Monday, January 26, 2015

[WoW] Is Raiding With 10 Harder Than 30?

So one of the things I've been mulling about in the back of my mind is in World of Warcraft, are current raids more difficult with fewer people? When you look at Highmaul, most mechanics require a static number of players to handle them, regardless of raid size:
  • Kargath - 5 players go to the stands
  • Butcher - Most strategies require the movement of 2 players
  • Tectus - Each shard requires a ranged player to deal with Crystalline Barrage
  • Brackenspore - Two players are utilizing Flamethrowers to keep moss at bay
  • Twin Ogron - Enfeebling Roar is distributed among 10 targets maximum (20 for Mythic) - WoWhead data is wrong on this.
  • Ko'ragh - Every time Ko'ragh goes into recharge mode, you need to use a (preferably) ranged player to use Nullification Barrier to soak Overflowing Energy
  • Imperator Mar'gok - Usually need 2 players to handle Branded
The more players you have, the less you notice the absence of players handling special mechanics. For example, in a 30-player raid, losing two DPS to handle Flamethrowers is trivial, because you're probably running 22 - 23 DPS, meaning at worst you're only losing 10%. In a 10-player raid, you're only running 5 - 6 DPS, so you've now lost 33% - 40% of your total DPS.

Now, granted, in the Brackenspore case they get a buff to offset DPS lost, but that's not always the case. Tectus at some point you're dealing with losing 5 ranged players to running around dodging Crystalline Barrage. In a 10-man raid, that means you may end up with all of your ranged DPS and healers running around only able to toss out instant-casts!

Similarly with Kargath, losing half your raid (more than half your DPS if you're currently running 3 healers) every time you get tossed into the stands is a significant loss of DPS on the boss. Twin Ogron's Enfeebling Roar becomes trivial to soak if you have a big enough raid, as your melee/tanks will basically auto-soak it by virtue of 10 players near the boss, with nothing extra required to handle it.

I understand why the developers/designers wanted to avoid numerically scaling mechanics: it creates optimal inflection points (like the ye olde 14 players of SoO Flex) that correctly or incorrectly, raids will balance themselves around. But arguably health on some of these bosses should not scale linearly. The difference between adding an extra DPS player on Kargath means you've added 50% extra damage during the stands phases with 11 players (having 3 DPS on the ground), versus adding an extra DPS player from 29 to 30 players, which would be an extra 6% or so damage during the stands phases (with 19 DPS on the ground).

Given each Chain Hurl you're expected to be in the stands for the 45 seconds, and starting at 1:30 into the fight you see a Chain Hurl about every 2 minutes, a 5-minute fight you're short DPS for 30% of the fight. Meaning for 10-player, you're short a total of ~15% DPS or so, compared to 30-player where that would work out to a ~1.8% DPS deficit. Adding an extra DPS to make an 11-player raid would reduce that to ~7.5%, and from there the number begins to decline.

Not quite sure what happened at Players = 20. I checked a couple logs and they both had that dip.
However, digging through WarcraftLogs, I found the health values for Normal Kargath across all raid sizes. It's clearly a linear progression (approximately 3.3 million health per player added). This suggests that fewer DPS players need to do a fair bit more DPS to make up for the lack of players. Thus, Kargath is quantifiably more difficult for 10 players than he is for 30. The more players you have, the less DPS each player needs to do in aggregate.

Granted, especially on lower difficulties, many raids just ignore the stands completely (healing through the extra damage), but the fight as designed expects you to send folks to the stands, and for initial kills that's likely the case.

The same pattern can easily hold for pretty well every mechanic I've listed in the preamble. I just wish I could filter WarcraftLogs' data for average ilvl for kills over a time period per raid size so I could either support or debunk this argument. In theory, if the average ilvl for a given raid size is smaller than a different raid size, if we have enough samples we could say that it is likely easier--caveat certain raid sizes for Mythic guilds, as they tend to be above 20 players on Heroic mode so they can gear their bench, so data may be slightly skewed.

How's Healing Affected?

Healing is a lot harder to quantify, as lots of damage tends to be optional. However, more healers means more overlap. When you're running two healers, that leaves with little to no room for error. When you're running with 5 or 6 healers, yes, there's more damage going out, but you can only amp up tank damage so much before we're back in two-shot territory circa MoP, and even if you say, double the damage taken between 10 players and 30 players, the fact that you have twice as many healers staring at your tanks' health bars means if one healer gets distracted for a moment (like having to run away from Tectus' mechanics), you're less likely to lose that tank.

Then there's also the age-old issue of losing players. Thankfully the battle-rez charging mechanic they added helps this issue quite nicely, but you still have the issue where at the end of the day, if you're down two players in 10-player raids, that's 20% of your raid, versus two players in 30-player raids, that's only 6.67% of your raid.

