Solid-State Roster: Foolishly Brilliant or Brilliantly Foolish?
by, 10-05-2010 at 04:01 AM (29489 Views)
Nearly two years ago, like thousands of Guilds in the WoW community, Vox was gearing up to begin raiding in the newly released WotLK, and we were presented with a unique decision: How best to manage our raid roster?Originally Posted by Albert Einstein
Certainly this isn't a decision unique to just Vox Immortalis, but it was unique for us at the time as we were drastically transitioning from a non-conformed Guild to a very tight, Strict 10-man, which suddenly meant there were choices to be made about how to manage and fill our raids in a nightly basis.
Probably like many long-standing Guilds, Vox went through many iterations and tried a variety of systems to handle raid rosters from 40-mans to 20-mans to 25-mans and down to 10 mans in Kara and Zul'Aman. We tried a "First-Come, First-Served" method where the active Guild Roster size was kept significantly larger than the maximum raid size, and on any given raid night, the first people to login and be ready to raid were invited in order until the raid filled up. We tried a "DKP-based" system whereby players were invited to a given raid based on their DKP earnings within a previous time period. We tried a "Static 3-Strikes" method where the raid roster was set in stone but any player who screwed up/missed a raid 3 times was removed and replaced. And finally, we tried an "Attendance History" system where invites were given based on previous attendance numbers, such that those that were available more often were raiding more often.
It seems futile to get into the specifics, but like any system, every method for managing raid rosters had its own set of unique issues and exploits, and when WotLK raiding was coming about and we were set to stick to a 10-man format, I once again realized a new system was in order for managing the raid roster, and ultimately, I settled on the "Solid-State Roster."
Hi, Peter. What's happening? We need to talk about your TPS reports.
The idea behind a Solid-State Roster is simple enough: The Guild maintains an active raid roster of exactly the number of slots for raid content (10 in our case), with the appropriate class/spec balance, and therefore all raid slots are guaranteed for those select players for any given raid. The term itself is just something I made up based on the idea behind solid-state drives in computing.
It should be immediately clear that, like all other systems, there are significant pros and cons to this Solid-State method, but the actual impact of those pros and cons were not apparent until the system was put into practice during WotLK and I am now able to look back and examine the effects as we approach Cataclysm.
That's a pretty fucking good milkshake. I don't know if it's worth five dollars but it's pretty fucking good.
The biggest draw for implementing the Solid-State Roster going into WotLK was the fact that it greatly simplified roster management for myself and the other Officers -- in fact, since the roster was static on a day-to-day basis, there was no management at all! Consequently, this meant a Guild with almost zero drama as there was no possibility for a dispute over who got into a particular raid and when.
Secondly, a Solid-State Roster provides a great deal of personal stability -- we all like to feel needed and important, and this is one of the most beneficial aspects of a static roster, in that everyone involved knows they are vital to the success of the raid and therefore, may arguably be more inclined to contribute more and give more effort.
Another major benefit was that with such a small raid of 10 players, a simplified solution such as Solid-State just made sense logically. It seemed pointless and even convoluted to complicate the issue unnecessarily, and nothing could be simpler than 10 Active Raiders = 10 People in the Raid.
Then we get into the more obtuse but still certainly important benefits, such as those pertaining to coalescence and morale. With a Solid-State Roster, it is much easier to create a group of individuals who get along with each other than with a roster that constantly shifts players in and out. Moreover, progression is faster (in most cases) with a static roster because the raid is already well-versed in the strengths and weaknesses of everyone around them and therefore can adapt quickly when needed.
Finally, with a Solid-State Roster, loot is distributed in the most efficient manner possible since it is not spread out unnecessarily, and therefore the overall strength of the raid accelerates faster than in a more sizable roster.
We're gonna need a bigger boat.
Now that I've sung the praises of the system in practice, there are also a number of issues that a Solid-State Roster will eventually bring to fruition.
By far the biggest issue is volatility: Any minor change in circumstance, schedule, skill, or otherwise for an individual in the roster greatly impacts the entirety of the raid. Obviously steps can be taken to lessen the impact, but suffice to say if even one player shows up late, retires, or otherwise is unavailable, the raid must absorb that impact, often by losing progression for days or even weeks, or at the very least cancelling a particular raid night altogether. Even planning ahead by making connections with other Guildies and friends as possible substitutes does not alleviate the situation, as in almost every case, any substitute will be under-geared and certainly unexperienced with the content and methods of the rest of the raid, to a degree that prevents most forms of active progression for that given evening.
