WotLK to MoP: A Trend Analysis of Raiding Difficulty.

**Pua** · 2013-05-30, 02:44 PM

[Author's Note]

This post has been dramatically reconstructed, and has been done so bearing in mind some of the statistical criticism and some particularly interesting data presented by other posters. It is inherently understood, both by the OP and the majority of the community supporting its conclusions, that the data isn’t perfect. That said, I’ve tried to keep comparisons as relevant as possible and explained why they’ve been made the way they have.

Remember: This is a discussion post.

Keep in mind its intended use while posting as constructively as possible – everyone will appreciate that a lot more. It’s also worth bearing in mind that Ghostcrawler himself saw the original version of this post and I’ve attached his commentary to the bottom.

Stay safe out there, and enjoy this work.

------------------------------------

As can be seen around the forums recently, much has been made of the “gap” between LFR difficulty raiding, and the normal setting on either 10-man or 25-man. The premise is formed on one of two opinions:

1) Normal mode is too hard.
2) LFR and heroic five-man dungeons are too easy.

Essentially, there’s validity in both camps and it’s a shame that a lot of shouting has gone on when a more meaningful debate could have taken place. If you want to track this discussion, it can be found at the following address; there’s a lot of good commentary in here, as well as the usual angry intercessions that accompany the debate. Part of those arguments have revolved around whether or not content now is harder than it was when the game was at its peak, subscriber wise. There is a growing opinion that raiding as a part of the game is fundamental, but the numbers are dramatically dropping and it’s continually increasing complexity is often cited as the major cause.

While using figures from WoW Progress, I’ve had a look at the fall-off rates to see where guilds appear to be giving up on the progression curve, which bosses seem to be causing it, and if the game is legitimately harder than it was during WotLK. Clearly, there are going to be a lot of subjective calls in here, and this is not meant to be an in depth analytical piece, but it’s a starting point for the community to get to grips with raiding trends and exactly which direction they’re going in.

To start with, a primer is in order:

- Do not expect unbiased views; it’s not that type of debate. People are rightly passionate about how they see the game, so I ask the moderators to show a little leniency.
- Some graphical representations are presented, but kept simple. This is to provide a visual idea of what we’ve been seeing since the game’s high point.
- 10-man normal is the setting used for most of this. It’s probably the most relevant, given that’s where the developers (and some of the community) believe the gap to be.
- Everything here is up for debate, but I’ll simply report posts that start the usual “no-lifer” to “scrub” intercession. I’d like this to be kept clean.

With that out of the way, let’s start with some numbers.

Wrath of the Lich King (Tiers 8, 9 and 10)

Clearly, we have to take a lot of these particular numbers with a pinch of salt. The progressive nerfs to instances, the length of time they were “current” for and the length of time that’s passed since they were current are all outliers that gut this particular set of numbers with regard to relevance; not to mention the split lockouts, different raid design and different community/guild structures. They do, however, fulfil a purpose when looking at who moved into Cataclysm and started raiding, so they’re presented here to be systematic. The first set of numbers comes from DisposableHero, and he makes the point far better than I did originally:

He started with a data set of boss kills, taking 2 months as a standard interval and looked at raid boss kills back to tier 9 after the boss being available for 2 months. There is some error in how he gathered this as he looked at the world rank of the last US guild to kill the boss 2 calendar months after the boss was released to determine the number of kills (and cut down on him manually scrolling through pages and pages of kills). This has a small margin of error that is not believed to be statistically significant. For tier 9, he took boss kills from 2 months after the launch of the instance even though the bosses were released 1 per week, as that didn't seem to slow anyone down anymore than just having to kill the earlier bosses in an instance. For Icecrown, with its staggered release, he recorded the boss kills 2 months after their wing opened.

This was done for 10-man kills and 25-man kills, and then multiplied out to get a number of characters with boss kills. From this, we can glean some interesting information; mainly the split between 10-man and 25-man, calculated from the percentage kills from raiders playing 25-man versus 10-man. The results shouldn’t surprise anyone, but here are some numbers for the first bosses of each tier (for tiers with multiple start bosses, the one with the highest number of kills was chosen):

Tier 9 - Beasts of Northrend - 68,838 kills, 48,883 10-man, 19,955 25-man.
By kills, 71% 10-man, 29% 25-man. By players, 49.5% 10-man, 50.5% 25-man.

Tier 10 - Lord Marrowgar - 79,523 kills, 54,968 10-man, 24,555 25-man.
By kills, 69% 10-man, 31% 25-man. By players, 47.25% 10-man, 52.75% 25-man.