About the only thing that is harder in bigger raids--aside from having to organize that many people and analyzing what went wrong is probably nightmarish--is space. Fewer players means fewer people overlapping for any mechanics that require spreading (i.e: Expel: Fire on Ko'ragh, or Pulverize on Twin Ogron), but that goes back to the fact that when you have 2 - 3 times as many healers, you need to amp up damage somewhere, and if you aren't increasing it on the tank, you're increasing it on the raid, and a lot of these overlap mechanics are where you'll see the increased damage taken to keep those extra healers busy.

Conclusion

So even ignoring the touchyfeelycraft around tank damage and healers, DPS at least shows a quantifiable increase in difficulty for smaller raid sizes that is currently not being taken into account by Blizzard's raid encounter designers if Kargath himself is any indication. Note that the more stringent DPS requirements means that if they're not being met, the strain will be passed on to the healers and their mana pools, as they'll have to heal for a longer amount of time.#WorldOfWarcraft, #Theorycrafting



UPDATE: It's clear between the discussions on Twitter and Bakoth's comment below, my point was lost in the minutia of my post. It's not just that raiding with 10 is harder than 30, it's that it's excessively more difficult.



The above graph is the DPS loss charted over the number of players, (and you can clearly see in the graph when I add an extra healer). For a 10-man, the players need to play disproportionately better than the players in a 11-man, and then again for a 12-man. The DPS loss is a curve. If it was linear, then yes, I would agree that all is fine and dandy.




This graph is the DPS deficit each remaining DPS would have to make up (this is mildly faulty given that the 3 players in the stands can increase their own DPS in the ~70% of the fight they aren't in the stands, but I don't really care to do the algebra when this will still illustrate my point accurately enough). This graph is clearly an inverse exponential curve.

The remaining DPS must make up a 9% deficit each for 10-player. But 11-player needs only to make up 5% each. 12-player, 3.21%. 13-player 2.25%. Once we hit 14 or 15 players, we're close enough to linear that I would say it would meet the goal of making things a little bit easier as your raid gets bigger.

Kargath, due to the nature of 3 DPS rather than 2 in most other cases, is a more extreme example. Luckily, he's the first boss of the tier, so this difference is largely steamrollered by the fact that, well, he's quite an easy boss on both Normal and Heroic. However, my point stands: static assignments disproportionately punish the smallest raids.

Is it a huge differential? To a Mythic guild bumming around in heroic, no. It's not. To a guild that's struggling to hit their DPS targets to begin with (take a hypothetical Kargath-like fight later in pretend Tier-18), you have fewer players, and all of them must improve a significant margin.

Is it even worth changing the mechanics? Honestly, probably not. But a simple tweak in health scaling at the lowest end would smooth that curve out a little. Leave it such that more players means an easier raid, if that's Blizzard's goal, by all means, have at it. However, the current style of mechanics aren't really solving the "inflection point" issue they had in SoO Flex. Instead, it's just pushed the inflection point such that optimal raid size is greater than 13 or 14 players.

33 comments:

  1. I've maintained for a long time that 10-player is the superior raiding experience. Since Cata merged the difficulties, the larger group size almost always made the game easier for everyone, especially for healers. Probably the only role that might need to do more in a larger group is the tank.

    Do they still reward more loot proportionally for the easier group sizes?

    ReplyDelete
    Replies
    1. Nah, loot distribution is now n / 5 pieces of loot, plus a % chance to drop an extra based on n % 5.

      Basically, if you have 17 people, you'd get 3 pieces of loot, and have a 40% chance (17 % 5 * 20% = 2 * 20%) pieces of loot.

      I'm not sure I'd go so far as to say 10 is the superior raiding experience. I certainly prefer it myself (I have difficulty keeping track of 25 people in the raid at once for on the fly analysis), but Tanks change the least arguably. They just take more damage, meaning they have to lean of external healing more.

      Rather, healers change the most (as you scale up, your healers will rely more and more on AoE healing as it'll be more effective), and DPS and healers tend to deal with more overlapping mechanics (ie: anything requiring spreading out is more difficult with 30 than it is with 10).

      Delete
    2. "Since Cata merged the difficulties, the larger group size almost always made the game easier for everyone, especially for healers."

      HAHAHAHAHAHAHA.

      Good one.

      I led a 10H guild from beginning of Cata to when they merged the sizes. We had it easier in raids, especially healers, particularly for the hardest bosses like Ragnaros and Lei Shen. Garrosh is about the only fight where 10 man was actually harder (technically some easy fights like Immerseus were harder on 10 man but...they're so easy it hardly matters). Well, I suppose 10H Madness was harder, but 25H Spine was way harder than 10H Madness.

      And this isn't even a matter of opinion -- can break down and analyze the numbers needed, coordination required, etc.

      Look at stuff like 25H Baleroc, even, where 25H guilds had to kite the boss like 30-45 seconds PAST the berserk because the DPS check was so insane, whereas 10H could kill it before berserk. And that's purely because the 25H version required so much more per person.

