Raids of Different Sizes Cannot be Tuned to Equal Individual Difficulties

[SobeBauxite]

lolmath.

Instead of math, try some history. Zul'Aman was a difficult raid instance. It was also 10-player. Hence, history shows us that 10 mans can be difficult. Stop trying to apply your wrath lol-10-mans-are-easy-mode impression onto a new expansion, and using some pretty terrible math to back it up. You know, the math that only takes one factor into account, and ignores class balance issues, lack of buffs, lack of more utility (lolstacking druid bresses, for instance) and quite a few other things people like to ignore when arguing 25 mans are always harder. I'm not saying 10s are easier than 25s, I'm saying they both have the capability of being as difficult as each other.

[diemor]

I think it's awesome that you tried to quantify 'epic fail' into an equation ;D

best.. quote.. ever.

[Fingo]

Let's see now. The common consensus is that either 10man content needs to be "tuned up" to 25man difficulty, or that 25man difficulty needs to be "tuned down" to 10man content.

Now, which one sounds more like Blizzard since Wotlk's release... Gee I wonder.

[Kheirn]

Let's see now. The common consensus is that either 10man content needs to be "tuned up" to 25man difficulty, or that 25man difficulty needs to be "tuned down" to 10man content.

Now, which one sounds more like Blizzard since Wotlk's release... Gee I wonder.

ICC 25 seems like a huge success.

Gee, I wonder what they will do.... : Really, do you think they would make 25-mans as trivial as 10-mans are now? The only challenge 10-mans offer are doing it Strict, using only 10-man gear. In that sense, you would come closer to how 25-man plays out in term of difficulty. It's kind of obvious that they want to maintain difficult raids while giving the players freedom of choice when it comes to raid size. I'm currently being "punished" for not wanting to do 25-man raids since I don't like the atmosphere in 25-man guilds. I'm all for this change and I eagerly await the challenges that will arise.

[The High Druid]

Surely equal difficulty here just means they are designed assuming the 10 man group and 25 group have the same levels of gear. Why does it need to be more complex than thet?

[diemor]

Okay.. I tried to follow but the variable's being too similar threw me off, but let me see if I got this.

You claim that 25 mans will always be harder than 10s, because if one person fails it wipes the raid. Now you could have summed this up in one sentence.

25 mans will be harder because you have 25 chances instead of 10 of getting an idiot in your group(24 and 9 if your confident in your own ability).

Problem is solved by not recruiting idiots.

if (raider == idiot) {
/ignore raider;
}
else {
/ginvite raider;
}
????
profit

[AuthurDent]

Fixt.

They will tune the encounters for 10 man.

25 man will see increased health/damage/attack frequency to compensate and take in account additional healing/b-rez ability or more peeps.

This is exactly the opposite of what is has been.. (Tuned for 25m then scaled back for 10 man)

10 man is the new focus for Blizzard as it represents the largest portion of raid content consumed. 25m efforts by Blizz will now be simple "bones" thrown to existing 25m guilds to more "gently" phase them out.

[ilialilov]

The place where your math goes wrong is the assumption that 1 man can fail a 25 man group. Widen that to 3 and the numbers will be a lot closer to each other.

[Lenea]

The encounters don't have to be exactly the same to be the same relative difficulty for 10 and 25 mans.

There are more people in 25 mans to co-ordinate, to assign roles and to deal with extra crap in the encounter which increases the difficulty. In 10 mans there are fewer people to take on specific roles, the impact of one person taking on an additional role can be more significant on dps/healing, and the loss of a single person can mean an inevitable or even immediate wipe - this increases the difficulty in a different way. An extra mechanic doesn't mean 25 mans are automatically harder, as you have 2.5 x the amount of people to play around with to deal with it. It isn't the SAME, but can be of a similar difficulty level (if tuned right).

10 mans don't seem challenging now as many people who run them also run 25 mans, and so technically 'overgear' the instance's tuning. This isn't going to be true in Cataclysm.

[sidphoenix]

Glad you're using some math here, but for better or worse the approximations of your assumptions are weakly linked to reality.

1: Each person's likelihood to fail is not independent of the other folks chance of failing, and in fact scales with it(why pugs<coordinated runs), and though this actually bolsters your cause that 25 would be significantly harder through logistics:
2: I know you want to treat the special case of individual failure=> raid failure, but since very very few encounters will be as tuned as LK, that case just doesnt come up in 25 man as much via Brez/ankh/distribution of buffs and responsibilities.

anyways, saying 25 is harder than 10 for extremely complex reasons does not imply that tuning them to be comparably attainable goals is impossible.

Example: You want the room to be a "little bit" darker and have a variable resistance knob for the light control. You can try to figure out the dimensions of the room and the illumination breakdown, quantify a "little bit", convert that to Lumens, and figure out for your light's efficiency and the dampening of whatever cover is on your light.

or you can be like blizzard and just turn the knob up or down until it looks right. Much easier, same effect. The trick is that through experiment it can be much much easier to find breaking points in the statistical treatment of success through PTR testing and live patching, than to try to theorize exactly how much the aura needs to tick for to require 1 more raid healer.

edit: spelling is hard

[AuthurDent]

This.

Bottom line is... best I can tell... if you had 25 skilled peeps... and 10 similarly skilled peeps enter in ICC with the same non-ICC gear... the only difference in difficulty is in getting 25 peeps online at the same time...

There is NOTHING more epic or difficult in 25 peeps doign their job than 10 peeps. Just that the problem lies in GETTING 25 peeps out of the scrubs that there are to fill the guilds out with.

Blizzard recognizes this.. and is shifting design priority to 10 mans... 25 man will just be scaled up to feel the same.

10 mans don't seem challenging now as many people who run them also run 25 mans, and so technically 'overgear' the instance's tuning. This isn't going to be true in Cataclysm.

[Pendergrass]

There are a couple of points:

The conclusion is that for a large class of mechanics (when failure changes are multiplicative instead of additive), adding more people inherently wrecks the ability to equally tune the encounter. Encounters will generally have a mix of factors, some of which are additive (more DPS, say, on Patchwerk) and some of which are multiplicative (not pulling an Orb on Princes, many tanking mechanics). My statement is that to the degree that multiplicative failures are present in particular fight, that fight cannot be tuned equally.

A key question is what does "tuned to the same difficulty mean"? Equal chances for success for the entire raid? Equally difficult for individuals?

97% compared with 95% doesn't seem like much. 3% compared with 5% could be huge. That might be the difference between not standing in the fire (on Northrend Beasts) and completely dodging it (which is possible, if you watch very carefully and have very good reflexes).

[cien576]

Raids of Different Sizes Cannot be Tuned to Equal Individual Difficulties

This note is in regards to the tuning of raids of different sizes to equal difficulties.

The conclusion of this note is that for raids of different sizes, except in a narrow category of raid mechanics, tuning of the raids to the same difficulty is not possible.

The key statistic which is explored by this note is the change of failure (PF_s) of a raid member in cases where the failure of a single raid member during a boss encounter will very probably lead to a failure of the entire raid during that boss encounter.

Examples of boss encounters where such is the case are numerous. For example, in Blood Queen, Blood Princes, and Thaddius.

There are other examples where a single failure does not necessarily lead to a complete failure, for example, Patchwerk.

I think you making an unfair projection of the current difficulty of 10 man raids v. 25 man raids onto cataclysm raiding. Inherent differences exist in the two types of raids which currently amount to 10 mans being considerably easier [because responding to certain mechanics like spreading out is simply easier in a 10 man environment], however, this does not mean that it would not be possible to create a 10 man version of the encounter that might have slightly different mechanics than the 25 man version in order to balance the difficulty level between the two.

Although I would most likely be in favor of Blizzard making both 10 only and 25 only raids (like in BC) because I think designing encounters with only one style of raiding is easier to execute and allows for more variety in available content, I will give them the benefit of the doubt and see how cataclysm is developed.

I do like that they are eliminating the need to kill a boss more than once in a week by placing all raids on the same lockout but I hope they have more content available then to compensate for content (even if it was repetitive) lost in eliminating the ability to do the highest tier raid twice in the same lockout period (once on 10 and once on 25).

Definitions:

PS_s = The chance of success of a single raid member
PF_s = 1.00 - PS_s = The chance of failure of a single raid member

PS_r = The chance of success of the entire raid.
PF_r = 1.00 - PS_r = The chance of failure of the entire raid
n = The number of members of the raid

Raid Failure Formula:

(1) PS_r = (PS_s)^n
(1') PF_r = 1.00 - (1.00 - PF_s)^n

I do not think these equations work at all especially because they oversimplify raiding. You also make another assumption which you did not list below in writing these equations. It is false as well as the other two below. It is:

"All raid members have the same chance of failure or success on a boss"

Obviously, all encounters are not designed equally for every player and some players will have a higher chance of failure than others

Your equations should probably be modified to be the average of the summation of every individuals chance to fail or succeed in the raid.

