Your FF download isn't working for me, but, off the top of my head, since the first spreadsheet gives numbers that match the simulation testing that's been done, I'm inclined to guess that it's correct and you made a mistake somewhere. From your description, my guess is that you're mishandling a conditional probability somewhere, but, again, I can't be certain until the FF link starts working.

New_Darkmoon_Trinket.xls
Hmmm... maybe try this link?

Hmm, Well unbuffed in pve dps gear I'm at 23.43% crit. Add raidbuffs and We're looking at 29-30%'ish critrate. At which point this trinket is becoming rather inferior.

It might work as a alternative to hourglass, but bloodlust brooch+dragonspine trophy, later tsunami talisman+dragonspine trophy and later the Black Temple trinket.

Thread's a month old but I have a relevant question:

Did anyone ever test this trinket to see if it procs when you fail to crit because your victim's resilience converted a crit to a non-crit?

I'm trying to decide if this trinket is worth replacing Blackhand's Breadth as a viable counter to resilient targets.

To be honest it looks like the best Ret pally trinket in game.

Note for fury warriors (regarding Flurry, cirt = next three swings 25% attack inscrease):

When you check the probabilities with/without trinket that three attacks will NOT crit in a row (with a base crit chance of 30%):

WITHOUT trinket: 34,3% chance that 3 attacks in a row dont crit
WITH trinket: 33.1% chance that 3 attacks in a row dont crit

If you want a rough esitmation you can just neglect the chance it has to stop working due to the extra crit and assume with X crit chance it has X chance to wear off. That means on average it'll stack up to 1/(1-crit) so on average you will have (1/(1-crit))/2. While this is VERY rough, it gives you a very good idea if it's worth modeling any further. For a mage with 33% crit chance I get this trinket worth somewhere around 25.5 crit rating - pretty useless.

Has anyone done extensive testing on resilient targets? Does it proc off non crits converted from crits due to resilience?

Afaik it doesn't proc then. Things only proc from non-crits that are converted from crits due to resilience when you are on the recieving end. Things like blessed resilience and enrage proc but not things like clearcasting from shaman after a crit and flurry.

Crits aren't converted. Resilience gives the bearer a chance to proc "on crit" abilities when they receive a regular hit. There is no conversion, so the post above me is correct.

The problem with using this trinket as a Rogue, even a Mutilate Rogue, is white damage. While it's true that with Lethality and Seal Fate crits are proportionally more useful, those talents only apply to Special Attacks. The issue then is that white crits will constantly knock off the stacks of the trinket, meaning that your specials will have much less chance to benefit from any significant increase in crit.

I think slow hitting classes would benefit the most. A rogue, as mentioned above, may well find himself wasting the stacks on white and/or offhand crits and their crit chance is already high as it is.

A destruction warlock aiming to keep improved SB up as much as possible could well benefit a lot from it. I'm wondering though, even being an affliction lock I spend a good chunk of my time spamming Shadowbolt... supposing I get this trinket, would it be worth dropping Immolate from my 'rotation' so that it doesn't consume the stacks? Improved Shadowbolt uptime boosts my DPS by a LOT, not sure that would be reason enough to drop the Icon or even the Blessings deck trinket for it.

Actually, I believe you're mistaken. Crit based procs were modified to occur on crits that were converted to non-crits by resilience.

http://www.worldofwarcraft.com/patch...patch2p01.html
I believe beneficial effects for non-victims (Nature's Grace) also still occur.

I've yet heard of a Rogue who has gone for a Seal Fate + Backstab build, they all go for a Mutilate + Seal Fate build, so usually 41/20 or something like 41/3/17.

Most Rogues are going to be sitting at around 24% - 26% crit rate with good gear, and "high end" arenas players have around 260-325 resilience, which if I remember right is roughly 6% - 8% reduced chance to be crit, which would end up giving you around 2%+ crit from the Darkmoon Card.

I don't really have anything to buy at this point so I'll pick it up anyway just to play with.

Remember that 2% crit is what the UBRS trinket used to give you ;p
It's 44.2 crit rating for a trinket slot - you can really get much better much more easily.

Well, going by the calculations in the first page it'll be closer to around 2.3% - 2.6% crit gained against a high resilience target, and unless I'm missing an obvious trinket thats not from a raid instance beyond Kara the most you get is 1.45% from the Hourglass.

As I said, fun to play with anyway.

Used to give you

That's what I meant...

I'm not mistaken. This has been extensively tested and discussed on these forums. Resilience does not convert crits into hits. What they did when they made the referenced change was to give crit-based procs a chance to occur from normal hits, based upon your resilience.

