"Cold Blood helped make rogues not care about crit as a stat."

[DisposableHero]

Not quoting your math for space, but that doesn't make sense. I could copy that same argument for attack speed with haste.

The reason it doesn't make sense is that 1% crit (not counting bonuses and crap which you didn't either) is 1% more damage (because it turns 1% of your 100% damage attacks into 200% damage attacks). All you showed is that crit gives you more damage in number the higher your non crit damage is in the first place. Its not like there is actual DR on crit.

I mean come on, obviously if I hit harder, turning that harder hit into a crit will end up with a larger hit than the guy who hits for less. All you really showed is that crit scales in value in relation to your other stats.

Rogue baseline damage values were essentially lower because of the naturally higher crit chances, devaluing crit as a whole. Crit's desirability is measured against haste, mastery, hit, and expertise because those are the stats we are most commonly trading it for or trading for it.

Each of those stats in assassination has a cumulative effect with the others. If your attacks crit more often then increasing the number of attacks with haste is more valuable, and increasing their damage through mastery is more valuable because these effects are multiplicative. Similarly if you attack faster and for more damage crit is more valuable. Having a huge amount of baseline crit chance devalues crit relative to hit, haste, expertise, and mastery, and therefor makes crit a less desirable stat.

[Wow]

So by adding 10% crit to a base crit of 0% versus a base crit of 15%, you saw the damage go up more in the first case. Since crit rating can be exchanged for other ratings, Rogue-A might actually like crit, but Rogue-B will probably not want it that much.

This works out because the lower crit one has more base dmg. But in our case mutilate does the same dmg now as it did before 5.0, backstab does about 30% less dmg but both lost the crit, Ambush was the only one that got any dmg boosts but its use is very limited so its affect on stat weights is mostly nothing. Things like Poisons getting 200% crits helped but then were countered by things like the loss of lethality and other bonuses to critting.

Cold blood is like ambush here so rare that it didn't make difference.

That's why overall crit didn't go anywhere they removed more that they added.

[Deleted]

I don't see how that makes sense. Having a built in 15% crit to mut from the old talents doesn't effect the value of crit unless you are getting your mut crit chance over 100%.

Crit's value diminishes the higher it becomes. A very extremely obvious example would be 100% crit chance, all crit past 100% wouldn't do anything at all.

Now let's lower this still very extreme example just a tiny bit. If you have 99% chance to crit, getting another 1% crit isn't going to do a whole lot for your damage. For that 1% to be worth anything at all you need to not-crit in a fight if that makes sense. But because the chance for you to miss that crit with 99% crit chance is very low.

It's all probability based, the more chance to crit you have the less additional crit is going to be worth.

I know the math behind it, but I can't really explain things that well. Your Cold Blood example however works in the same way as the passive 15% crit chance to Mutilate.

[Squirl]

Removing all of our crit damage modifiers makes us value crit more? And why is it so important that we value crit...

Cold Blood was a pretty boring move anyway. I don't miss it. But their logic is inconsistent and flawed.

[Verain]

Not quoting your math for space, but that doesn't make sense. I could copy that same argument for attack speed with haste.

NOPE!

Because haste stacks MULTIPLICATIVELY. If adding 10% crit to a mutilate with 15% base crit made it crit, not 25% of the time, but 26.5% of the time, then you'd have an argument (to a degree). None of the "you swing faster" buffs are the equivalent of haste rating, because haste rating stacks with it. For instance:

Slice and dice increases your attack speed by 40%. So if you just walk up and right click, you attack at speed, say, 1. If you slice and dice, it's speed 1.4. 40% boost. If you had 20% attack speed from haste, then right clicking would be 1.2, but putting slice and dice up is 1.68 (1.2*1.4), not the 1.6 you would get if you just added 20% and 40%.

BUT PRETEND IT DID WORK LIKE THAT-

You would quickly realize that having a base amount of haste rating would actually devalue haste rating. In fact, this is true now- if you had 1 million haste rating, 200 haste would be far worse than 200 crit for combat.

All you really showed is that crit scales in value in relation to your other stats.

You almost got it here- having a huge fucking amount of crit on moves is essentially the same as having a massive pool of crit rating. Since ratings stack multiplicatively with each other but additively with themselves, you can see how having a huge amount of one (in all cases this is crit, none of the others have this issue) makes that rating (crit rating) junk.