But determining an individuals chance of success or failure is hard to do and it changes as they become more accustomed to a fight. Even given your definitions it is possible that PS_r where n=10 can equal PS_r where n=25 can become equivalent, in which case doesn't your argument fall apart? Just as difficult meaning same chance of raid success...individual players may have a higher or lower level difficulty associated with an encounter in 10 as compared to 25

Assumptions:

1) The chance of a single raid member succeeding or failing is independent
of the chance of any other raid member succeeding or failing.

2) All raid members must success for the success of the entire raid.
(The failure of a single raid member will fail the entire raid.)

These are two bad assumptions, the first one is not entirely true because one person's mistake can easily lead to the mistake of another. The second one is also not true because it is very possible to kill a boss after raid members have died or 'failed' especially with things like battle rez, shaman ankh, and soul stone in the game.

Inverting the raid failure formulas (1) and (1') gives:

(2) PS_s = PS_r^(1/n)
(2') PF_s = 1.00 - (1.00 - PF_r)^(1/n)

Working from a target overall chance of failure of 0.50 (50%), this results to PS_s(n = 10) = 0.933 and PS_s(n = 20) = 0.966, and PF_s(n = 10) = 0.067 and PF_s(n = 20) = 0.034. Note that PS_s(10) = PS_s(20)^2 because of the particular selections for n.

Working backwards will obviously reveal disparity in what is required from each individual in the raid given your assumptions, however, as I said above your assumptions are all invalid at least to some degree.

That is, depending on the the particular mechanics of the boss encounter, tuning the fight to provide the same level of difficulty for individual raid members leads to a different chance of success for the entire boss encounter. Conversely, again for particular mechanics, having the same overall chance of success for the encounter
requires that the encounter be set at a different level of difficulty for each raid member.

Notes:

The selection of raid sizes (n = 10) and (n = 20) does not limit the result: The key relationship is the difference in raid sizes.

The selection of raid mechanics to single point failures is limiting, but is a valid approximation of many current raid mechanics.

Not true at all, an encounter does not have to be set to the same level of difficulty for each raid member in order for an encounter to be just as difficult on 10 man as it is on 25 man especially because there is a disparity in level of difficulty depending on your role in the encounters of today. Your oversimplification of raiding is really the issue here and what has caused you to draw that conclusion. The nature of a 10 man raid is entirely different from a 25 man raid and Blizzard will probably opt to design 10 and 25 man encounters with that in mind.

It is unrealistic to say that blizzard can create encounters which are equally difficult for everyone involved in them and then to say that they would have to make a 10 man version and 25 man version while maintaining that equality in difficulty for everyone involved. This philosophy could only result in patchwerk style fights and the game would be boring. 10 man raids obviously will require more from each individual than a 25 man because of the style of the raid but that is not to say that the raid cannot be designed in a way that is just as difficult if not more difficult than the 25 man version of the encounter.

[Fingo]

The place where your math goes wrong is the assumption that 1 man can fail a 25 man group. Widen that to 3 and the numbers will be a lot closer to each other.

Please go work on either 10 or 25man versions of heroic Lich King, and tell me that one person cannot wipe the group.
Myself, I've completed 10man HM LK, but my guild is working on the 25man version still.

http://www.wowarmory.com/character-s...er&cn=Jacelynn

My main, if you cared.

[sicness]

lolmath.

Instead of math, try some history. Zul'Aman was a difficult raid instance. It was also 10-player. Hence, history shows us that 10 mans can be difficult. Stop trying to apply your wrath lol-10-mans-are-easy-mode impression onto a new expansion, and using some pretty terrible math to back it up. You know, the math that only takes one factor into account, and ignores class balance issues, lack of buffs, lack of more utility (lolstacking druid bresses, for instance) and quite a few other things people like to ignore when arguing 25 mans are always harder. I'm not saying 10s are easier than 25s, I'm saying they both have the capability of being as difficult as each other.

Except you're comparing a sole 10 man raid to raids designed for 10 and 25 man...

There's no doubt the some encounters are easier and harder on either mode, but there is no doubt that it's harder to get 25 people organized instead of just 10 and that is significant.

[Herecius]

The OP is basically just giving mathematical proofs for what most already know; 25man encounters are often made harder than 10 not because of tighter enrages or harder-hitting bosses, but because you have 15 more people to potentially blow up the raid with any ability that is single-targeted to a random member and is vital in how its handled.

Blizzard, if they want to live up to their rather far-fetched promise, is either going to have to make 25s laughably forgiveable, thus making them 'even' with 10s, or make 10s EXTREMELY tightly tuned, which would land us back in the weird Naxx/Ulduar cases where 10 was actually considered tougher than 25 by most.

