Shield Block vs. Shield Barrier: The Numbers

[Breccia]

Kebess, you bring up some good points. I personally would hope that the damage on Heroic isn't double regular (which is what your numbers compared to Espada's suggest), considering the difference in ilvl isn't that big. Current raid logs suggest that a boss on Heroic does about 50% more than normal, excluding other raid mechanic changes like Zon'ozz. But a boss capable of 3-shotting a well-geared tank does fit the history of the last two expansions and I would not be surprised if it happened again. Disappointed, yes, but not surprised. This is the damage range I was talking about in my original post, the point at which Shield Block starts to take more total damage off the table.

I have no objection to healers preferring a slower rate of health decay vs a fast one after a resting period. The numbers I ran had nothing to do with that at all, and it's refreshing to get another viewpoint on the subject. Yes, the damage stopped by a "flat fee" such as Shield Barrier will always make a bigger change than a proportional reduction like Shield Block, and that could be a bigger problem than the pure amount of damage stopped. It could especially be important when it comes to comparing efficient versus fast heals, and at the start of a new raid tier, OOMing healers could very well be an issue.

The problem you bring up, unfortunately, pains a pretty bleak picture. 100% uptime on Shield Block is impossible, and our CTC is being gutted on purpose. The average total coverages I'm seeing hover around the 50-60% range. No matter how you spend your defensive rage, there WILL be at LEAST a 3-second period every 15 in which you cannot Shield Block, and based on the rage numbers I theorycrafted, it could easily be 5 or 6 seconds. There are unavoidable periods in which we are "defenseless" often in combat, and in that time, a boss could get two full-strength unblocked, unbarriered hits in a row around 4 to 10% of the time. Even with the absolute minimum gap of 3 seconds per 15 (which is difficult, from what I've heard), and a short boss fight of, say, 3 minutes (which also sounds low), you will see such a BAM-BAM two-hit combo about half the time. Extend the fight to six minutes, it becomes 65% chance or higher. If taking two hits in a row like that is fatal to the tank, then you're going to lose a lot of fights to the RNG Fairy, regardless of his skill or tactics, and I don't know what to tell you.

If such massive melee attacks make it into the game, you're going to see warrior tanks being spikey no matter what they do. And probably paladins too, based on their abilities. Maybe druids and possibly DKs. Monks, that stagger thing...no idea, haven't figured that one out yet.

We'll have to wait and see how much damage they actually do.

EDIT: Most of the heaviest-hitting blockable melee in DS comes in short bursts. The Stomp debuff of Morchok is a good example. So is 5+ ball stacks for Zon'ozz and being hit by the second Devastate by Blackhorn. A smart tank should save Shield Block and/or Shield Barrier for those times, rather than spamming it recklessly. This would put the tank in the situation you describe, namely, using Shield Block to ride out the short span of heavy damage rather than Shield Barrier to absorb just the first hit. There aren't a ton of Cata raid bosses who constantly did enough melee damage to two-shot the tank with back-to-back unblocked hits, and quite frankly, since we can't maintain defensies 100% of the time, I don't think there should be in MoP, either.

[kopcap]

Hold on. I'm not sure you can back that last one up. With current nerfs, he's hitting about 40-45k in normal and 65k in heroics in 10man. Grab a worldoflogs at random and see for yourself. Remove the current 25% nerf and that doesn't even let him break 100k.

I don't know how you are looking, ie http://www.worldoflogs.com/reports/m...=blocked+%3E+0

In other words, an un-nerfed boss, carrying around a big damage buff (which he will not have for very long), on the hardest possible difficulty. Since my main point was that Shield Barrier is better for all but the hardest of the hard bosses, I don't think this is the best example to use here.

Extremes are always great because they often show immediately if your model is any good at all. Second, this does not need to be anywhere close to 180k. Your model already fails at 80k. You just can't realize that 80k is nothing even in current content, let alone MoP. Third, that's only the tip of the iceberg. There are plenty of underlying fallacies that are simply miniscule to the main problem at hand, ie that a 40k reduction on 1/4 hits is the same as a 10k reduction on 4/4 hits from a survivability stand point. This is simply untrue.

I don't remember the adds hitting that hard, and I was the add tank for my raid groups for both normal and heroic, but maybe that's because I am a warrior and blocked just about everything

CTC was way out of reach back then and SB would only cover half of their uptime at absolute best. Worm would trigger pretty much instantly the moment that the second add reached you. Then you needed to chain cds to survive or risk getting GCDed.