[Stede]

Cold Blood was on a pretty long CD, and most rogues used it - you guessed it - on CD. It'd be great to have it back for PvP now that assassination is kinda viable and we'd all love to have some more burst options.

But no, they're not sentient. They don't even want to discuss it. They just wanna be assholes and make anyone who questions their awful design choices look dumb. Because they have never prioritized blance in PvP. This is the kind of crap I'd expect to hear from a dev or CM who knows nothing about rogues beyond what they dug up on an old wowhead tooltip before manufacturing such a bullshit answer.

---------- Post added 2012-11-15 at 07:12 PM ----------

I might have missed something, so could you elaborate when you say they've tried to fix the problem of crit? I don't recall anything being done to our crit-mechanics, apart from the removal of the +crit damage talents(and the removal of Cold Blood, I suppose).

And yes, Shadow Blades seems very tacked on. Again, outside of Combat, its purpose seems very unclear. It's been causing quite a bit of trouble for me, as Assassination, because I often end up capped on combo point and energy while I already have the Envenom buff up and running. I'm often forced to either clip the Envenom buff, to waste energy, or to waste combo points; with just a few crits, the cooldown completely breaks the fluidity of the rotation. It's probably more manageable as Sub, seeing as you can just Eviscerate to your hearts content, but it's quite annoying to deal with as Assassination.

SB is great for getting the necessary rotation up for Sub quickly. But for that, the CD is far too long, imo.

[Squirl]

Crit, hit, and expertise are the only binary stats. The stats just increase the chance that a move will flip in that that direction (crit, non-miss, non-dodge, respectively). Haste and mastery always work. Therefore, the value of crit is determined by the probability that it will flip your attack to a crit.

It's really simple to think about on an individual attack or on average, but when you're talking about the value of increasing/decreasing the probability, then things get complicated. More complicated than most people think. It's more math than can be explained here easily.

EDIT: There's also the rating tradeoff that people are talking about. This seems mostly irrelevant in practice as we get crit from agi, so our crit will never be so low that it will jump above other stats. We also almost never get enough of our best stat that the next best stat becomes better. There just isn't enough gear itemized like that with high enough ratings.

[Kaustos]

LOOOOOOOOOOOOOL WHAAAAAAT? WTF? Can GC REALLY be this oblivious? I cannot even compose a proper reply to this statement I'm so awestruck by its stupidity. Cold Blood had zero impact on stat values. Maybe it had to do with the fact poisons, which was a huge chunk of assassins damage, only crit for +50% and werent even benefiting from agi's crit? HMMM?

[Sesshou]

Rogue baseline damage values were essentially lower because of the naturally higher crit chances, devaluing crit as a whole. Crit's desirability is measured against haste, mastery, hit, and expertise because those are the stats we are most commonly trading it for or trading for it.

Each of those stats in assassination has a cumulative effect with the others. If your attacks crit more often then increasing the number of attacks with haste is more valuable, and increasing their damage through mastery is more valuable because these effects are multiplicative. Similarly if you attack faster and for more damage crit is more valuable. Having a huge amount of baseline crit chance devalues crit relative to hit, haste, expertise, and mastery, and therefor makes crit a less desirable stat.

I don't see how that first paragraph is relevant to what I said whatsoever... I also basically said what you did at the start of the 2nd. Baseline crit increases the value in those other things, but it isn't devaluing crit itself.

Crit's value diminishes the higher it becomes. A very extremely obvious example would be 100% crit chance, all crit past 100% wouldn't do anything at all.

Now let's lower this still very extreme example just a tiny bit. If you have 99% chance to crit, getting another 1% crit isn't going to do a whole lot for your damage. For that 1% to be worth anything at all you need to not-crit in a fight if that makes sense. But because the chance for you to miss that crit with 99% crit chance is very low 1% crit still isn't worth much.

It's all probability based, the more chance to crit you have the less additional crit is going to be worth.

I know the math behind it, but I can't really explain things that well. Your Cold Blood example however works in the same way as the passive 15% crit chance to Mutilate.

Ok so are you hit capped? Because you could sit 1% under hit cap and for that last 1% hit to be worth anything, you would have to actually miss. That argument makes no sense...

If everything else is held constant, going from 98-99% and 99-100% are the same in that 1% more of your attacks will land for double damage. If I hit for 1k and have 98% crit chance, the expected value for my attack is .98 x 2k + .02 x 1k = 1980. If I hit for 1k and have 99% crit chance, the expected value for my attack is .99 x 2k + .01 x 1k = 1990. If I hit for 1k and have a 100% crit chance, my expected value is 2k. As you can see, each crit % I add gives me the same return as long as other things are held constant.

