Etuy wrote:
Hum. Not 100% certain I agree with gbaji on that one (which is rare in itself).
CA and LR give you the chance to avoid an attack or attacks. And CS and ID mitigate the damage each attack does. Over a long enough period of time the amount of damage prevented is the same.
Yeah. And if they were treated the same in the actual damage calculations, you'd be right. However, unless they've changed the way the forumlas are done (which I haven't heard of, but admittedly haven't been actively been keeping up on), avoidance counts for roughly twice the "real" damage reduction over time as mitigation.
The reason is the way the combat formulas work. Actually, let me preface this by stating that the reason we believe the formulas work this way is the result of parsing the relative effectiveness of avoidance and mitigation. When parsed, against *most* mobs, avoidance is about twice as effective. The conclusion is that the damage formula first calculates a chance to hit, but then uses the result of that (how well you hit) as a factor when calculating the damage done. Thus, the benefit of avoidance comes into play in both parts of the equation, wheras mitigation only counts in the second part.
Let me explain a bit more clearly. First off, recognize that there are two components to AC. Avoidance and mitigation. There are *also* two components to ATK (part that goes against avoidance and part that goes against mitigation). Let's assume a situation where both parts of both ATK and AC are in balance with eachother. How this ends out working is that let's say a random roll from 1-100 is rolled to see if you get hit. A 1-50 is a miss (since we're equally balanced) and a 51-100 is a hit. The key part is that the result of that roll is then passed on as the base value for the damage roll. So if you rolled a 75, then 25 points of "hit value" are passed to the mitigation result. This one's done as a ratio though. The mitigation AC and ATK against mitigation are compared. The relative "goodness" of the result acts as a multiplier to the base value passed in by the initial to hit roll (the Avoidance roll). That's then applied to a table to determine the value of the hit.
What we find then is that if you increase your avoidance value, you get hit less often, but the resulting hits will *also* be of lower average value. There's some "odd" discrepancies in the hit tables though (for some reason max and min hits never seem to be affected), but all the in between values get shifted in probability. So the curve moves to the left and you get a lower average hit value over time.
This by itself isn't super significant, but is measurable. However, there's another factor. Damage from mobs is not a straight roll (it is when you hit them, but not the other way around). Mob damage has two components. A "base damage", and a "mitigation damage" part (those aren't the terms used, but I don't feel like digging through steelwarrior to remember what exactly they call them). Essentially, the Base damage is *always* done on a hit. The mitigation portion is split into 20 parts, each a set number above the last. This is the table I spoke of earlier. The probability curve is broken across these 20 possible results, a magic die is rolled, and that determine the additional damage done.
So a mob might have a base damage of 50, and a "+5" mitigation table. That means that each of the 20 results will increase the damage done by 5 points. The result is a damage range from 55 to 150, with damage *always* going in multiples of 5 damage (55, 60, 65, 70, 75, 80, etc... up to 150). Increased mitigation only affects the range on that table. So right off the bat, 1/3rd of the damage range *can't* be mitigated with the CS line of AAs. This, combined with the small shift effect caused by the avoidance pass-through to the mitigation roll, is why parsings almost always show avoidance to be better then mitigation.
Obviously, this changes based on the mob. Some mobs have higher base damage then range, or vice versa. Steel warrior has a table of common boss mobs with their relative values, and a resulting disc suggestion for warriors tanking them. Basically, a ratio below a certain point means you use defensive (increases mitigation). A ratio above a certain point means you use evasive (increases avoidance). However, for day to day exp mob tanking, every parsing post I've ever seen has shown conclusively that you are about twice as well off taking avoidance before mitigation.
Quote:
However, tanks in exp groups generally die due to damage spikes, rather than sustained battering. And I always think of mitigation as preventing those spikes, whereas avoidance can be a bit more flukey. So avoidance MIGHT avoid that big swing that finishes you off.... but mitigation prevents the swings previously from putting you in position where you could be one rounded.
This really realy depends on the situation. You're right that damage is "bursty". However, the bursts are almost always the result of getting a large number of average hits in a short period of time rather then a couple really big hits. Most mobs don't have such a huge range of damage that the difference between a min hit and a max hit will kill you. You die far more often by a long stream of hits.
This is easily seen with simple parsing. Interestingly enough, damage done is far less "bursty" then rate of hits. The next time you do a long solo fight, stop and go back and look at your damage window (easy to do if you're like me and filter "my hits" and "hits against me" into a single window (yes, this isn't a true parse, but just something to illustrate the point). Count out any 5 hits by the mob against you (5 is just a random number, take any you want). Look at the spread of those 5 numbers. Take an average. Compare them to the average for the whole fight. Odds are, any 5 hits in a row wont be far off the average hit you took during the whole fight (again, not a great parse, but good enough for this).
Now look at the comparison of hits. A simple way to do this is when you're soloing since it's just you and the mob. Compare the white lines to the red lines (your hits and his). Notice that there are sections where you'll see 10 red lines in a row, and sections where you'll see 10 or more white lines in a row. This is not super conclusive, since we're not putting timestamps and such in there. I'm just suggesting a simple and dirty way to see that the relative hits for and against can vary wildly in ratio. If you want, you can certainly run a parser, or just open up a log and do some quick manual checking. You'll see that the rate at which you get hit over any 10 second period of time (picked that randomly, but relevant to the time of a CH for example), will vary wildly. By simply glancing at a log output, you can clearly see that the damage you take over a 10 second period of time can vary wildly, but the average "per hit" over that same 10 second period doesn't. Clearly, it's not the mitigation that causes the burst, but the avoidance, Higher avoidance helps you avoid getting those 8 hits in a row that'll kill you.
Feel free to check this yourself though. Don't just take my word on it. This is just my observations from looking at damage outputs from many many combats. I *always* see huge variation in rates of hits over time within any particular fight. I very rarely see a huge variation in average damage across any arbrary "small" set of hits. It does happen, just not as often as the bursty numbers of hits...
Neither is perfect, of course. And you should aim towards getting both as quickly as possible. However, for exp mob tanking you will see much more bang for the buck going down the avoidance line first.