      "Do they still reward more loot proportionally for the easier group sizes?"

      Nope, 10H got less loot per person since it was easier (both in difficulty and coordination).

      That's not to say 25H Baleroc was harder than 10H Ragnaros, mind you, just that 10H Rag < 25H Rag. Both were still the two hardest fights in Firelands.

      Delete
    3. Load of crap, the only reason why 25 would be harder is the fact that it's harder finding 25 great players as opposed to 10.

      Raided DS both 10 and 25 and in comparison 25 was childsplay the only reason we even wiped ever was due to the noobs we just kinda brought in to fill the ranks

      Delete
    4. "the only reason why 25 would be harder is the fact that it's harder finding 25 great players as opposed to 10."

      Or you could actually read what I wrote and notice how 25H often literally required like 5-10% more DPS per person on tight DPS checks. Or realize that doing the Rag triangle was harder with 25 than 10. Rag 10 was still hard...just not as hard as 25.

      "Raided DS both 10 and 25 and in comparison 25 was childsplay"

      You did 25H DS prior to the nerfs and thought H Spine was child's play, eh? That's the "load of crap" here. 25H Spine was the hardest fight in there followed by 10H Madness.

      Delete
  2. "Thus, Kargath is quantifiably more difficult for 10 players than he is for 30"

    This is entirely intentional. Blizzard wants to encourage people to bring more people into groups -- not enough raid leaders, too many people wanting to raid. They want to promote larger social groups rather than a 10 man clique that doesn't want to bring anyone else.

    That 10 man clique can still exist, but it'll be a bit harder because Blizzard wants it to grow in size.

    This same thing happened back in SoO -- except boss HP/damage didn't increase linearly, it increased at a decreasing rate. Same result, more people equaled easier.

    ReplyDelete
    Replies
    1. I buy the idea that 30 should be easier than 10. That's not that big a deal, makes sense, even if I don't really like it. However, for the smallest raid sizes (10, 11, 12), it's not just harder, it's disproportionately harder. See the addendum/update to this post for details.

      Delete
    2. Yeah, but you're still not factoring in tanks/healers when you say stuff like this:

      "The difference between adding an extra DPS player on Kargath means you've added 50% extra damage during the stands phases with 11 players (having 3 DPS on the ground)"

      If the raid is 2 tanks/3 healers/5 DPS like you said, then at that point you have 1 tank/2 healers/2 DPS on the ground, right?

      Well, for starters, that tank is about 75% of a DPS, so you really have 2.75 DPS on the ground. So already that means adding a DPS is only a 36% increase, not 50%.

      Then those healers are going to be bored out of their minds with 2 of them to heal 1 tank and 2 DPS (people are in the stands clearing out those mobs in this scenario so there's little raid damage) so one or both will/should be throwing in some DPS, which is probably another 0.25 DPS or so. That brings us down to gaining 33% more DPS at that specific point in the fight on Kargath specifically, not 50%.

      Something I'm also curious about is whether the mob HP in the stands is static due to 5 people or if it scales up as well -- meaning those 5 people need to stay up longer in a larger raid size.

      But even laying that aside...

      "However, for the smallest raid sizes (10, 11, 12), it's not just harder, it's disproportionately harder."

      I'd honestly say that's a very good thing. Get people out of the mold of "Find exactly 10 people and go." You should be willing to take anyone who's capable to a normal or heroic group, but you still see groups saying they're "full" with 10 or 11 or 12 people.

      And of course I assume you remember the "Can't go past 14 people or we get 25 man mechanics" disease in SoO. The sooner we can cure people of that plague the better.

      "Hey, I have a 570 friend who wants to do SoO Flex with me, can you invite him?"

      "LOL nope we don't want 25 man mechanics."

      On another note, I found this twitter comment by you to be interesting and probably true, not something I had really put into a coherent thought before:

      "Mythic-level raiders won't notice this difficulty bump in raid size. 15% increase in difficulty on Heroic is negligible to them"

      Delete
    3. From experience I can tell you when progressing, those healers are definitely not bored out of their minds. Our first forays into Highmaul, we struggled to meet the healing demands of folks on the ground, be it from gear and/or skill. Now that it's on farm, we could 2 heal it easily, but none of our healers have a DPS off-spec. Bring the player, not the class indeed.

      As I mention, and as you bring up, if you're over-skilled for a given difficulty level, this is not noticeable. However, the grand majority of the WoW-raiding populace is not.

      True on the Tank DPS, that's an excellent point. Mind you, statistics on Warcraft Logs are showing tanks at around 60%, not 75%, of single-target DPS, but that's definitely still not negligible in a small raid. That would certainly offset the issue to a degree.

      I mention the SoO inflection point numerous times. And I'll state it again. They've not removed the inflection point. They've just moved it such that optimal size is now greater than 13 players.

      Mayhaps with the tank values added in, that inflection is reduced to 12 or so. I disagree that raids smaller than that should be disproportionately punished, but I think that's a value we shall have to agree to disagree on. Yes, fine, bigger raids can be easier than smaller raids, but for smaller guilds that don't have the roster to jump past that 10 - 11 player barrier (perhaps because they're just IRL friends and family) shouldn't be unduly punished by innate designer mechanics for choosing to raid with their friends and family.

      Delete
    4. "those healers are definitely not bored out of their minds"

      I'm talking specifically about when the group of 5 was in the stands and killing the adds. That means there's no raid AoE (since the adds up top are fighting the group that got thrown up) and the tank damage isn't enough to keep two healers occupied.

      In all seriousness, what were they possibly healing at the time? 2 healers in a group of 5.

      Regarding Tank DPS, here's our most recent Butcher kill:

      https://www.warcraftlogs.com/reports/fP1Tc3ymrzZYndpa#fight=9&type=damage-done

      Our average DPSer did 32.6k. The average tank did 24k. That's 73.6%, which is pretty close to 75%.

      Regarding the infliction point, unless I missed something, you said the following:

      "it creates optimal inflection points (like the ye olde 14 players of SoO Flex)"

      14 players was *never* an actual inflection point of significance. There *were* no major mechanical shifts there. 18 was an inflection point, got the third Sha prison. 25 was an inflection point, got the 4th Sha prison and all Garrosh MCs. But 14? Nope.

      To quote Blizzard...

      http://us.battle.net/wow/en/forum/topic/10894620000?page=2#23

      "all of the abilities with actual breakpoints going from 14->15 are things with very minor impact (Protectors' Shadow Word: Bane, Nazgrim's Bonecracker, Hisek's Multi-Shot, etc.). There are indeed a couple of more impactful breakpoints on Garrosh's Touch of Y'Shaarj and Sha of Pride's Imprison, but those both occur at raid sizes well above 14, and randomization isn't the correct solution to either.

      The 14-player "magic number" is actually a fairly interesting social dynamic, since there never was anything particularly special about the number, and now there really is objectively no advantage."

      Going from 14 to 15 was no different from going to 10 to 11 or 20 to 21 or whatever. It wasn't some optimal stopping point. Saying 17 was the magic number would have made far more sense, or 24 people.

      "They've just moved it such that optimal size is now greater than 13 players."

      The optimal size has *always* been more than 13 players, that's the thing.

      "I disagree that raids smaller than that should be disproportionately punished, but I think that's a value we shall have to agree to disagree on."

      I'm not saying I think that smaller raids should be disproportionately punishing. I'm saying that I think Blizzard should design bosses for larger groups -- and if it winds up being a bit harder for groups that insist on being exactly 10 people, then I don't think drastically changing the mechanics of the fight to compensate is worth it.

      "perhaps because they're just IRL friends and family"

      Sure, but doesn't it seem that they have another friend or two they could bring in if they're really having that much trouble? It's not like they need to maintain a raiding roster or something, just have people broadcast on BTag "Hey, could use another person or two for N/H Highmaul if anyone's interested."

      One last note: we're mainly talking about Kargath -- and he's a fight that's so easy that it hardly matters if you raid takes 5 minutes versus 6 minutes to kill him.

      The other main example is Tectus -- which is, what, the healers/ranged kiting the barrage for 10ish seconds twice during the fight? Tanks get hit for less anyway with few people and even with like 16 people all of the healers could get targeted.

      Just doesn't seem like we've seen this be an actual big deal that is a hindrance to groups yet.

      Delete
    5. RE: Tank DPS + Skill Disparities
      You're running under some faulty skill-level assumptions. Remember, Normal and Heroic are not tuned for the best-of-the-best players. One could argue Normal isn't supposed to be tuned for even the top 50% of raiding players (note, not all players, just raiders). Though, that's a question only the WoW devs could answer.

      Your anecdote about the tanks being 73% of your DPS is fine. You're Mythic players. The data for raids in aggregate still shows at the 50th percentile of tank single-target DPS in Normal for successful kills of The Butcher (the closest thing to Patchwerk we have this tier), it's closer to 60%.

      My point being, when tuning for raids, *especially* at the lower difficulty levels, you're not tuning for ideal or maximum conditions. You're tuning for what players at that level are capable of, hence my point about Mythic players not being able to notice even a 15% increase in difficulty on Heroic. Mythic players are generally the top 10% of the playerbase (or better!). You're outliers when it comes to tuning lower difficulties.

      Mind you, in our case, our healers were struggling, it was our first week progressing, folks were way undergeared, tanks and healers still getting a feel for raiding this expansion, and a number of other things. We still downed him eventually once we rebalanced who was healing what--the stands on Normal need almost no healing at all, versus the tanks and healing the raid back up sine the healers were behind from before getting tossed into the stands. Once we got our shit together, so to speak, the fight was indeed trivial. But you cannot use your experiences as a Mythic raider as a proxy for the experiences of a much lesser-skilled raid working on a raid that's two difficulties below yours.