Clearly, this shows that 25-man guild activity in the heady days of WotLK (prior to the shared lockout) was the preferred option; the most interesting part is that at least 50% of people raiding were raiding 25-man content prior to Cataclysm, and the rest of this post will highlight just how far this percentage has fallen.

By the end of the expansion, the Icecrown numbers looked like this for 10-man:

Lord Marrowgar: 84k
Lady Deathwhisper: 84k
The Gunship Battle: 83k
Deathbringer Saurfang: 83k
Festergut: 81k
Rotface: 80k
Professor Putricide: 75k
Valithria Dreamwalker: 75k
The Blood Prince Council: 72k
Blood Queen Lana'thel: 70k
Sindragosa: 62k
Arthas: 48k

These numbers point to the fact that those starting the instance by defeating Lord Marrowgar then went on to kill Deathwhisper, the Gunship and Saurfang with no more than a 1% drop in kills. The first drop of any note is, unsurprisingly, found at Festergut who saw 3% of Lord Marrowgar’s total fail to kill him. Another 1% drop goes to Roftace, and then a 6% drop to Professor Putricide. At this point I jumble the order of the bosses slightly given that the progression here wasn’t linear, but the same number of people killed Dreamwalker before a 3% drop to the Blood Princes and another 3% drop to Lana’thel. Therefore, out of the first 10 bosses in the instance, there was never any more than a 17% drop in kills (again, given the state of the numbers, read into this what you will).

Then we come to Sindragosa, the first real drop from one boss to another, as we shed 22k kills and a 9% drop. Lastly, the final drop to Arthas, rounds us out at 57% of those who killed Lord Marrowgar; a 17% drop from Sindragosa.

1) Killing Lord Marrowgar meant you were killing up to and including Deathbringer Saurfang, and most likely Rotface and Festergut.
2) Up to Sindragosa, the overall drop in kills was 17% over 10 bosses.
3) The biggest drop was Sindragosa to Arthas, a drop that consisted of the same 17% that was shed from Lord Marrowgar to Blood Queen Lana’thel.
4) A total of 57% of guilds that started the tier managed to clear it.

Before moving on to tier 13, just take a look at the number of players killing Lord Marrowgar when the game was at its peak. Raiders, of any stripe (and even when the game was reputedly easier at level cap), have never breached 100,000 guilds.

That’s a sobering thought.

But with all this borne in mind, please understand that in Icecrown we’re talking about a tier that was extremely long, was progressively nerfed comparatively early, and that 25-man guilds had players that were doing all sorts of wonderful things.

Basically, do not read too much into these numbers. The only thing that I would take from Icecrown is that the system of progressive nerfs kept people going, certainly until Sindragosa and (particularly) Arthas. That’s probably the only widely agreeable conclusion, though an argument can be made that Arthas himself, when compared with the rest of this post, was the most challenging end boss the game has seen on normal mode.

Cataclysm (Tiers 11, 12 and 13)

To start what we’re seeing in Cataclysm, I’ll again defer to DisposableHero and his numbers for first boss kills in each tier:

Tier 11 - Magmaw - 46,027 kills, 41,843 10-man, 4184 25-man.
By kills, 90.9% 10-man, 9.1% 25-man. By players, 80.0% 10-man, 20.0% 25-man.

Tier 12 - Shannox - 31,528 kills, 27,500 10-man, 4028 25-man.
By kills, 87.2% 10-man, 12.8% 25-man. By players, 73.2% 10-man, 26.8% 25-man.

Tier 13 - Morchok - 42,861 kills, 38,350 10-man, 4511 25-man.
By kills, 89.5% 10-man, 10.5% 25-man. By players, 77.3% 10-man, 22.7% 25-man.

The first thing to note is that the number of people raiding at all crashed. More players had killed Arthas on 10-man than managed to defeat Magmaw in total, and this is a staggering statistic considering tier 11 saw the game’s highest subscriber number to date (or since). The drop in numbers from Lord Marrowgar to Magmaw is a nigh unbelievable 42% using the two-month interval period. Given the high number of players, the lack of any time-consuming grinds prior to raiding and the fact there was nowhere else to go in the PvE endgame (LFR hadn’t been introduced), we can unquestionably conclude that raiding in Cataclysm became significantly harder than it was in WotLK, and this took its toll on the participation rate. The numbers drop moving on to Shannox to the tune of around 14,500 kills (roughly 32%), so these were the guilds that we could argue simply gave up along the way. There is simply no trend similar to this prior to Cataclysm. Coming back into tier 13 and Morchok, we get quite a significant jump back up the way and this is likely attributable to the gearing hump of Firelands nerfs while it was current as well as the Hour of Twilight five-man dungeons, an easier boss overall, and the advent of LFR.

By the end of the tier, final Dragon Soul numbers looked like this for 10-man:

Morchok: 55k
Yor'sahj the Unsleeping: 54k
Warlord Zon'ozz: 54k
Hagara the Stormbinder: 54k
Ultraxion: 53k
Warmaster Blackhorn: 52k
Spine of Deathwing: 52k
Madness of Deathwing: 49k

Similar to Icecrown, if you were killing Morchok at the start of the instance it meant there was more to come. What’s staggering is that it pretty much meant everything was to come. Again, the same bias with numbers needs to be borne in mind as with Icecrown, but there is no more than an incredible 9% drop between those killing Morchok who then went on to kill the Madness of Deathwing. The biggest drop is 5% and happens between Spine and Madness which, as with Arthas, would imply that the hardest boss of the instance was the last one. Another parallel with Icecrown, naturally, was the progressive nerfs that did their job very effectively once again.

1) Only 6,000 guilds that managed to kill Morchok ended up unable to kill Deathwing. 6,000.
2) From a percentage point of view, there was a tiny 9% attrition throughout tier 13.
3) The biggest drop was between Spine and Madness, with no other jump managing to shed more than 1% of guilds.
4) Despite the dramatically lower numbers starting Dragon Soul (approximately 35% less), Deathwing was still killed more times than Arthas, who also had a longer shelf-life.

I mentioned the 35% drop in raiders from Icecrown to Dragon Soul, but this was a peak for Cataclysm as players returned for more forgiving content, the advent of LFR and, quite likely, more five-man dungeons.

Taking Cataclysm as a whole, it's hard to decide which boss to start on so we'll average out Magmaw, ODC, Halfus and the CoW: Total 68k.

This looks like a remarkable drop from WotLK (it is, for reasons posted above), but we have to take the shared lockout into consideration before saying "OMG". Again, we need to average the bosses in Firelands for their total, given the fact you could start almost anywhere: Total 62k.

Not bad, really, not too bad at all, this is just shy of a 9% estimated attrition rate from one tier to the next. This comparison between tiers 11 and 12 is pretty good, given that the shared lockout affects both raids, subscriber numbers were relatively stable (the bigger drops were to come) and LFR isn't there to cushion either tier. Moving on to Dragon Soul and Morchok time: Total 62k.

This number is fundamentally important because it proves that despite the bigger subscriber loss going into the Dragon Soul patch, as well as the introduction of LFR, the number of normal mode kills does not drop. Yes, it’s understood that we’re talking about a very long patch with progressive nerfs, but the nerfs only prove that people will continue to raid if the difficulty gets the effort to reward ratio correct. LFR is NOT the cause of normal raiding numbers being in the dire straits they currently are, because we didn’t see that happening when it was introduced.

I think we need to be very careful about what we say regarding LFR, and for many reasons. If we say LFR is there to let people see content, then it does its job. But is it providing meaningful progression?

Well, no. No, it's not.

Progression can be about:

1) Spending time with friends.
2) Helping less gifted players progress their characters.
3) "Organised" group activity is fun.
4) Serving as a stepping-stone to bigger things.

Unfortunately, none of this is covered by LFR or punishing normal mode raiding. Blizzard brought everyone into this during WotLK, then decided to design their entire endgame model around an extreme (and extremely vocal) minority of players who all wailed that "challenges are more fun".

Clearly, these numbers are showing the exact opposite and Blizzard need to be having another look over where their priorities lie with World of Warcraft.

Still, here comes the pain.

Mists of Pandaria (Tiers 14 and 15)

What’s extraordinary about this graph is the fact that Mists of Pandaria saw a bounce in subscribers, yet less players managed to kill the very first boss of this tier than did the final boss of the previous expansion. Yes, you read that correctly; less. Once more, let’s look over the numbers provided by the peerless DisposableHero:

Tier 14 - Stone Guard - 32,685 kills, 30,362 10 man, 2323 25 man.
By kills, 92.9% 10 man, 7.1% 25 man. By players, 83.9% 10 man, 16.1% 25 man.

Tier 15 - Jin'Rokh - 26,775 kills, 24,775 10 man, 2000 25 man.
By kills, 92.5% 10 man, 7.5% 25 man. By players, 83.2% 10 man, 16.8% 25 man.

Overall, that leads to some not too surprising results. The switch to equal item level loot in 10-man and shared lockout shot 25-man guilds in the heart, as the number of 25-man kills of the first boss dropped by a factor of 6 between tier 10 and tier 11, and they went from an even split of the raiding community to a 20% minority. It has also slipped further this expansion, being close to 15% of the raiding community now.

Interestingly, but also not surprisingly, the percentage of 25-man kills increases as you get later into each tier, suggesting that 25-man guilds continue to lean to the hardcore end of the spectrum. This difference is most pronounced in tier 15, as only 16.8% of players killing Jin'Rokh are 25-man raiders, but almost 28% of players killing Lei-Shen are 25 man raiders.

Here are the facts and figures for 10-man kills in tier 14:

Stone Guard: 47k
Feng the Accursed: 46k
Gara'jal the Spiritbinder: 44k
The Spirit Kings: 44k
Elegon: 38k
Will of the Emperor: 34k

Imperial Vizier Zor'lok: 39k
Blade Lord Ta'yak: 35k
Garalon: 30k
Wind Lord Mal'jarak: 27k
Amber-Shaper Un'sok: 23k
Grand Empress Shek'zeer: 20k

Protectors of the Endless: 28k
Tsulong: 26k
Lei Shi: 25k
Sha of Fear: 22k

The first wall that appears here, Elegon, sees a drop of 13% from the Spirit Kings (20% from the Stone Guard). Just to put that into perspective, those four bosses managed to shed more guilds than everything in Icecrown before Sindragosa combined, and the entirety of Dragon Soul. This is an incredible statistic. Not only did the tier start with probably the roughest opening encounter in WoW history, it managed to hammer out the biggest percentage to date before even a quarter of the tier had been completed. There are 2% more kills on Imperial Vizier Zor’lok than there are on Elegon, implying many players switched to kill him before Elegon, but certainly not Blade Lord Ta’yak who sits at 75% killed from those who defeated the Stone Guard.

But just take a look at those Heart of Fear numbers.

Ta’yak defeats 7% of those beating Zor’lok, while Garalon kills another 11% after him. Another 6% are beaten by Mel’jarak, 9% by Un’sok and 7% are undone by Grand Empress Shek’zeer. A miserable 42% of guilds managed to officially unlock the Terrace of Endless Spring after defeating the Stone Guard. Horrendous. In fact, amusingly, guilds seem to start making inroads into the Terrace via other means rather than sticking with Amber-Shaper Un’sok (more guilds killed Protectors of the Endless, Tsulong and Lei Shi than defeated Un’sok and Shek’zeer).

1) This is the first time we start to see massive jumps and “brick walls” appear in normal mode raiding. Elegon himself puts paid to more guilds than the entirety of tier 13.
2) The Heart of Fear is a one-instance wrecking crew. Of the guilds that started the expansion by managing to defeat the Stone Guard, it’s managed to kill over 58% of them.
3) The ‘attunement’ for Heart of Fear is bypassed, allowing more guilds to kill the Sha of Fear than killed Grand Empress Shek’zeer.
4) 75% of the final tier instance was less punishing than Amber-Shaper Un’sok; the Heart of Fear accounts for an average mortality rate of over 7.6%.

Those claiming that the Stone Guard was an unusually punishing boss for a first tier seem correct, but the sheer brutality found in Heart of Fear accounts for 58% of guilds dropping out as a result of it. There’s a jump of 2% to the Sha of Fear which clearly implies those managing to kill Shek’zeer (a pitifully, pitifully small number) finish by clearing the tier in its entirety.

I need to catch my breath after that. The wreckage caused to guilds by tier 14, particularly in the Heart of Fear, is really quite alarming. However, let’s move straight on to the Throne of Thunder:

Straight away we can see that almost half (22k or 47%) of guilds starting the expansion have fallen by the wayside coming in to 10-man Jin’rokh.

Jin'rokh: 25k
Horridon: 20k
Council of Elders: 17k
Tortos: 15k
Megaera: 14k
Ji-Kun: 13k
Durumu: 11k
Primordius: 10k
Dark Animus: 10k
Iron Qon: 9k
Twin Consorts: 9k
Lei Shen: 7k

Alas, things are not going to get any better. Horridon, well discussed as the first wall of this tier, manages to dust himself off to the tune of another 20% of guilds killing the monster before him. Again, going back in time, Throne of Thunder has managed to kill more guilds in a single boss than Icecrown managed prior to Sindragosa, or the entirety of Dragon Soul. One boss. Alas, the Council of Elders and Tortos between them work the miracle of trash-canning another 17% before Megaera kills off another 6%. Should you manage to kill the hydra, you’re looking at another 15% of guilds being killed off before Iron Qon who, should you kill him, guarantees you a crack at Lei Shen.

Lei Shen himself, by the way, who just over a quarter of guilds managing to kill Jin’rokh have managed to kill to this point in the tier. This is not good news.

1) Essentially, the Throne of Thunder is limited solely to those who survived the decimation of tier 14 and cleared it.
2) Horridon accounts for a full 20% of guilds that kill Jin’rokh, becoming the single-biggest stopping point of this entire paper, other than Arthas.
3) Council of Elders and Tortos, between them, do to guilds what the vast majority of content can’t do in a full tier.
4) Lei Shen stands tall as the hardest to kill end boss, by percentage, of any boss looked at in a single tier, yet the drop between him and the preceding boss is the smallest. This implies that Lei Shen is actually quite easy, it’s just getting to him that’s hard.

And just so we’re clear – of the 47k guilds that started this entire expansion by killing the Stone Guard, only 7k (a terrifying 14%) have managed to follow that up with a kill on Lei Shen.

To round out all of this, and to provide a TL, DR if you didn’t want to stick around for the whole post, DisposableHero presents the following piece of work which, arguably, is the best part of this entire thread (his conclusions are at the bottom):

“The really interesting piece of data though was when I took the total number of players with kills, and normalized against the reported subscriber count for the quarter the data was taken. This gives us a picture of what percentage of the player base participates in normal mode raiding. Here are the results:

Tier 9
Beasts - 8.59%
Jaraxus - 8.39%
Faction Champs - 8.04%
Twin Valks - 7.86%
Anub'arak - 7.59%

Tier 10
Marrowgar - 10.12%
Deathwhisper - 9.90%
Gunship - 9.89%
Deathbringer - 9.62%
Festergut - 5.92%
Rotface - 5.15%
Putricide - 3.57%
Blood Prince Council - 4.38%
Blood Queen - 3.02%
Valithria Dreamwalker - 4.51%
Sindragosa - 2.96%
The Lich King - 1.24%

Tier 11 (ready for the shocker)
Magmaw – 4.36%
Omnitron – 3.80%
Maloriak – 3.05%
Atramedes – 2.42%
Chimaeron – 2.03%
Nefarian – 0.69%
Halfus – 3.70%
Valonia and Theralion – 2.97%
Ascendant Council – 1.55%
Cho’gall – 1.43%
Conclave of Winds – 2.71%
Al’Akir – 0.94%

Tier 12
Shannox – 3.38%
Beth’tilac – 3.02%
Rhyolith – 2.78%
Alysrazor – 2.16%
Baelroc – 2.59%
Majordomo – 1.95%
Ragnaros – 1.15%

Tier 13
Morchok – 4.87%
Zon’ozz – 4.52%
Yor’sahj – 4.68%
Hargara – 4.43%
Blackhorn – 4.00%
Spine – 3.52%
Madness – 3.31%

Tier 14
Stone Guard – 3.58%
Feng – 3.31%
Gara’jal – 3.04%
Four Kings – 2.90%
Elegon – 1.84%
Will of the Emperor – 1.54%
Imperial Vizier – 2.66%
Blade Lord – 2.07%
Garalon – 1.55%
Wind Lord – 1.28%
Ambershaper – 1.02%
Grand Empress – 0.83%
Protectors of the Endless – 1.20%
Tsulong – 1.07%
Lei Shi – 1.05%
Sha of Fear – 0.84%

Tier 15
Jin’rokh – 3.59%
Horridon – 2.67%
Council of Elders – 2.35%
Tortos – 2.02%
Megaera – 1.79%
Ji-kun – 1.65%
Durumu – 1.40%
Primordius – 1.36%
Dark Animus – 1.24%
Iron Qon – 1.11%
Twin Consorts – 1.09%
Lei Shen – 0.80%

So, based on this data I propose the following hypothesis. A delta in the difficulty and tuning of the bosses in normal mode raids has a significantly larger effect on participation levels as a percentage of total subscribers than LFR has had. As evidence, I present the data above and make the following observations. In WotLK, where we know the development intent was for the raids to be easier and more accessible than it has been in the previous two expansions, the participation level for raiding was around 10% of the total subscriber base. When more difficult raids were introduced in Cataclysm, participation plummeted to 4.3%. When LFR was introduced in tier 13, participation rose from its low of 3.38% in tier 12 to 4.87%. This rise also corresponded with a substantial decrease in overall difficulty.

A return to difficult normal mode raiding in tier 14 saw a drop in participation by almost exactly the amount that it rose by in tier 13. That is, an observable segment of about 1.5% of the playerbase will participate in easy raids, but not in hard ones. Interestingly, this segment of the playerbase is double the size of the “has killed Lei Shen normal” segment of the playerbase, and is likely also significantly larger than the number of players with any heroic raiding experience.

Finally, I have to echo the OP’s conclusion that Heart of Fear was a miserably overtuned instance, sporting collectively the absolute lowest participation of any raid instance since tier 9 with one exception, and that exception was the instance that was gated behind it.

