I don't expect it to not make a difference, but Castle obvious does if we look at how tohit buffs have been altered.

There are toons capable of self capping their tohit (ie pb+aim+bu) and there are toons with no tohit (ie lots of troller builds). In such an environment defense can either be hugely overpowered meaning you never hit them, or it can be hugely negated.

In came the jar of mollasses and that blaster now hits ~20% tohit buff with those same powers. Now the haves are barely better than the have nots.

Acc is in a similar position though not as blatantly strong due to its multiplicative nature rather than its additive nature. Except in the case of elusivity where it is additive vs acc.

With a range possibiltiy of 0-200% acc you get vs 43% elusivity
(1-.43)*.5 = 28.5% tohit chance
vs
(1+2-.43)*.5 = 128.5% chance to hit.

How you are going to even out the haves vs the have nots without something as confining as the DR curve put on tohit I'm unsure.

The result outside of a paper analysis in actual implementation is that either I hit you and you die and go home crying on your defense toon ******** about how def/elusivity sucks and Castle stole your lunch money, or I don't hit you and I die and go home crying about how OP'd def/elusivity is.

I've never heard any sort of consensus among players about how often you *should* get hit, in fact the answers given when Castle specifically asked about some defensive scenarios varied wildly.

The disconnect is this: If elusivity is properly implemented, building +acc into your toon will shift your net hit rolls a few % one way or the other but there's no amount of +acc you could IO with that would make elusivity not matter.

What is true for +tohit vs defense is not true for +acc vs elusivity.

But with +DEF and the old mechanics, that was easy. With Elusivity and the new mechanics, it can be made really really hard. Separate from the Elusivity calculations above, we could place a shallow DR curve on Accuracy strength that didn't kick in until about 180% Accuracy (+80%) and made it difficult to get much higher than about 250-300% accuracy.

200% accuracy (+100%) is not a problem, because that's still about in the range where accuracy affects both avoidance and non-avoidance sets (since everyone starts at base 50% tohit, everyone can be hit at least about twice as often). Its accuracy above that level that is the problem, because it ceases to have an effect on anything other than avoidance sets, no matter how small. With invention set counterbalancing, we can creep that up to about 250% accuracy or so, but then we need to start diminishing the return on investment for ultrahigh accuracy. But only at those stratospherically high values, and not with as sharp a curve as most of the DR parameters specify.

Its not a problem for people to try to build to hit as often as they want to. The game only needs to make the costs reflect the benefits and the limits of the game. I think the biggest problem with DR is that the devs tried to use it to keep everyone in the "average" range, when they should have used it to make getting out of the average range expensive. That's a subtle, but important distinction. By making hyperaccuracy expensive instead of impossible, it could be balanced against the opportunity costs of buying other things, like more damage, rather than focusing strictly on the issue of whether it is absolutely fair to acquire such accuracy against an avoidance set.

Short version: Elu + the right DR means *some* people will do as you describe, and buy accuracy to hit at all costs, but many others will do the cost/benefit analysis and decide its not worth it. If the game presents that choice in a way that people choose both ways, rather than everyone thinking that one of those choices is obviously the best one, that would act to help balance the PvP environment.

Usually, when it comes to PvP, the trickier but more important thing to balance is not the effects of things, but rather the choices presented to the players. That makes the "arms race" a diversity race, and not a magnitude race. Its ok to buy an edge against something if it also costs you a vulnerability (or less of an edge) to something else, as long as those strengths and weaknesses are moderate and within the ability for the game environment to counterbalance reasonably well. Its not necessary, and usually not desirable, to force everything to be exactly equal.

I don't think the premise that you can balance around expense holds true at all. If a given strategy is better and it is possible then that is where most builds will cluster no matter what the "cost" because in the end the cost is only time and play money--and we know well that the time factor can be gamed down to a minimal level.

Unfortunately that reasoning is all too common in dev circles... "well, we know this is potentially OP'd but we thought the price tag would balance it." That never works.

I agree that you can't balance on time to acquire capabilities - which is really all that "expense" translates to in this game - but the "price" that Arcanaville is talking about is not influence cost, but opportunity cost. As in, because you slotted for lots of A, you don't have any room to slot for much B - no matter how much influence you spend on enhancements.

@macskull, @Not Mac | XBL: macskull | Steam: macskull | Skype: macskull
"One day we all may see each other elsewhere. In Tyria, in Azeroth. We may pass each other and never know it. And that's sad. But if nothing else, we'll still have Rhode Island."