      RE: Inflection Points
      Agreed that there wasn't actually an inflection point at 14, but inflection points did exist (I am aware the 14 was an illusion perpetuated by hearsay), therefore the precise inflection is immaterial to the argument.

      RE: Friends and Family
      You miss the point of raiding with friends and family. Bringing different people or more people changes the social dynamic, sometimes immensely so. Blizzard offers 10-player raiding. If they don't want people raiding with 10 players, they should up the minimum raid size--and perhaps they should, maybe the minimum should be 12 players, or even 15, it's an interesting discussion. It would certainly offset the issue brought up here.

      RE: Kargath vs. Other Bosses
      Agreed that currently Kargath is the most egregious example this tier, though I cannot agree or disagree that it's been or not been a big deal hindrance to groups without more data, which we don't currently have. However, that being said, I do agree that Kargath himself is trivial enough that it *probably* doesn't make a huge difference. The larger point being is that if designers aren't careful about how they make these encounters, requiring static numbers of players to handle special encounters impacts smaller raids more than larger raids.

      Delete
    6. "You're running under some faulty skill-level assumptions."

      I really don't think I am, though, for several reasons.

      First, I'm thinking about this more in some of the really terrible normal PUGs I've been in on my Horde priest.

      Second...

      "The data for raids in aggregate still shows at the 50th percentile of tank single-target DPS in Normal for successful kills of The Butcher (the closest thing to Patchwerk we have this tier), it's closer to 60%."

      That metric is skewed, though. Why? Because you're more likely to have the highest DPS recording the normal kills because they want to rank, and given that there's far more DPS than tanks in a group you're much more likely to have a high DPS in the group bothering to record the fight than a high damage tank in the group bothering to record the fight. So the fights that ARE recorded are skewed towards higher DPS than what is typical for normal groups.

      That's also not getting into the fact that it's more likely for an "elite" DPS to step into a friend's guild's normal run to help out than a tank. I know on a personal level that I've helped a decent chunk of people out in normal Highmaul on Balkoth as a DPS (and sometimes healer), but usually they have two tanks, haven't taken my warrior.

      "We still downed him eventually once we rebalanced who was healing what--the stands on Normal need almost no healing at all, versus the tanks and healing the raid back up sine the healers were behind from before getting tossed into the stands."

      This is something I don't understand -- when you say "healing the raid back up" you mean "healing half the raid back up?" I mean, you have 2 healers on the ground healing their group of 5 (with only the tank taking damage) and 1 healer in the stands healing their group of five (with only the tank taking damage). And both halves should be at about the same overall level of HP (from damage taken before the toss).

      I am sincerely asking what your healers were possibly healing at the time except possible previous damage which would also then apply to the stands group.

      Regarding inflection, what is your argument, then? This is all stemming from

      "Instead, it's just pushed the inflection point such that optimal raid size is greater than 13 or 14 players."

      when the optimal flex raid size, from the start, was 25 players. So I'm not seeing how it pushed the inflection point anywhere, really.

      "You miss the point of raiding with friends and family"

      I'm not sure I do, though. What would that same group of people do if they only had 8 or 9 people in their F&F group who wanted to raid? I mean, we're assuming that this F&F group has *exactly* 10 people able to raid. Not 9, because then they apparently cancel the raid. Not 11, because then it's much easier. *Exactly* 10.

      It seems like an insanely low probability scenario to not only have exactly 10 but then to not have anyone who has a good enough friend who's able to come who won't upset the social dynamic.

      On top of that, what happens when/if someone doesn't want to raid in that group of exactly 10 F&F? The whole group stops raiding for good? Or...do they try to make more friends, either in RL or in-game, to raid with them?

      And if it's the latter, then...why can't they be trying to do that process now rather than waiting for someone to quit?

      Delete
    7. "However, that being said, I do agree that Kargath himself is trivial enough that it *probably* doesn't make a huge difference."

      I'd go beyond that and say that if needing to play 5% better overall on Kargath is what is preventing a group from killing him that they have no chance against any other HM boss.

      "The larger point being is that if designers aren't careful about how they make these encounters, requiring static numbers of players to handle special encounters impacts smaller raids more than larger raids."

      But it all comes down to *how much* they impact the smaller raids, right? In SoO the bosses flat out scaled less with more people. If the boss has 100 HP, adding an 11th person might have given him 8 more HP while adding the 25th person might only have added 4 HP. The problem with that system is that adding the first few people past 10 didn't add all that much, though it did keep getting better. If anything you got increasing returns on adding people.

      Now, the boss's HP scales linearly but smaller groups are impacted more by this kind of stuff. You get more of an immediate benefit for going past 10 but it tapers off a bit more.

