First things first:
Are raids broken?
I think they probably are.
Not specific raids, but the architecture of raiding in general, as specifically implemented by World of Warcraft and Rift – the two MMOs I’ve raided in most recently. Raiding is designed to have certain difficulty levels; in Rift, some raids are 10-player raids and some are 20-player raids, while in World of Warcraft (at least in its most recent incarnation; it started out like Rift), each raid has a 10-player and 25-player version, plus difficulty levels within the player-count “flavors”.
The problem with this model is that there is always going to be some distance (and animosity) between players at different difficulty levels – “casual” players are derided by “hardcore” players for not being willing to put the time and effort into “real raiding” (and in some cases told that they shouldn’t even be allowed in the instances), while “hardcore” players are mocked as “no-lifers” by “casual” players. Moreover, it’s a system that fundamentally disadvantages players who don’t get an early start in raiding or skilled players who only have the opportunity to play with 10 other people, because by and large, when raiders are putting together a group for a given raid, they seem to want to trivialize it as much as possible.
(That is, incidentally, completely understandable – because raids require items beyond player skill, like food buffs and potions, each player death or group wipe is a set of consumables wasted, and so the fewer player deaths and group wipes a raid can manage, the better for everyone – or at least everyone who’s on farming duty.)
How to fix it
The most frustrating thing about WOW raids is that they almost have it – but not quite. The tiered-difficulty system still has the problems mentioned above. The solution is not to remove the difficulty – but to remove the tiers.
To fix raids, difficulty should be progressive, taking cues from more casual games like “Wii Sports” – or even from the games’ own leveling process. In the competitive games in Wii Sports, each player starts out at skill level 0, and – when playing against the computer – is matched with an opponent just a little above 0. If the player defeats that opponent, their skill goes up depending on how well they played, and the next opponent is just a little higher than their new skill. This goes on until a player plateaus at their actual current skill level – but over time, as the player improves, that number keeps climbing up, until they ultimately reach the peak of their physical and mental ability to play the game.
Raids in MMOs can take a similar tack. Assign each character – or even each spec – a “raid skill”*, which indicates how good they are at defeating bosses of the current level; average the raid skill levels of each member of the raid; and use that number to determine the initial difficulty of the raid. As the raid’s skill level goes up, the raids’ difficulties go up, and the value of the items dropped by the raid goes up as well. (This is nothing new in World of Warcraft, at least; we’ve seen that Blizzard can scale gear, thanks to WOW’s heirloom items – and we’ve seen that they’re okay with giving out multiple versions of the same item with different stats, thanks to Cataclysm’s three-tier approach.) At the top levels, allow raids to acquire unique mounts, legendary and artifact items, special achievements, etc.
As a further improvement, allow the leader of raid groups below a certain skill level to reset the raid as many times as she wants, but limit loot to the first boss kill per week. (In fact, given that loot is per-player in Mists of Pandaria, this should be extended in World of Warcraft to per player per week, so that players who join on the third or fourth run aren’t cheated of loot.) This lets low-level raid groups get practice and improve their play without giving them an excessive loot advantage.
The end result of this is that raiding will, interestingly, both become far more accessible and remain challenging far longer; entry-level players will be able to see the entirety of the content at a very low skill level (and receive very low-level loot for their efforts), and high-level players will constantly be challenged, because their next run will always be just a little bit harder than their last.
Thoughts?
This idea came to me today, and I’m sure I haven’t thought it through completely; I’d like your feedback, because I’m positive there’s something I’ve missed.
* Yes, I’m aware of how much this sounds like GearScore; the difference is that this number reflects the player’s actual ability, and not just her ability to acquire gear.
3 Comments »
RSS feed for comments on this post. TrackBack URL



Comment by Masith — April 14, 2012 @ 4:24 AM
Interesting idea and its great to see people thinking outside the box but I have a few questions/problems:
1) Do you anticipate guilds clearing every week and the encounters just getting a little harder each clear? I worry that this would result in every tier being a continuous farm. I also wonder if players will still get the rush from killing a new boss when it is just slightly harder than last week.