Using these numbers I have assigned a difficulty score to each instance which is the average of the % of the playerbase who has killed each boss in the instance. Based off this, the following is the ranking of raid instances from hardest to easiest, since tier 9. I have excluded two instances from the discussion being throne of four winds, whose difficulty is heavily skewed because a lot of people skipped this instance in progression as it didn’t offer particularly compelling loot, and only included 2 bosses. The second is Terrace of the Endless Spring, as participation in this instance was hampered by the gating behind Grand Empress. With that said, the following is the difficulty of raid instances since tier 9, from hardest to easiest:

1) Heart of Fear – 1.57%
2) Throne of Thunder – 1.62%
3) Bastion of Twilight – 2.41%
4) Firelands – 2.43%
5) Mogushan Vaults – 2.70%
6) Black Wing Descent – 2.72%
7) Dragon Soul – 4.03%
8) Icecrown Citadel – 5.82%
9) Trial of the Grand Crusader – 8.09%

The last interesting data observation is that while difficulty has gone up substantially, the deltas in difficulty boss to boss have gone down. If you look at the drop in percentage of the playerbase killing the boss over the boss before you can find the biggest difficulty jumps. The boss with the largest jump in this area is Nefarian in BWD. Other heavy difficulty jump bosses include Al’Akir, the Lich King, Sindragosa, and Ragnaros. The largest in MoP was Elegon, showing approximately half the difficulty spike observed for Nefarian, and less than Rotface, Festergut, Putricide, Ascendant Council, and many others. For a scale for this, Nefarian clocked in at a 65% drop in kills over previous boss, Al’Akir 63%, Lich King 54%, Sindragosa 47%. Elegon clocked in at 36%, Lei Shen 27%, Horridon and Garalon 25% each. Most bosses fall into the 7-12% range. Every single boss in Heart of Fear is over 17% with Garalon being the worst offender. Twin Consorts ranks by this metric as the easiest boss this expansion and second easiest ever at 1.97%, only being out ranked by the Icecrown Gunship battle’s hilarious 0.1%.

In many ways, ladies and gentlemen, I think that rounds out our analysis quite nicely.

Raids have unquestionably gotten more difficult, LFR is not to blame for the drop, and this design direction is serving a tiny percentage of the playerbase, considering we’re not even talking about heroic mode raiding.

Worst of all, this design direction is killing both the raiding community, and the game.

Summary

What this article is designed to do, specifically, is to show the abnormally high attrition rate that the current endgame model is enforcing, while putting to bed any argument that the game has remained as easy as it always was. At this moment, most of these numbers come from the 10-man normal mode statistics, as this is traditionally where new and casual players are going to start their raiding adventure. Clearly, since the launch of Cataclysm, guilds are being obliterated by the most hack and slash raiding encounters the game has ever seen.

Obviously, the Throne of Thunder isn’t finished. Those numbers will go up by the time it does. But we’re still looking at an absolutely frightening number of guilds that are getting to see Blizzard’s best content and, quite frankly, this exclusion has got to be looked at properly. Look at these numbers for a moment. Just look at them. In a game with over 8 million subscribers, LOOK at who the current endgame model is aimed at.

Next to nobody.

If you're in a guild that only really does raiding as a group activity, you'd best hope you're good at it. Coming in to MoP, I think a lot of guilds hit the T14 "wall" (Stone Guards, Elegon or Heart of Fear), which left members ditching for guilds they might have liked less, or just falling in to LFR. Assuming the latter, those players really have to ask why they're bothering with a guild at all if it's not providing them anything.

That lack of social cohesion is, potentially, at the root of all this. When 25-man guilds were at their peak, so were subscriptions; this implies (and I mean IMPLY, not outright prove) there's a link between large guilds and whether players stick or not.

Look at it like this:

A 25-man guild has more room for redundancy (less cancelled raids), more room to try players out (who might be new) and, quite simply, more players online (increases the likelihood of players doing "something" together).

I honestly think the destruction of 25-man guilds is the biggest contribution to the subscription loss, but difficulty is what’s killing raiding.

Tiers 14 and 15, to most reasonable people who've raided for a long time, are self-evidently more difficult than anything before them, purely by going off the number of mechanics per boss, the number of mechanics that are essentially one-shots and the sheer output requirements. No realistic person is going to dispute that. There are a couple of things to note here, however:

1) This post highlights the drop per boss in both Mogu'shan Vaults and Heart of Fear to be substantially more damaging than anything before it.
2) The percentage drop in raiders, since MoP started, is substantially higher than it's ever been. Substantially.
3) Models that use one-month clears will be the guilds with enough skill to do clear an instance in one month - which isn't many. This will make the numbers after a month look better.
4) The top end of the playerbase has improved over time, in some cases drastically, and so have things like mods or UI's.
5) Classes, in general, are now far more challenging to play optimally than they've ever been and this has a significant impact.