      Delete
    8. This is actually a discussion I was interested in continuing (though it seems you're not).

      Delete
    9. I'm unsure as to what else there is to continue.

      I have no data to support or debunk your skew hypothesis, and even if it were skewed, I still agree that the tanks do soften the blow (though it also means you're missing 3.75 DPS for 30% of the time instead, so it doesn't soften it as much as your positing). 60% vs 75% was just me quibbling.

      As to our deficiencies at the time, yeah, all right, we had a shitty first week and we pushed past him. As I mentioned, we were under geared and learning the fight. People were taking damage. Once again, running under faulty skill level assumptions. Normal mode is for scrubs Balkoth, and yeah, though we got Ahead of the Curve for Heroic ToT and SoO, a year off made us lazy and enforced bad habits, so we got bad. We're currently working that off and getting our skills as a raid back.

      Inflection points did actually exist at 17 raiders, by your own admission. Blizzard wanted to smooth them out. I argue they haven't if they're not careful. You disagree. We'll have to agree to disagree, because while I agree Kargath in practice is really easy, the data still shows, even addressing the issues brought up, that 10, 11ish people are stuck shouldering a disproportionate burden from a DPS perspective.

      As to F&F, sure, even 11 people is still more difficult, but if you're raiding for social reasons, you sometimes don't want to ruin that dynamic. From experience, I can tell you that raiding with 13 people as our guild is right now is a fair bit different than the 10 we did last expansion. Yeah, it's nice to have the wiggle room, and I like our new players, but it still feels a lot different. It's amazing how adding those 3 players changed the dynamic significantly. For us, it's not a bad thing, but I could see why others might not want that change. I don't have anything concrete for you here besides my own experiences, as this is about feelings rather than numbers.

      Delete
    10. Raiding isn't binary. Yes, there's either you kill it or you don't, but how well you do informs future fights as you suggest. A tiny raid struggling to get Kargath down doesn't necessarily mean they'll have no chance against future bosses, but they'll definitely need to pick up the pace more than raids who don't struggle. But keep in mind that each raider having to pick up 5% more RAID DPS is like 5k more each player. Even if we halve that to 2.5k, that's a huge differential than asking a bunch of DPS to up theirs by 500 DPS, which would be the case once you started hitting 13 - 14 players.

      But yes, the point is ultimately how much the static mechanics impact smaller raids vs. larger raids. I'm arguing perfectly linear scaling is a poor decision. I'm not arguing the it should be SoO either. It would be simple to offset this and still maintain a general "more people is more better" curve for DPS requirements at least, but it requires the designers to adjust health based on the mechanics of the raid, which they clearly are not currently.

      Delete
    11. ER, the Ahead of the curve comment should be new Heroic, old Normal. Gah, that still trips me up.

      Delete
    12. Apologies if I come off as brusque. It's been a frankly shitty week and I shouldn't be taking it out on other folks.

      Honestly, it's not as if I meant to ignore you, I just didn't have anything new or different to add to the conversation. Sorry.

      Delete
    13. "Apologies if I come off as brusque. It's been a frankly shitty week and I shouldn't be taking it out on other folks."

      No worries, hope it's been improving.

      "though it also means you're missing 3.75 DPS for 30% of the time instead, so it doesn't soften it as much as your positing"

      Well, let's say each DPS does 10 DPS and we'll ignore healers.

      70% of the time we have 65 DPS.
      30% of the time we have 27.5 DPS.

      Average DPS is 53.75.

      If we add another DPS...

      70% of the time we have 75 DPS.
      30% of the time we have 37.5 DPS.

      Average DPS is 63.75.

      So 18.6% more group DPS for adding 10% more people.

      "People were taking damage."

      I realize there's a skill gap but I am literally being 100% honest when I ask what people are taking damage from.

      There are only three sources of damage on the fight:

      1, Kargath
      2, Flame Pillars
      3, crowd

      1, Kargath is meleeing the tank
      2, presumably people aren't literally standing in the flame pillars as they spawn. Maybe they were.
      3, busy attacking the group that just got thrown up.

      That's why I'm confused -- unless people were literally just chilling in the flame pillars (and maybe they were) I'm trying to figure out what other damage would be hitting the raid.

      Ultimately it's not a huge deal, we're talking about whether heals can contribute during that 30% of time, but I am honestly confused, not trying to be insulting.

      "We'll have to agree to disagree"

      The only thing I've been disagreeing on is your statement of:

      "They've just moved it such that optimal size is now greater than 13 players."

      Which is just factually false. Both because 25 people was the optimal raid size in terms of tuning and 17 or 24 were really the only semi-significant break points prior to that. So it's not like 15+ people being much better than 10 people is new for WoD, which is what you seemed to be saying.

      "I don't have anything concrete for you here besides my own experiences, as this is about feelings rather than numbers."