2) Is there a risk of an everlasting increase in difficulty? Each week the difficulty goes up you also get better gear so you can kill stuff at an even higher difficulty and so get even better gear next week etc..
3)Would guilds be reluctant to bring in new players with a low “skill score” because it would mean their average score would drop so the encounters would drop worse gear?
4) If the difficulty increased at a faster rate than gear inflation could account for thus solving problem 2 wouldn’t guilds only stop increasing difficulty when it was already too hard and they couldn’t kill anything anymore.
5) How would competition between guilds be defined?
[Reply]
Comment by Twice — April 14, 2012 @ 5:44 AM
Three major issues with incremental leveling systems like this:
1: Coarseness. How fine do you chop up the rewards? If you don’t chop the rewards up into a number of levels equivalent to the number of internal difficulty levels, you get a number of asymetric rewards – and it gets worse the farther apart your “item reward levels” and “internal raid difficulty levels” are in absolute numbers. (3 levels of gear and 15 levels of difficulty will be much worse than 5 and 10, respectively)
2: Assumed losses. The Wii model is designed to keep you around 1:1 win/loss until your skill plateaus. Can you imagine the screaming if Blizzard implemented a system that would keep EVERY raider at a 50% kill rate until the cap out at, effectively, Heroic difficulty?
3: Aggregating skill ratings. Even assuming equal skill, character Skill Ratings will vary as a function of unavoidable circumstance. (Bob has to bring his Shaman to heal the first few weeks, and now that the normal healer is back Bob’s Druid tank is 20% behind the Rating curve of the rest of the Guild…) If Blizzard doesn’t include a min/max Skill for a particular level of difficulty the whole progressive difficulty thing goes poof unless an entire Raid ALWAYS progresses together. If they do, Bob up there gets benched because his Druid can’t play with the big boys through no fault of his.
I think the idea is great, but implementing it would be a nightmare. And I just can’t see how it’d work.
[Reply]
Comment by shamus — April 15, 2012 @ 6:43 AM
I think there are two problems with the idea. Firstly, scaling of encounters isn’t sadly as simple as moving a slider on boss health/damage. The latest run at this from Blizzard gives us three levels of raiding: LFR, Normal, Heroic. Each step up not only gets harder in terms of damage/healing required but also adds features to the fight. This changes the progression from being linear to having jumps. Even without the added features, there will be thresholds where damage jumps from not killing you to killing you. (e.g. fading light on Ultraxion).
Okay, that’s not insurmountable perhaps, but certainly quite tricky.
Then comes the second issue – how do you measure players’ skill? In theory this is simple for dps, kind of doable for healers, and tricky perhaps for tanks. Except even then there are problems. For example, healer throughput is bounded by the amount of damage people take. If people keep standing in fire then healers stats will look good (or everyone will die). Once people stop sanding in fire suddenly healers stats drop. You have to start tracking why people died. Was it due to avoidable damage (again, fading light is a good one here), or was it due to not getting enough heals. If the latter, which healer is to blame? Even evaluating dps is tricky as it can vary so much between fights. Tank and spank, vs. multiple targets, vs. target switching and burst dps (e.g. Spine). Single target, multi target and aoe play quite differently and one person might be uber at single target, okay at aoe but suck at multi target.
Then of course you throw the whole lot together with 10/25 people of varying skills…
Except, I wonder, looking at the arena for comparison. People have individual and team scores. Raid team score could be set initially based on average player score and then adjusted based on how quickly you kill a boss and how many people were still alive at the end. Maybe even work out a score based on how long till wipe and when people died. Then you could try and work out some individual stats for dps/healers based on length of fight and healing done vs. raid damage taken, and adjust them based on the raid score adjustment. Though … healers are a bit of an issue due to a good healer being able to hide the stats of poorer healers to an extent. No idea how you might score tanks. Probably mitigation based and a lower weighted dps score.
So maybe we can score people. Raid tuning though I suspect would still have to be fixed tuned rather than auto scaled, but who knows.
[Reply]