Edit: Also when I said haste, I meant from rating, not mixing sources.

[Verain]

Don't bring up the agi->crit thing. The one dev who browses mmo-c rogue forums will ignore the rest of the thread, march into GC's office and bellow

THE ROGUES DON'T WANT AGI TO MAKE CRIT ANYMORE SIR

And that will be our last change for 5.1.

---------- Post added 2012-11-15 at 07:22 PM ----------

The math, by the way, is NOT "too complex to explain here". The math for the whole rogue rotation is, but something as easy as "If mutilate crits 15% of the time base, you get less out of crit rating than if it crit 0% of the time base". Granted, I haven't been able to explain this concept to Sesshou, but presumably I'm not breaking it down intuitively or something.

[Sesshou]

Don't bring up the agi->crit thing. The one dev who browses mmo-c rogue forums will ignore the rest of the thread, march into GC's office and bellow

THE ROGUES DON'T WANT AGI TO MAKE CRIT ANYMORE SIR

And that will be our last change for 5.1.

---------- Post added 2012-11-15 at 07:22 PM ----------

The math, by the way, is NOT "too complex to explain here". The math for the whole rogue rotation is, but something as easy as "If mutilate crits 15% of the time base, you get less out of crit rating than if it crit 0% of the time base". Granted, I haven't been able to explain this concept to Sesshou, but presumably I'm not breaking it down intuitively or something.

No, you don't... The expected value for your attacks does not get diminishing returns for increased crit. I just showed you that...

Edit @ Squirl
I don't see how it is too complicated. It is just an expected value problem. You multiply the probability of the event with the value of the event should that occur and then add them all together to get expected value.

[Verain]

As you can see, each crit % I add gives me the same return as long as other things are held constant.

But it's NOT constant. In my example, the two rogues- Rogue-A and Rogue-B- were both balanced to average 1000 hit per attack. Rogue-A did it with 0 base crit, Rogue-B did it with some base crit. The base crit on those moves was never free, and it raised the mulitplier coming from crit from, say, 1.0 to 1.15. This means that while you can still add 10 to the multipliers, it should be obvious that going from 1.0 to 1.10 is going to give you more bang for your buck than going from 1.15 to 1.25. Hell, going from 1.0 to 1.10 is 10% more damage- going from 1.9 to 2.0 (going from 90% crit to 100% crit) is only a bit more than a 5% boost. The damage increase is smaller.

But instead of being DESIGNED that way, you could be that way from stats. If you have a massive stack of haste rating, your crit rating will be better than haste, etc. The point is that having stuff baked in reduces the value of that, if you have a rating that just adds to it additively.


Can you tell me what is confusing about this? I'm clearly failing at this explanation

[Deleted]

Ok so are you hit capped? Because you could sit 1% under hit cap and for that last 1% hit to be worth anything, you would have to actually miss. That argument makes no sense...

Not critting is just a little bit of missed damage, while a miss from not being hitcapped is zero damage and some lost resources. But I'm not able to fully explain the differences when comparing hit to crit.

If everything else is held constant, going from 98-99% and 99-100% are the same in that 1% more of your attacks will land for double damage.

No, because there is a higher chance for your attacks not to crit at 98% compared to 99%. This is where probability kicks in.

Squirl's post perfectly explains how higher base crit will diminish the value of additional crit.

[Verain]

No, you don't... The expected value for your attacks does not get diminishing returns for increased crit. I just showed you that...

Go back to my top example, with the two rogues both balanced at the same 1000 per hit! It's not about "diminished returns". Holding everything equal, as you point out, each value of crit is the exact same. The point is, stuff isn't equal when they design the class! They don't balance us and THEN put 15% crit on mutilate, they balance us ASSUMING that, and this part of why crit rating is poop (and has been). Other things being stuff like "poisons crit for garbage damage".

[Anthas]

I get the impression that Blizzard just wants us to stop bothering them about rogues. :-/

[Shaley]

I saw this quote this morning and had a good chuckle. I always like to start my day with a laugh.