      I understand that, and I'm not talking about performance concerns or anything. I'm talking about the harsh realities of raiding. If you're a group of 8 F&F...you need a minimum of two more people to raid. Doesn't matter if it changes the social dynamic, you either get 2+ more players or don't raid.

      Or if you have 10 F&F and then someone quits...you need to bring in a new person. Doesn't matter if it changes the social dynamic, have to do it or not raid.

      Do you see what I'm saying?

      "But keep in mind that each raider having to pick up 5% more RAID DPS is like 5k more each player."

      Er...I think there's a typo here? Even on Mythic Butcher most people only do about 32-33k so picking up 5% more raid DPS is about 1.6kish each.

      "It would be simple to offset this and still maintain a general "more people is more better" curve for DPS requirements at least, but it requires the designers to adjust health based on the mechanics of the raid, which they clearly are not currently."

      You mean tweak the scaling for every fight, other words, and alter it based on the mechanics? So that Butcher may gain 8% more HP per player while Kargath gains 5%, then 6%, then 7%, then 8%, then 8% on out or something?

      How would you "fix" this for Tectus given that it's not a health/damage issue?

      Delete
    14. RE: Your DPS thought-experiment
      Yup. So DPS goes up significantly by adding another, but health goes up linearly. That just continues to support my argument.

      RE: Kargath
      Kargath also hits the raid. His charge is an AoE that does about 40k to all affected parties on Heroic, and 30k on Normal. That makes up about 20% of damage done.

      There's also Iron Bombs getting tossed out, dealing damage and leaving DoTs. That makes up for about 40% of the damage done in our parses. That stacks up relatively quickly on melee especially. Iron Bombs still go out until the Bombers are picked up in the stands (and they spawn a good 30 - 45 seconds before the Chain Hurl).

      Between those two sources, which makes up about 60% of the total damage going out that you missed, if your raid isn't good at being spread out (and frankly there's not a lot of space if you don't want to be in a fire pillar when it pops up), it can be somewhat significant to catch up on. It's not a super tonne. Two good healers can easily two-heal it at low raid sizes if your raid is good at not clumping during the fight.

      RE: Raid Inflection
      I'm saying that with static numbers of player dealing with significant mechanics, a sharp inflection is created. It's not a wishy-washy "More DPS = More good" statement, which I agree has always generally been the case.

      RE: The Harsh Realities of Raiding
      Yup, if you only have 10 people, and one quits, you need to find someone else. And that sucks. It's still different than adding 5 players to your roster.

      Delete
    15. RE: 5% more raid DPS

      Kargath-10N has 31.35M health. Let's say you need to kill him in 7 minutes before your healers run out of mana. You need ~75k Raid DPS to pull that off. If you lose 17% of your total raid DPS to the stands (6.5 DPS players--5 and 2 tanks, 3.75 lost to the stands for 30% of the time), that's ~12.75k total. The two remaining DPS (and tank) will need to make that up, meaning about 4.75k each (and 3.25k for the tank).

      Kargath 11N has 34.69M health. 7 minute kill means ~82.5k DPS required. 15% of your DPS lost to the stands, ~12.4k to make up between 3.75 players remaining, so ~3.25k each (and 2.65k for the tank)

      Kargath 12N has 38.76M health. 7 minute kill means ~92.3k DPS. 13.2% DPS lost to stands, ~12.2k to make up between 4.75 players remaining, so ~2.5k each (and ~2.2k for the tank).

      Kargath 13N has 41.63M health. 7 minute kill means ~99.1k DPS. 11.8% DPS lost to stands, ~11.7k to make up between 5.75 players remaining, so ~2k DPS each (and ~1.7k for the tank).

      Kargath 14N has 44.46M health. 7 minute kill means ~105.9k DPS. 10.7% DPS lost to stands,
      ~11.3k to make up between 6.75 players remaining, so ~1.65k DPS each (and ~1.4k for the tank).

      The difference between 10N and 14N is nearly a factor of 3. And it just gets smoother from there. I don't mind needing to be better on a smaller raid, but that's egregious.


      RE: Tweaking the Scaling for Every Fight

      Yes. The sarcastic response is, "I call that tuning." **shrug** Blizzard has fights that scale. They need to be tuning for all raid sizes now. Not an easy job, but some analysis of the fight can certainly help determine good numbers. Make the fight fun, fix the numbers after.

      For Tectus, I agree, not quite as easy. What happens if all your healers get nailed? Or all your DPS? The nature of the mechanic suggests health tuning is probably not appropriate, or if it is, perhaps split the difference between outgoing damage and boss health. Maybe temper the RNG such that you can never have all your healers targeted (assuming enough targets available). It's a good question.

      Kargath is easy. It's just pure numbers. Other fights, not so much.