If I can make a build (and I can, quite easily) that will hit you at the same frequency regardless of whether you have elusivity or not then it indeed does not matter from an end user point of view.

You'd need about 150% elusivity to prevent me from just building in global acc and hitting you like you have no elusivity. I don't mean "no elusivity" from a mechanical standpoint, but from an end user standpoint.

The build sacrifice necessary to do it is irrelevant, because if you are granted such a high value of elusivity that requires me to overcome that will become the new paradigm for pvp. You need look no further than i13 forts for evidence of that. People were rapidly constructing builds to counter them. If it was allowed to persist it would have happened too, however you'd have an even higher barrier to entry and a new paradigm for the meta game to counter.

At the same time that 150% elusivity would absolutely devastate lesser builds, which the devs seem really devoted to protecting.

edit: another example of a new paradigm where build sacrifice is irrelevant is kb protection. 41 points is ridiculous, but it is pretty much necessary if you want to compete.

They likely used it to average out performance because that is way easier to (theoretically) balance around compared to widely variable values. It is exceedingly difficult to balance defense in an environment where people range from 0% tohit buff all the way up to ~180%. Accuracy is in a very similar boat, but more constrained by the base tohit value of 0.5 so that it is half as rewarding. That said you can generate nearly twice as much acc as tohit buffing so in actual practice it can still vary considerably.

They have done much the same thing with average resistance because it is much easier to balance expected defeat rates when people are taking on damage at a more predictable rate. A squishy with 0% res has a drastically different expected survival time than a tank with 90% res (even ignoring hp). If you apply a mechanic that adds res to the squishy and subtracts it from the tank you bring the damage mechanic into a narrower range and it is (theoretically) easier to balance.

Castle has done this with nearly ever mechanic in pvp now. And yet builds still emerge as definitively stronger than others because until you constrain everyone to a variance of nearly 0 it is inevitable that something will be better than something else. The nature of pvp mandates that the better option will be what survives.

Without being rude you aren't the first person to suggest that a magic pony would solve all problems.

If something is even a tiny bit better pvp will flush it out and it will become more widespread than the choice that is a tiny bit worse. The more prevalent the slightly better choice is the more people start building to counter it. There is likely no way to avoid that cyclical evolution short of reducing everyone to the exact same abilities.
I agree, but this was for better or worse far more of a factor in i12 pvp than it is now. For a host of reasons that system did not flourish and in an attempt to average everything out through dumbing it down the only result has been explicitly an arms race.

How this relates to elusivity: I'm not saying to do away with it as it is a layer in the balance equation that ironically is very elusive. I'm just saying that unless Castle throws his hands up in defeat like he did with tohit buffing and reduces everyone to a very narrow spectrum then elusivity is either going to be too good vs casuals, too weak vs experienced players and very rarely just right. I think we can all agree that the value assigned to it now of 10% is just a place holder while Castle waits for the magic pony to arrive and help him out.

My second point is that hitting it with a DR curve as restrictive as tohit is more of a sign that it can't be balanced than a sign of proper balance moving forward; unless proper balance is viewed as brawl+sprint and a total lack of diversity.

True, but my magic pony only claims to solve one problem, and it comes with specific numbers.


But even by your own admission, there is another factor: people do not like to miss. Elusivity is designed to change the game so that what is numerically optimal and what is necessary to achieve the best possible to hit are not the same thing. That's significant, because now you can't assert both that people will just do whatever it takes to hit as often as necessary, which sounds like the tohit ceiling (see below) and will do whatever is necessary to get the maximum possible performance in PvP.


I don't understand this statement. The only real way to build enough accuracy to hit me at essentially the same frequency with or without Elusivity is to drive tohit to the 95% ceiling with enough accuracy that tohit stays at the ceiling with or without Elusivity. I think you should explain this statement: I provided example numbers for SR that you could use: 15% defense, 43% Elusivity.

Actually, now that I read further I think the problem may be a mathematical misunderstanding on your part:

That's not how the math of Elusivity works. Accuracy and Elusivity take effect separately, not additively:

Without any accuracy, net tohit with 43% Elusivity is:
(1 - 0.43) * 0.5 = 0.285

With 200% accuracy, meaning +100% accuracy, net tohit is:
(1 + 1) * (1 - 0.43) * 0.5 = 0.57

At least, that is how I described it to work and how I tested it to work when it was first implemented. I haven't specifically looked at the implementation recently, but if its doing what you're implying in your math that is a horrendous bug that crept into the implementation. The 200% accuracy person gets to hit the Elusivity character twice as often, which is about what his 200% accuracy allows him to do to everyone else, which is the basic intent of Elusivity. Its not intended to neutralize the advantage of high accuracy, its just intended to reduce the impact of high accuracy (actually, specifically high tohit in combination with high accuracy) to have a similar impact on avoidance sets as non-avoidance sets. 200% accuracy is still allowed to be better than 0% accuracy vs all targets regardless of mitigation types, which is also intended.