[sicness]

The place where your math goes wrong is the assumption that 1 man can fail a 25 man group. Widen that to 3 and the numbers will be a lot closer to each other.

Except that it's very common, especially on hard mode encounters, that one person can wipe the raid group. Just as is common that 1 person failing in 10 man won't wipe the group. So it goes both ways and saying that more responsibility is put on players in 10 man isn't accurate.

[DSRilk]

Your options are make the room smaller or increase the space needed to spread out

Or they can change the area of the effect, which is extraordinarily easy.

Assumptions:

1) The chance of a single raid member succeeding or failing is independent
of the chance of any other raid member succeeding or failing.

2) All raid members must success for the success of the entire raid.
(The failure of a single raid member will fail the entire raid.)

These assumptions are incorrect, and you're missing one huge point: Going forward, raids will be designed with the goal of making 10 and 25 man equivalent.

Here are the assumptions you can make:

1. Blizz is assuming that because 10 man gives equal loot, you won't be running 25 man raids unless you like 25 man raids. This means that they can tune for a team of 10 reasonably good people and 25 reasonably good people, as opposed to now, where it's 10 reasonably good people or 10 reasonably good people and 15 droolers.

2. Importance of individual failure will scale based on 10 and 25 people. E.g. if you can succeed while losing 2 DPS in a 10 man, you can succeed while losing 5 DPS in a 25 man.

3. Blizz can easily change the AoE and number of targets of spell effects, the magnitude of spell effects, and what spell effects are used in each version among other things. And doing so, they will make it equitable for each version.

How can you make these assumptions? Because going forward, they are going to design raids based on the idea that they will be equally difficult. So yes, it is arguably impossible to make 10 man and 25 man equivalent if one person can wipe the entire raid, which is why you won't see that on both versions. I believe the major error in your hypothesis is that you seem to be trying to figure out how they would take existing raids and make it equitable for 10 and 25 man teams by only tweaking the damage and health. And yes, that would be extraordinarily hard. Given that's not the task they are undertaking, however, your conclusions are unfortunately erroneous.

[Pendergrass]

These assumptions are incorrect, and you're missing one huge point: Going forward, raids will be designed with the goal of making 10 and 25 man equivalent.

That must be demonstrated. (And is not even a conclusion: The goal is to reach equivalence. That is not to say that the goal will be reached, only that the designers will be trying.)

There are a couple of ways of reaching equal overall failure rates:

Reduce the impact of a single player failure: On PatchWerk, as a DPS, except for low threat tanks, DPS failure is additive. Equivalently, make the raid mechanics very easy for the additional people. (This must be true for the extra tanks and healers, or you run into the initial problem.)

Double (or slightly more than double) the encounter length. The chances of failure then works out to be the same.

The question then becomes, is it the same encounter, or even the same difficulty?

Even in non-hard modes, there have been quite a few encounters with single player failures: Melee too close on KT; Pulling aggro on Sindragosa; Failing to move on Thaddius; Failing a bite on Blood Queen; Tripping a canister (or failing to avoid the orange ooze) on Putricide; Pulling the orbs on Blood Princes. Failing a taunt rotation (lots of bosses).

[vesseblah]

The conclusion is that for a large class of mechanics (when failure changes are multiplicative instead of additive), adding more people inherently wrecks the ability to equally tune the encounter. Encounters will generally have a mix of factors, some of which are additive (more DPS, say, on Patchwerk) and some of which are multiplicative (not pulling an Orb on Princes, many tanking mechanics). My statement is that to the degree that multiplicative failures are present in particular fight, that fight cannot be tuned equally.

A key question is what does "tuned to the same difficulty mean"? Equal chances for success for the entire raid? Equally difficult for individuals?

The point where your assumption of perceived vs real difficulty and the math goes seriously wrong is that you assume identical numbers for both raid sizes, of course it's not going to work.

How they resolve the issue is pretty simply tuning the 10-man abilities slightly harder to avoid to simulate the crowds in 25-man format. In reality if you look at the raw numbers and take 25 players of exactly equal skill into 10 or 25 man raid the 10 man raid would be slightly harder, but the percieved and effective difficulty would remain virtually identical due to the increase chance of failing in 10-man. This makes tuning abilities like Defile on LK fight totally reasonable to reach the point where the difficulty is virtually identical regardless of raid size.

Same principle applies to just about every imaginable thing you'll bump into while raiding. Adjusting either raw dps or timers, you can get the effective difficulty to a point where normal raid will not percieve any difference and wipe equally many times even if the actual numbers are vastly different.