      Delete
    16. And also RE: 5% more raid DPS

      I do realize that the other 3.75 players can make up some of that deficit over the 70% of the time they're in the fight, so perhaps instead I should amortize them in, too. Let's see what happens if we get more accurate, maybe it's not so bad:

      10N -> 3.75 * 0.7 + 2.75 = 5.375; * 1 / 12.75k = 2.37k per player
      11N -> 3.75 * 0.7 + 3.75 = 6.375; * 1 / 12.4k = 1.95k per player
      12N -> 3.75 * 0.7 + 4.75 = 7.375; * 1 / 12.2k = 1.65k per player
      13N -> 3.75 * 0.7 + 5.75 = 8.375; * 1 / 11.7k = 1.4k per player
      14N -> 3.75 * 0.7 + 6.75 = 9.375; * 1 / 11.3k = 1.2k per player
      ...
      30N -> 3.75 * 0.7 + 19.75 = 22.375; * 1 / 11.4k = 0.511k per player

      Given that the 50th percentile DPS of folks killing Kargath N in early December was ~17k (and that's not accounting for all the inflation from stands DPS), we're talking a DPS differential for every player of 14% in 10N versus 7% in 14N, versus 2.3% in 30N.

      So basically, DPS players in 10N need to play 7% better than 14N, and 12% better than 30N. That jump from 10 to 14 is huge.

      Delete
    17. Also, this hotfix is a perfect example of what happens if Blizzard isn't careful with how they design raid content with static numbers of people:

      * Gruul now uses Petrifying Slam against 5 targets in a 10-player raid, scaling up to 10 targets in a 30-player raid (used to be 8 targets on all difficulties).

      Petrifying slam on 80% of your raid at 10 man? Lolwut.

      Delete
  3. I think this partially depends on the type of difficulty Blizzard wants to achieve. A fight might be relatively difficult for a 25 man group to pull off but might be easy for an individual. Using the example of people being chased by mechanics; if each individual has more time to stand around and do their thing the fight might be easier individually, but tighter DPS requirements might make the fight more difficult to the group as a whole. Raiding with fewer people might also mean fewer resources such as battle rezzers (including in people who can actually cast it), healing cooldowns, etc. Presumably Blizzard knows this, but I think Balkoth is correct in that they do want to subtly encourage larger groups by making bringing more friends overall a benefit. Meeting their stated goal of "Bring the player, not the class" requires you actually are encouraged to bring the player after all.

    ReplyDelete
    Replies
    1. "Meeting their stated goal of "Bring the player, not the class" requires you actually are encouraged to bring the player after all."

      Indeed.

      Delete
  4. Unrelated, but I prefer smaller raid sizes. My ideal MMO would be mostly dungeons, sometimes small raids, and rarely larger ones (for events usually).

    ReplyDelete
  5. One additional thought. Many guilds have "good" players who tend to handle the more difficult mechanics. This approach means that smaller guilds need a larger % of "good" players, whereas the larger raids have more space to "hide" people.
    All these effects reward guilds to invite "one more DPS", I suspect intentionally, to increase the inclusiveness of raiding.

    ReplyDelete
    Replies
    1. "All these effects reward guilds to invite "one more DPS", I suspect intentionally, to increase the inclusiveness of raiding."

      And yet sadly most PUGs, at least, are blind to this : /

      Delete
    2. I feel like raid size has always had an impact on the difficulty tuning of encounters. I know this is a hot topic amongst the rose tinted goggles wearing community, but 40 man raiding really did only require about 25 people to pull their own weight in order to be successful. Blizzard intentionally includes group coordination in their difficulty, and having more players generally increases the amount of coordination required for certain mechanics (Replicator Force Nova on Imperator for example). If there was no incentive to bring more players people generally don't, because space and mechanics often become easier to deal with in smaller groups. Another example as mentioned is Koragh's Expel fire, Twin's Pulverize etc etc.

      I feel like larger groups should be compensated due to the fact there are alot more factors in play than sheer dps requirements. I mean the traditional "gear check" is about the only type of encounter which dps is actually an issue, as most fights have quite lax enrage timers, hell Mar'gok doesn't have an enrage timer at all as far as i know.

      Delete
    3. You could 20 man MC in blues in Vanilla if those people had a blue.

      "If there was no incentive to bring more players people generally don't, because space and mechanics often become easier to deal with in smaller groups."

      Not to mention social issues like having to deal with more people (which also had more potential for drama).

      Delete
  6. All of highmaul that can be done with 10 is a joke anyways why is this even a thing? You really have troubles with kargath jus bcuz 1/2 your raid is in the stands? is that what is holding you back from a kill? srsly?

    ReplyDelete
    Replies
    1. Uh, no. In fact, Kargath has been on lockdown for us for ages. This was simply an exercise to explore the mathematics behind different raid sizes.

      Delete
    2. H Imperator isn't a joke regardless of whether it's 10 or 30.

      That's not to say *I* find him difficult (I lead a two night guild working on M Imperator) but he's certainly hard enough to not make him a "joke."

      Delete