One part of the magic pony that I think a lot of people overlooked the first time around when the Elusivity idea was discussed, so it was amplified in all future discussions, is that it wasn't a defense "buff." That implies defense was just made stronger. It didn't do that explicitly. Defense might have gotten better under certain conditions, but it didn't get universally stronger. More importantly, under the right selection of numbers tohit *also* becomes more effective. Under the original mechanics, high avoidance could be generated in only one way: with high defense. Tohit could only be countered with even more stacked defense but because there is a tohit floor, high defense designed to counter high tohit would also completely neutralize low tohit. So it wasn't just Defense that was broken under the original mechanics, it was tohit as well. More precisely, both moderate defense *and* moderate tohit could be simply made to have literally zero effect.

Under Elusivity, the intent is that low avoidance has a low effect, high avoidance has a high effect. Low tohit has a low effect, high tohit has a high effect. Elusivity doesn't judge which numbers are fair, and which are unfair. It just makes sure the numbers always have an effect. The specific question of whether a defense set has too much or too little avoidance is a question for the powerset designers to address by setting the numbers to the desired values.

1. Any melee in-set defense toggle or epic/patron defense toggle. These give elusivity to the same positions or types they provide defense for in PvP. There are some exceptions - AoE defense buffs (Dispersion Bubble, Mind Link, VEAT Maneuvers, etc.) do not provide elusivity, nor do some in-set defense powers such as Cloak of Darkness or Heightened Senses. Personal Force Field provides elusivity in both the in-set and epic/patron versions, however.
2. Generally there is no way to stack elusivity because it is only provided by in-set toggles. The exception to this is Energy Aura, which has elusivity granted in Overload in addition to the elusivity found in the other defense toggles in that set.
3. No.
4. No.
5. Yes - see number 1. For example, Power Shield gives defense against fire, cold, energy, psi, and negative in PvP, therefore it grants elusivity to the same. Evasion gives defense against AoE in PvP, therefore it grants elusivity to the same. It's important to note that the types of damage a power protects from in PvP are not necessarily the same as what it protects from in PvE. Power Shield provides psi and negative defense in PvP but does not in PvE.

@macskull, @Not Mac | XBL: macskull | Steam: macskull | Skype: macskull
"One day we all may see each other elsewhere. In Tyria, in Azeroth. We may pass each other and never know it. And that's sad. But if nothing else, we'll still have Rhode Island."

With regard to the magic pony: the cost/benefit analysis doesn't come into play. Hitting someone in pvp is the most important factor for a number of reasons. There is no benefit currently implemented or likely to be implemented that makes hitting someone less desirable relative to something else. Which is why what you said is a pipe dream and will never happen.

With regard to acc: When I say 200% accuracy I mean +200% accuracy not +100%. If I mean +100% then by proxy me also stating 0% accuracy would mean a power that can under no conditions ever hit.

To be more specific the range you are dealing with in pvp realistically goes from:
+60% acc (casual zoner) all the way up to ~+220% powers like KO blow on an IO'd toon)

The affect of acc is automatically cut in half by the 50% tohit chance vs PC's so you have an effective range of 160/2 = 80% accuracy variance that you will encounter in a pvp environment.

The variance is too large for elusivity to be balanced to not feel retardedly overpowered vs a casual and majorly underpowered vs an heavily built toon. Which once again is pretty much exactly what we saw when it was first implemented.

It needs a DR curve that cuts the variance in half which would mean putting a ceiling of about +140% acc (so 2.4 acc) before a single value of elusivity stops the extremes.

Once the extremes are eliminated you could pick a singular value for elusivity that would perform more reasonably.
ie 25% elusivity
So for the casual they'd hit a character with elusivity
1.6*.75*.5 = 60% of the time before def and tohit is factored in
The top end build
2.4*.75*.5 = 90% of the time before def and tohit is factored in

Vs Now with no acc DR
1.6*.9*.5 = 72%
3.2*.9*.5 = 144% (capped)
Vs i13 with no acc DR
1.6*.7*.5 = 56%
3.2*.7*.5 = 112% (capped)

*I shut my brain off earlier with regard to 150% elusivity, the number in my head was 50%, not sure why I threw a 1 in front of it. Sorry 'bout that.