[Breccia]

Well, the first way I would look at it is to say "Yep, that damage is higher than I said. Closer to 90k on average than 100k, but way over the 65k I was going with. I must have been looking at warriors spamming Shield Wall or something else similiarly careless on my part."

Then the SECOND way I would look at it is to say "Oh wait, is that a 2-second swing timer? It is, isn't it? My model was based on some comments, including GC's own comments, about a 1.5 swing timer with no attack speed debuff, and all of the associated numbers follow this. This changes rather a lot in the specific context we're talking about!"

And then I would re-run the entire calculation, INCLUDING the slower Rage generation of Shield Block because of the slower attack speed, assuming a 2.000 swing timer (which means no Thunderclap debuff, which is a dumb thing to do because I'm pretty sure bosses are adjusted based on an attack-speed-debuff in current content, but I'm not 100% convinced so I'm erring on the side of caution) and find that D, in such context, jumps to over 95,000, which is, incidentally, about the same as the average on that log you showed us. If I were instead to pretend that the boss damage values WERE adjusted around an attack speed debuff, i.e. a swing timer of 2.22, then D jumps further, over 105k, a number that I doubt even you can defend as "nothing even in current content".

And then I would go back to see if Morchok and 0-stack Zon'ozz were using the same swing timer. At first glance, I think they are.

Yeah...I'd like to act all smooth and claim I did all that on purpose, but truth is I didn't. Thanks for bringing it to my attention. You just made my point even stronger.

Oh, and don't go trying to steal Kebess' thunder about the "a 40k reduction on 1/4 hits is the same as a 10k etc. etc." thing. I fully admitted he had a very good point, you don't get to steal it hours later and pretend it was your "main problem" all along when it wasn't even hinted at in your first three bitterly sarcastic posts.

Finally, your "tip of the iceberg" comment suggests you have 36 more posts coming, truth is, it doesn't matter anymore. Ghostcrawler responded directly to my concern with the possiblity of changing the way Shield Barrier scales, which would make all of this math modelling no longer apply. I made my point, it's being looked at, and quite frankly, I can't ask for anything more than that even if nothing changes.

[kopcap]

Then the SECOND way I would look at it is to say "Oh wait, is that a 2-second swing timer? It is, isn't it? My model was based on some comments, including GC's own comments, about a 1.5 swing timer with no attack speed debuff, and all of the associated numbers follow this. This changes rather a lot in the specific context we're talking about!"

Ye, 2 sec swing timer can be a real shock for someone who never looked at a log before. Its not like 90% of bosses are like this after all. How does Blizzard dare disagreeing with you imaginary numbers?

INCLUDING the slower Rage generation of Shield Block because of the slower attack speed

Ok, would you share the delta with us and how you got it please?

If I were instead to pretend that the boss damage values WERE adjusted around an attack speed debuff, i.e. a swing timer of 2.22

They always are and that's the whole reason you see the numbers that you see.

hen D jumps further, over 105k, a number that I doubt even you can defend as "nothing even in current content".

Yet this is arguably the least demanding tank fight of the entire expansion bar Baradin Hold.

And then I would go back to see if Morchok and 0-stack Zon'ozz were using the same swing timer.

Glad you picked the most appropriate examples. Once you figure out "0-stack Zon'ozz", the rest of the fight is cake I guess.

Oh, and don't go trying to steal Kebess' thunder about the "a 40k reduction on 1/4 hits is the same as a 10k etc. etc." thing. I fully admitted he had a very good point, you don't get to steal it hours later and pretend it was your "main problem" all along when it wasn't even hinted at in your first three bitterly sarcastic posts.

Are you accusing me of reading his manuscripts? Really? o.O

Finally, your "tip of the iceberg" comment suggests you have 36 more posts coming, truth is, it doesn't matter anymore. Ghostcrawler responded directly to my concern with the possiblity of changing the way Shield Barrier scales, which would make all of this math modelling no longer apply. I made my point, it's being looked at, and quite frankly, I can't ask for anything more than that even if nothing changes.

I am glad for you. Would have helped if you did a little more research before trying to theorycraft next time. Topic has been covered to death both here and on TS months ago. And this is not a pun.