How to prove this - take someone with a 0% crit rate (e.g. dot applications), stack your resilience gear and have them attack you. You will proc your "crit-based" attacks even though the attacker can't crit you in the first place. The procs are occurring off normal hits, based on your own resilience.

well, I decided to check if this trinket could be any useful to warlock, so I made two things:

1) my own XLS that calculate and show effect for different crit%: Crusader_Wrath.XLS
2) program that simulate same thing

I managed to get exactly same numbers in my XLS calculation, and in simulation, therefore suggesting that both are correct. Numbers are similar to those posted before for higher crit% numbers, but are slightly different for low numbers:

Interesting thing about XLS is that I managed to make calculation for one crit% in one row, without need for additional sheet with complete table for every crit% to calculate it. But in order for it to work, you need to turn on 'Iterations' in Tools->Options->Calculation->Iteration (mark checkbox, and make sure its 100 or more in 'maximum iterations')

BTW, my conclusion is that trinket is not good enough for demo warlock, since value of one crit point is about 0.6 dmg, so at average 15% crit this trinket would only be equal to 40 +dmg - much less that Darkmoon:Crusade, which for similar stats is equal to 77 +dmg. Even for Destro lock (who with similar other stats have crit valued for 0.72 dmg) this trinket would be under 50 +dmg equivalent, so still much less than Crusader.

Modeling the card as a change in the expect chance to crit is really the wrong way to do it becaues it does not change the expect chance to crit on a single spell. The card changes the average time between crits without changing the actual crit. You can pretend that this is just a change in the expected value by equating it with whatever actual crit rating would be required to net the same change in expected time between crits but this ignores the mechaincs of the card. With the card equiped the probability curve for expected time between crits changes shape, not values. Changing the average crit rate and changing the probabilty of long strings without crits are not the same thing.

Given the mechanics of ISB, where increased crit chance increase the chance ot proc AND increases the chance for ISB to end early, it seems like more investegation is needed to see if the card can be useful in downgearing crit on other slots and still maintaing a decent rate on ISB/sec.

I'll come back when I have harder data but I'm hoping a few others try and tackle this card from a differnt perspective than just change in effective chance to crit.

I dont agree exactly on this one tetra;p

Best way to model this card is to model along only thing card actually affect, and that is chance to crit.

But you mentioned "The card changes the average time between crits without changing the actual crit" - where first part is correct, and second part is not . namely, card does change (reduce) time between crits, but exactly due to fact that it increase crit%. And to be more precise, card does not reduce average time, since it is not related to speed of casting, but reduce number of needed hits in order to crit.

My XLS model work exactly in that way - I calculate change in number of hits between two crits, and calculate 'effective crit rate' based on that changed number of hits.

Also, notice that I said in my post that I made simulation program to check numbers. In that program I do not have any "model" or complex "calculation" - I exactly replicate stated effect of card, ie i count hits and crits, and if non-crit i increase crit% for 17 points, and if crit i reset to basic crit%. Again, my simulation numbers match exactly my calculation numbers from XLS.

Regarding effects like ISB, this card really increase number of crits in given time period, and really decrease number of spells (for example, SBs) between two crits, so it is very accuratelly modeled as increase in crit%. I agree that we should consider secondary bonus effect of this spell, ie increase in crit% will also increase ISB uptime. But that is true for ANY crit% increase, so it is safe to say that effect from this card is same as effect of item with 3% crit ( if you have 15% crit to start with, of course - since effective crit% bonus from card is reduced as your total crit% increase). Also, I dont think that "increased crit chance increase the chance ot proc AND increases the chance for ISB to end early" is correct - increased crit% does not increase chance for ISB to end early. It is increased number of warlocks in raid that increase both things (ie ISB proc and ISB cancelation), resulting in that limit or 'cap' that I was mentioning when I made simulation for ISB. But increaseing crit%, with same number of shadow users, only increase ISB uptime.

But since for any such item with crit% or crit points I check its value anyway by using crit_to_dmg ratio from my XLS, (example above was 0.60 for demo), and in that ratio both my stats and ISB effect of crit are calculated, when i say that card is equivalent to 40 +dmg, I also have its bonus effect on ISB calculated in that number. But main point here is: even if increased crit% have completely different effect on ISB than both mine and your version here, or if increased crit% has some 3rd effect, it wouldnt matter - as long as we express value of card in crit%. If , after that, we are able to accuratelly value crit% from ANY item, it would equally apply to evaluating this Card.


So, if you want to state effect of Card in most natural value, dependant on least number of variables (ie dependant only of your crit% stat), show it as crit% bonus (like those 3.01% if you have 15% crit). If you correctly convert that crit% bonus into +dmg bonus (just like for any other crit% item), other side effects should be included there, including increased ISB uptime, different talents (ruin or not) etc. In other words, if player have 15% crit, this Card value is 3.01% crit for ANY class and talent build and +dmg stats etc. For one specific class (like warlock) and specific build (like Demo) and specific other stats (like 1200+shadow, 10%hit), that particular 3.01% crit is valued as equivalent of 40+dmg. For other build (Destro) it is 50+dmg. For other class (Mage) yet more different - but all those differencies are due to 1% crit having different effect on +dmg (or DPS or...) for different cases, and in all those cases (as long as they all have 15% crit without card) , this card value is 3.01% crit.