Lastly, taking certain numbers in a vacuum doesn't tell you anything useful because they’re not telling you where people are getting stuck or where they're flat out giving up. Guilds are happy to be progressing without killing final bosses, and we saw that throughout WotLK.

Guilds in Mists of Pandaria are simply GIVING UP.

This has never happened before in such numbers. The only logical explanation for this is that the raids are harder than they've been before, and that's what the Lead Systems Designer (who's been in the game since WotLK beta) is also saying.

I stopped full on progression raiding in Cataclysm, but started my raiding journey in The Burning Crusade. Some of the people I've played with in that time, who cleared content up until 4.0, would simply not survive in this raiding scene. Cataclysm itself culled a significant number of raiding guilds, leading MoP to having less "casual" guilds from which to make the numbers look worse.

Also statistically, looking at the ditch of guilds in MoP has the least variables by default. The only other relevant comparison is probably tier 11 to tier 12 (for the outlier), and this is for the following reasons:

1) The playerbase behaviour is likely to be similar.
2) The content length (when T15 is complete) will likely be similar.
3) The gear catch-up mechanics were similar (very few).
4) LFR status was similar (either both had, or both didn’t have).
5) All raids spoken about have similar lockouts.

Between T11 and T12, guilds that didn’t complete final bosses still made it to the next tier; in general, they didn’t give up. Between T14 and T15, guilds that didn’t complete the final bosses DID NOT make it to the next tier. The only explanation is that they’ve given up, because no other explanation is rational. I’ve no doubt the numbers will go up by the time T15 finishes, but do we honestly think it’ll go up by a lot?

What, though, is making these bosses so hard? Mechanics and tuning are two different things; mechanics are binary parts of an encounter that are either done correctly or incorrectly, while tuning is how punishing it is to get them wrong. I’m of the opinion that mechanics are more complex than ever and tuning is too tight (a double whammy), but we have to be careful when we say things like that.

Many speak of one-shot mechanics, but I've deliberately avoided this debate until now, largely because "one-shot mechanic" can be a bit misleading. I mean, taking solely Lei Shen, some believe there are no "one-shot" mechanics in the fight.

Yet Thunderstruck WILL kill anyone caught in it, Decapitate or Fusion Punch WILL kill a tank who's unprepared, and Static Shock WILL kill someone who's not grouped up. Will Lightning Whip kill someone outright if it hits them?

In any event, we're looking at four one-shot mechanics on that fight alone.

Comparing that to Arthas, this might lead you to think there are definitely less. But bring up the question of Necrotic Plague - it's not an obvious one-shot, but it WILL kill someone if not dispelled in time. A more poignant example might be Defile which, again, won't one-shot anyone, but WILL wipe a raid if even one person stands in it for too long.

That's what I mean, and why I want to avoid using the term "one-shot mechanic". It can be very misleading.

I think it's probably safe to use a formula like this:

If you don't do A, the result will be B to the player and C to the raid.

Even this is fraught with danger. Let's say we're talking about Heigan's dance mechanic, just as a blunt example.

If you don't dance, the result will be death to the player.

There's no C here, at first look. Then again, let's assume B killed the sole tank for the encounter and you have no warlock, death knight or druid. The result to the raid, therefore, is a wipe.

We can then bring in our erstwhile mentioned chum, Necrotic Plague.

If you don't dispel Necrotic Plague, the result is death to the player and potentially Necrotic Plague on another raid member.

Arthas also had something else that Heigan didn't - a Berserk. So even assuming that the player was killed by Necrotic Plague and it didn't jump to anyone else, it could still cause a wipe prior to the end of the encounter by DPS being insufficient.

Now we start to see why I try to avoid getting into this.

Yet...

Let's try and distill a mechanic from the other perspective; namely, what MAKES a mechanic kill people.

I'd say, on average, it's the time to react as well as the coordination involved to handle it properly and the result to the player that makes a mechanic dangerous. I would leave out the discussion on the result to the raid, because it's almost impossible to figure out how an entire encounter will play out after an accident. It might matter a lot, it might not matter at all.

1) Time to react: Obvious, really. Do you get a second to sort it, or ten seconds?
2) Coordination: Does a single player handle it, or does it need a group?
3) Result to the player: Does it kill players, or simply just hurt them?

This is the type of direction we probably want to go if we want to debate how punishing mechanics are - basically, when things go wrong and you don't handle it properly, what the average outcome is. Given that raiding is harder than ever in MoP, I'd imagine this is where we'll find the answer as to why guilds are giving up and renting a permanent residence in LFR. Naturally, though, we still have to be careful about specific mechanics that won't conform to exactly that formula. It’s in interesting corollary, but combat ressurections used to be the sole haunt of druids – it’s now a capability found in three classes.