Another problem with a Solid-State Roster is recruitment; or more specifically, the timespan gap between the moment we need a new recruit and the moment the slot is filled. This issue is caused in large part by the fact that most people will not actively choose to join a Guild where he or she cannot raid regularly just in hopes of getting a full-time raid slot sometime in the future. That said, a large part of this particular issue was due to the limitations of the Strict 10-Man system, whereby we were unable to invite players who would be willing to wait for an open slot while raiding with pugs or friends, since Strict 10-Man rules limited the number of 25-man achievements by guild tag.
I do not blame the 10-man Strict community for those limitations, as certainly that was the only reasonable option and of course Vox, as a Guild, actively chose to participate. However, suffice to say those limitations contributed a great deal to making recruitment with a Solid-State Roster much more difficult than it likely would be in a post-Cataclysm environment, where anyone and everyone can be invited to the Guild without regard of their achievement or outside-raiding status.
Finally, a Solid-State Roster can present issues of player burnout, whereby an individual feels a greater sense of obligation to continue playing past the point where he or she might otherwise have stopped. At the very least, Solid-State requires a great commitment from each individual and a player that might be content to raid 15 hours a week for a year might find raiding for 20 hours a week in a static roster to be too much and want to stop after 3 months. However, this idea about player burnout is very subjective and thus not very factually definable, so while I can't state with certainty it is actually a negative (as the "personal stability" benefit mentioned above is virtually a mirror counterpoint), it does seem a logical assumption.
What do you mean funny, funny how? How am I funny?
Ultimately, having used the system for nearly two years, the inevitable question I'm having difficulty answering for myself is: Does a Solid-State Roster work? The answer? "Yes..."
"...and also No."
I think a Solid-State Roster can (and does) work wonderfully for certain small raids, and when everything is clicking, it is a thing of beautiful, raiding nirvana. There was a span of time midway through Ulduar that Vox went nearly three months without a single hiccup in our raids, every scheduled raid started and stopped on time, everyone was present and accounted for, and we just blew through achievements and content, bursting our Old Faithful right through Yogg's outer crust.
On the other hand, we've also had sudden, often severe slumps and setbacks. Even though I didn't explicitly state it above, the true downfall and issue with a Solid-State system is us -- the players; the human beings. Imagine a computer that controlled a robotic arm and hand, using very basic English-language instructions, was given the following task:
1. A row of 20 identical wooden blocks lay before you.
2. A single bucket is on the ground below you, between the blocks and your location.
3. Starting from the center, in the most efficient manner possible handling only one block at a time, place a total of 10 blocks into the bucket and return to your starting position.
If we run our little basic computer program, the robotic arm will recognize that the most efficient manner possible means to keep arm movement distance to a minimum and in other words, increase the speed of the task. Therefore, the robot will pick blocks closest to the center and work outwards. If we numbered our blocks from left to right as #1 through #20, the robot would start with either block #10 or #11, as they are equidistant from the robot (the exact center is between the blocks), and therefore either choice is the most efficient. So, if #10 was picked first, then the second selection is block #11 as that is now the closest block. The third choice would then be between #9 and #12, as both are equivalent again, and so on until all 10 blocks are picked. At the end of the process, we're left with a row of blocks #1 through #5, a gap of 10 missing blocks, and at the end blocks #16 through #20.
If we setup the blocks again and run our program again, one hundred or even one million more times, will we ever get a different result?
The reason for this is simple: The computer was told to use the most efficient selection method every time, so even though the block #1 is identical to block #10 or #11, the computer will never actually pick block #1, as it has no reason to do so. Even though the result of 10 wooden blocks in a bucket would be the same, it serves no purpose, and in fact hinders itself, by picking blocks not in that middle chunk of 10 every single time.
Consequently, if we were able to ask a computer the most efficient way to setup and manage a World of Warcraft raid, it would absolutely tell us something like a Sold-State Roster. That is, it would find the first 10 players to match the criteria of our instructions, and it knows using those 10 players, every single time for every single raid, is the most efficient method. It would never randomly decide to find an 11th player to replace one of the original 10, because creating an 11th variable in a 10-maximum-variable system is pointless and inefficient.
From that perspective, if we were all computers we'd easily see every raid using a Solid-State Roster of some sort, but because we are human beings, we throw a wrench into things every once in a while. This certainly doesn't mean every player participating in a Solid-State Roster must be a robot... but it helps. The more variables or unknowns for an individual (be in schedule, job, family, friends, whatever), the harder it is to fit that person into the conformity that a static roster demands. As with any team sport, you are only as good as the people around you.
Total Trackbacks 0