Defense or Dodge - Here are the graphs

The problem is, the formula is not totally linear. Defense versus dodge is easy, just keep the ratings roughly equal as you improve your gear. Parry is trickier. As your gear level increases, the proportion of parry that is beneficial actually rises. At 700 defense rating + 700 dodge rating, 200 parry rating is more valuable than 100 defense rating + 100 dodge rating. At 1400 defense rating + 1400 dodge rating, it would be closer to 500 parry rating for maximum avoidance.

Honestly, I doubt we'll be driving our ratings anywhere near that high. We need to remember to balance our stamina and strength gear contributions as well.

I updated the original post with what I think are decent guidelines for increasing your avoidance. A quick note, 700 defense rating, 700 dodge rating, 200 parry rating, and 200 block rating comes out nicely above 102.4% with a bit of extra to account for estimated miss DR. (assuming I have the right numbers, as I'm at work and can't look them all up)

If someone could give me the proper base and talented numbers, I can update the spreadsheet to include them (I've got block factored into my current version that I haven't had time to upload).
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

It's not that a formula can't be made, just that it would be somewhat unwieldy. My gut tells me that spreadsheets may be easier to work with, but I can play around with it a bit and see what comes out.

In other news, I'm in the process of rebuilding my spreadsheet to include more information. It should end up being a one stop shop for comparison of two pieces of plate tanking gear. It should work for all three plate tanking classes if I can get all of the info in it. I'm a bit torn as to the format that things should take. I may just do the web based thing in the end. I just need to get all of the data looking the way I want it first, then I can worry about the final presentation.

I'm trying to integrate EH, TTL, and the like into the spreadsheet. It should make for far more interesting comparisons than pure avoidance. Pure avoidance is interesting as far as theorycrafting goes, but choosing between avoidance and EH is far from obvious at the moment. We can really only go with what appears to be good until there's been a good bit more number crunching done.
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

We can certainly derive formulas for that purpose and narrow scope simplifications may be feasible. As I think back to my original post, that was my original plan, but I went with the spreadsheet because I was feeling lazy and didn't want to think.

As far as spreadsheets are concerned, 90% of the time when someone wants my advice (being the primary resident theorycrafter in my crowd) all they really really want to know is which piece of gear is better. I spend a lot of my time in WoW receiving whispers from my fiancee with two items linked as she turns around and says to me, "which one is better?" Spreadsheets are just a convenient way for me/others to be able to answer that question without actually thinking. (now I just need to make a spreadsheet for every class... she's got her rogue and dk at 80, hunter at 77, and a whole lot more waiting in the queue at 70.)

[edit] and to be certain, my fiancee is an excellent wow player (better than me at most classes), she just can't stand the theorycrafting and leaves that to me. (and she somehow managed server first rogue/belf/horde to 80 on Shadow Council)
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

Have done the same calculation, and do not necessarily come to the same conclusions.

Probably because I started the analysis from a slightly different premise, which was looking at the amount of avoidance gained per rating point.

First off, it is pointless to look at the contribution from dodge rating without already taking into account the fact that we have to be crit capped.

So, for all intents and purposes, that means starting the analysis at a defense rating of 689 and including our base agi (I included the leg armor enchants in my calculation as I cant see any tankadin not using them).

This means that before you even get one point in dodge or parry rating, you are already seeing DR on ALL ratings (except block)

By the time you are crit capped, defense rating has changed from providing ~.0244 pure avoidance per point to providing ~.0220 pure avoidance per poiint which, is almost exactly the same contribution you get from dodge rating.

Now, I will caveat this by stating that it assumes there is no DR on contribution to miss% as no one has had time to determine if there is a cap on miss, and if so what it is.

Effectively, this means that dodge/defense rating are ~equivalent once you are crit-capped and that is without considering what defense contributes to block-cap.

Any direct parry/dodge rating you get will detract from defense contribution to avoidance on a point by point basis.

Frankly, until there is documented proof of a cap on miss, I have trouble recommending dodge rating over defense rating at any point in the curve, unless you scale defense to incredible levels, which current gear doesnt really allow you to do.

I do agree with the emphasis that parry rating is to be avoided as it suffers way too much from DR.

If you stack any real amount of parry, you will definitely benefit more from dodge rating than defense rating.
_________________
Garath Lvl 80 Paladin, Gorefiend

While you are for the most part correct, there is a lot more interplay between the ratings than I think you realize. We operate under the assumption that you are already uncrittable. If you are not yet block-capped, the value of defense is significantly inflated due to the block contribution not being affected by diminishing returns. Once you are block-capped, the game changes entirely.

There is most certainly a diminishing returns effect on miss, and this has been stated as fact in a blue post. Ziggyunderslashone did some data collection with several defense levels, with the largest samples being around 20000 attacks (which is statistically significant to a 99% confidence level). His values for the diminishing returns coefficient had some variance, but we can be fairly certain that the true DR coefficient for miss is close to the value he determined.

There is a complex interplay between dodge, defense, and parry ratings. As your defense rating is already around 700, the DR effects (especially on the miss and parry contributions) lead to dodge rating being your most valuable avoidance stat. (if we are talking pure avoidance; that is you're block capped with block rating) Dodge rating ceases to be the most valuable rating at around the time you have equal dodge and defense ratings (at which point the two are roughly equal). Whether this is by blizzard's design or coincidental, I can't say. I should also point out that the balance is not at 100% equal, but it is a simplification of my results in the graph in the OP.

Around the point when you have 700 defense and 700 dodge, there is some value in parry rating. The DR on dodge and defense are significant enough that it is feasible to stack small amounts of parry rating for maximum avoidance gain (somewhere around 200 rating).

My spreadsheet linked in the original post will let you play with the numbers and verify this for yourself. The formulas are based on whitetooth's work (for Ratingbuster) and I believe he has verified that they match the character sheet. My only addition to his work was the integration of ziggy's miss DR numbers. There is some level of uncertainty in the miss numbers, so don't take them as rote, but they should be reasonably close to the actual numbers in game.

There is room for improvement in the miss DR calculations, but as of yet no one has offered to spend hundreds of hours taking hits in the name of science.

[edit] I should clarify that the key to the numbers working out the way that they do is in fact due to the fact that not only does miss have diminishing returns, but is affected by diminishing returns significantly more than even parry.
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

[quote="Elsie"]

[quote="Trixia"]

Mob level does matter in the miss calculation, as does dual wield and the occasional mob buff, which makes me leery of using random wws parses. I have seen talk of using pvp to gather data, and assuming that there aren't any differences in the mechanics I think it's a promising way to do it. That would keep everything constant. Perhaps I can borrow my fiancee's rogue for a few hours.
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

Yes, the math is all known, but a random wws parse won't tell me what level the mob is (and I'd have to look up the dw and any specials for each mob). It's just cleaner to keep the data to one source, so that is what I'd prefer to do. Also, assuming the incoming hit mechanics are the same in pve and pvp, I like the idea of using a rogue because their mechanics are well understood.

If anyone takes rogue or other dueling data before I do, remember that when dual wielding special attacks and white attacks have different miss rates, so stick to white attacks (I think blade flurry is ok to use though). Avoid any procs that might affect hit/miss rates. Using naked rogues removes hit rating from the picture, so if using multiple rogues, same level, weapon skill, and points in precision should be ok.

[edit] forget everything about using pvp hits to determine miss rates. I forgot that there are ongoing discussions about melee hit caps appearing to have changed. (not to mention the wide variations in the reported caps that people are seeing)
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank