Flexible Raiding by Rigid Numbers

Ion Hazzikostas, the lead encounter designer of World of Warcraft also known as “Watcher” on the Blizzard forums, jumped into an interesting thread about Flex Raiding and Magic Numbers. MMO-Champion has a convenient Blue Post watcher that sums things up.

Flex: Magic Numbers

The gist of it is that some people have noticed that certain mechanics become easier or harder depending on the number of people you take. For example, on Sha of Pride the number of prisons goes from two to three when you bring eighteen people. Norushen is another encounter which penalizes you heavily for bringing extra players, because as Watcher mentions, health of the bosses scales up, but you can’t actually stuff more people through the test realm to get purified because there’s a set number of orbs.

Bashiok argues:

“If a raid leader is attempting to build a rigid group to attempt Flex and is telling members of his own 25 Normal team to sit out because some lower number is optimal, that person is overthinking it.”

Which for Flex raiding is certainly the intent of the feature. And would be nice to have, but practice diverges from theory. The difficulty spike from two prisons to three on Sha of Pride was too much for our raid. When you have eighteen people, suddenly three prisons is a lot to deal with, and given the penchant for it to choose our more active raid members, it honestly stonewalled us for nearly two hours before people started to drop, at which point when we passed that magic number and had seventeen people, we one shot it from there.

Now, not all the fights suffer from this problem. For example, the entire second wing, excepting General Nazgrim, was relatively easy regardless of the number of players. At that point it just became a healing/DPS throughput game, and we had more than enough there.

Nazgrim himself was a bit tougher, though to be fair our first set of attempts had eighteen people, six of which were healers, and a few DPS below the minimum DPS threshold (more on that in a bit). The next week our DPS:Healer ratio was far better (nineteen people total, four healers), and we blew the enrage timer out of the water. We probably could’ve used another healer given how few people were alive by the end of that. But I would argue that’s still well within reason for Flex. You’re still expected to bring the appropriate number of players for each role, so that was on us.

But when the mechanics themselves unduly punish you for bringing a sub-optimal number of people? That sucks. Watcher is correct in saying “Boss mechanics inevitably involve some breakpoints (alas, there's no way for Malkorok to create precisely 3.7 Implosions),” but I wonder if there’s some other way to smooth out that difficulty spike when you wander over into 25-man mechanics. Or perhaps it’s just a matter of eighteen people being slightly too few to handle 25-man mechanics well. Perhaps they shouldn’t scale those linearly, and instead start them at twenty players? You still have the problem of a magic number at nineteen to twenty, but at least with the two extra players at twenty, mechanics like prisons or implosions become easier to deal with because you have enough people on the board to actually handle it.

Flex: Real Numbers

The other half of the story is Watcher has given us approximate numbers to play with in terms of boss health and damage scaling, so now we can actually calculate the minimum DPS threshold before a DPS player becomes a burden on the raid rather than a help. 

Watcher states:

“Thok, for example, has 355 million health with a 10-player Flexible raid. If you add an 11th player, his health will go up to 390 million. I.e., you've increased your raid's damage output by somewhere between 14 and 20% depending on how many healers you were running, but the boss's health only went up by 10%. Adding that 11th player will also make Thok melee for about 3% more damage, which will barely be noticeable.”

Now, I may be incorrect in extrapolating the 10% increase in health across all bosses, but I figured I’d use an easier boss with no adds to do our calculations. Specifically, Iron Juggernaut. Single mob, single health pool. Wowhead states that Iron Juggernaut has 361 million base health in Flex, and a ten minute enrage timer.

If we add nobody, we’re talking needing approximately 86k DPS per player (tanks included) for a basic 10-man raid consisting of two tanks, five DPS, and three healers, and that’s to kill it in ten minutes precisely. (361 million, divide by 10*60 to get DPS for the raid, divide by 7 for players).

For every player added, we add 10% to the boss’ health pool, which is 36.1 million health per player, which amounts to ~60k DPS per added player to ensure they’re helping the raid rather than hindering.

Except that’s not enough! When we add healers, they aren’t really helping DPS at all. If we take into account adding a healer every five DPS (so that we’d have five healers for a 25-man raid, sufficient for Flex if they’re good to great healers), you’d have to up the amount of damage done by each player 20%, bringing our total baseline requirement to be 72k DPS per new player. And that’s just to hit enrage.

Now, 72k DPS isn’t that much to a relatively skilled player. Heck, I ran in on my 502 ilvl Frost Mage and pulled 92k on that fight in Flex, according to our World of Logs parse. But Flex is also billed as something where you wouldn’t leave your friends or family behind if their output or skill is subpar. So what do you do when you’re hitting enrage on Nazgrim over and over again because you have three players pulling 30k each along with those six healers? Turning a healer into a DPS--even a subpar DPS--is probably better, because then you still have the body, but you’ve made up slightly for the DPS deficit and don’t need to boot anybody.

I can tell you that people get incredibly frustrated quickly when they run up against an enrage timer. At that point you know you’re executing the mechanics correctly, because if you weren’t you’d have been dead well before enrage (like watching five Ravager axes whirling through the raid within a minute of the fight starting), but then it becomes a question of “How do you eke out more effective DPS?” Either players have to improve (which is not impossible, but see the conversation between Milady and myself on my Motivation post), or folks leave. Healers or DPS who are below the minimum threshold, as each player leaving is less health on the boss and the adds.

I think there needs to be a bit more patience walking into these raids. There’s still progression of a sense. They’re definitely a lower difficulty level than normal, but definitely more difficult than LFR, which is great, so folks should expect wipes going in (and do), but when you hit that enrage timer wall over and over again, it’s not just the boss that gets enraged.

Flex: The Reality

Let me be clear, Flex has been absolutely awesome. That level of difficulty between LFR and Normal was sorely needed, and has been great. Folks are jazzed at being able to raid in a more relaxed atmosphere, and the fact that we don’t need exactly ten people is fantastic. People can join in or leave as the night goes on. It’s really, really awesome. But there’s still those rough edges that cause frustration, and hopefully mechanics can be tweaked, either now, or in future designs, such that we can smooth out these frustrations. It’s also nice to participate in a raid where I’m not the one raid calling (though old habits die hard and sometimes I accidentally take over that job, whoops!).

As Bashiok states, we may be “overthinking it,” but as raid leaders who want to ensure some amount of success to our raiders, our job is to overthink it. We just need to provide Blizzard the feedback to ensure that we don’t need to anymore.