Are people dying more?

Moving on, another point of note for new players is the expectation of how well they’ll play their class. My view is that classes are now harder to play optimally than they’ve ever been, and the difference between a good player and a bad one has made tuning extraordinarily difficult. Maximising GCD’s, prioritising the right spells, stacking cooldowns, lowering movement, efficient use of resources and encounter specifics are all borne in mind by the top players, but never those at the bottom who also may not be using VOIP or a boss mod (pretty much mandatory for raiding nowadays).

Gemming and enchanting, as well as learning a solid rotation, is something I don't think the game teaches particularly well and it's basically the type of feedback that Blizzard can consider when it comes to designing the levelling part of the game and how statistics are presented.

I often talk about a friend of mine who levelled a hunter and was going for intellect gems. While you may laugh, it made perfect sense to him because he wanted his Arcane Shot to do more damage. To a longer term player it sounds like a daft conclusion, but to a newbie it looks reasonable. Another player I did a workshop with on Argent Dawn was trying to gem his hit rating to 100% and didn't really grasp the context of expertise at all.

Again, to someone not in the know, what makes more sense; wanting to hit 7.5% of the time or wanting to hit 100% of the time?

I know which I'd choose.

So, there are some things to take from this.

- Gem slots could be made red if you're putting in something that contradicts the item (spirit gem in agility mail, for example).
- Hit and expertise could be amalgamated with players aiming for 100% because it makes sense; the character sheet could show that.
- Gear without an enchantment could have a red border with a tooltip saying what's missing.
- The character sheet could border certain stats in red or green depending on what you want to stack or want to reforge out of.

All of these things could make it into a new expansion that would make the transition to endgame simpler for new players - and none of it need be hard or time consuming. I also think the levelling curve needs looked at so that a spec starts to play like it does at endgame a bit earlier, but that's not common across all classes. Just for the record, I also think the entire endgame model needs an overhaul but that's a tale for another day.

The proof that the current endgame model is broken, is overwhelming. Putting everything behind punishing content has clear to see effects. Yet, there is the argument that if we were to stop making heroic raids (which a tiny percentage of people are seeing), this would have a knock-on effect below that level as players have less to aspire to. But I’m not saying we should remove all hard content, merely that the presentation should be different.

Let's say a 10-boss tier had five bosses (gatekeepers and end boss) that had heroic modes. Given the tiny percent of the population that likes mega-hard content, that's plenty. It's also self-evident that those heroic raiders wouldn't go anywhere, for two reasons:

1) They have nowhere else to go.
2) They have nothing else to do.

I agree that if the hard content was removed entirely there would be repercussions beyond merely a few anonymous Internet heroes at the top, but the game doesn't need 12 heroic modes every tier. Nothing we've seen suggests that it does. In fact, there’s lots of evidence to suggest that dungeoneering is far more popular than raids, and the subscription drop in a world of less or fragmented dungeons is difficult to ignore.

We can argue that challenging five-man content shoves people away, but that's not what happened during The Burning Crusade heroic dungeon era; certainly not in general.

You can remove these "heroic modes" from the gearing curve by making their bosses drop the same loot as they do on normal, with the exception of one drop from the end boss. Take them out of the queue, and they become fun and challenging for guilds rather than the "bullrush-through-like-ass-on-fire-with-invis-pots" that challenge modes currently are.

They can then be used, assuming normal and heroic dungeons for each tier, to substitute as catch up while still providing far better content for friends and families than LFR can ever hope to achieve. So long as the loot isn't better than the previous tier (similar to the Zandalari dungeons), no harm is done to the progression gearing curve.

There's next to no downside, other than development time, but too much of that is being spent on too small a percentage of players (raids) and the subscriptions are showing that. What I'm effectively suggesting is marginally smaller raiding tiers, with far less time spent tuning the damn things, and more people are brought into meaningful endgame via what is quite probably the most popular content this game has ever produced.

Dungeons.

To wrap up, then.

My intention is not to say “this is too hard, nerf everything”. My point is purely that these huge gaps between boss kill percentages is implying that almost every boss is a wall (the Heart of Fear… Jesus), and it’s utterly destroying a raiding community that’s at its own throat. Nigh on a quarter of players that started the ascent up Icecrown Citadel bothered to turn up and face the first boss in the Throne of Thunder.

This has got to stop.

Nobody’s winning, and the game is suffering due to Blizzard trying to please next to nobody with their endgame design intent.

Thanks for reading – any and all feedback, assuming it’s constructive, is welcome.

----------------------

"How do you explain the drop off in raiders?"