[Sesshou]

But it's NOT constant. In my example, the two rogues- Rogue-A and Rogue-B- were both balanced to average 1000 hit per attack. Rogue-A did it with 0 base crit, Rogue-B did it with some base crit. The base crit on those moves was never free, and it raised the mulitplier coming from crit from, say, 1.0 to 1.15. This means that while you can still add 10 to the multipliers, it should be obvious that going from 1.0 to 1.10 is going to give you more bang for your buck than going from 1.15 to 1.25. Hell, going from 1.0 to 1.10 is 10% more damage- going from 1.9 to 2.0 (going from 90% crit to 100% crit) is only a bit more than a 5% boost. The damage increase is smaller.

But instead of being DESIGNED that way, you could be that way from stats. If you have a massive stack of haste rating, your crit rating will be better than haste, etc. The point is that having stuff baked in reduces the value of that, if you have a rating that just adds to it additively.


Can you tell me what is confusing about this? I'm clearly failing at this explanation

Again, your example fails. OMFG yes crit is more valuable if your non crit damage is higher. Obviously the rogue with the higher base hit gets a larger gain. You can't compare those two as you changed other things around. All your example can ever show is that crit increases in value as other factors increase your damage.

Not critting is just a little bit of missed damage, while a miss from not being hitcapped is zero damage and some lost resources. But I'm not able to fully explain the differences when comparing hit to crit.



No, because there is a higher chance for your attacks not to crit at 98% compared to 99%. This is where probability kicks in.

Squirl's post perfectly explains how higher base crit will diminish the value of additional crit.

I know hit has a larger penalty. That is irrelevant. The original point was that at 99% crit you have only a very small chance to not crit and therefore can 'ignore' it, but you can't because it does occur just like that 1% miss chance if you go under hit cap.

You people... where probability kicks in? Do you know what expected value is? I showed you it is the exact same... I had the probability "kicked in" above.

Edit:
"For that 1% to be worth anything at all you need to not-crit in a fight if that makes sense."

Is what I was talking about. The exact same can be said for 1% hit. It isn't worth anything either if you can land all of your attacks without the last 1% hit just like if you can land all crits the last 1% crit isn't worth anything. Thats why you want to use expected values.

[Wow]

But it's NOT constant. In my example, the two rogues- Rogue-A and Rogue-B- were both balanced to average 1000 hit per attack. Rogue-A did it with 0 base crit, Rogue-B did it with some base crit.

You seemed to have missed this.

This works out because the lower crit one has more base dmg. But in our case mutilate does the same dmg now as it did before 5.0, backstab does about 30% less dmg but both lost the crit, Ambush was the only one that got any dmg boosts but its use is very limited so its affect on stat weights is mostly nothing. Things like Poisons getting 200% crits helped but then were countered by things like the loss of lethality and other bonuses to critting.

Cold blood is like ambush here so rare that it didn't make difference.

That's why overall crit didn't go anywhere they removed more that they added.

If they did it right what you said would be all good but they didn't do it like that.

[Deleted]

I know hit has a larger penalty. That is irrelevant. The original point was that at 99% crit you have only a very small chance to not crit and therefore can 'ignore' it, but you can't because it does occur just like that 1% miss chance if you go under hit cap.

I'm not saying you can just ''ignore it'' I'm saying 1% crit has far less value at 99% basecrit compared to 0% basecrit.

You people... where probability kicks in? Do you know what expected value is? I showed you it is the exact same... I had the probability "kicked in" above.

It starts at 0.0001%* crit technically, but it diminishing effects become more obvious past 50%.

Small obvious example:
1% crit gives 1% more damage at 0% base crit.
1% crit gives 0.5% more damage at 50% base crit.
1% crit gives 0% more damage at 100% base crit.

The more crit you have, the less it will add to your damage. This happens because the probability of not-critting becomes lower making that 1% crit you got not do anything.

*that was a random meaningless number. My point was, it starts the moment you get any crit at all.

[Squirl]

The math, by the way, is NOT "too complex to explain here". The math for the whole rogue rotation is, but something as easy as "If mutilate crits 15% of the time base, you get less out of crit rating than if it crit 0% of the time base". Granted, I haven't been able to explain this concept to Sesshou, but presumably I'm not breaking it down intuitively or something.

I suppose "prove" would have been a better word. Probabilities, man. They're the least intuitive math concept for me, but I suppose I "explained" it as well as I can.

Also, your explanation is simply a statement. It doesn't explain anything. I tried my best, but if the people in this thread don't get it, they will just have to trust us.