Tuesday, January 12, 2010

How the Dungeon Finder beat Gearscore

Imagine for a moment that you could for any given group in World of Warcraft use a miraculous addon which would exactly tell you the overall power score of that group. Being miraculous the addon not only counts measurable things like level, class, talents, group composition and gear, but also harder to impossible to measure things like skill and cooperation into account, and transforms all that into a single number. As that number tells you absolutely everything about a group, you can now compare it to a given difficulty level of a dungeon, and come up with an accurate prediction of that group in that dungeon. The scale would look something like this, from lowest group power score to highest:
  • Group wipes on trash
  • Group manages trash, but wipes on bosses
  • Group can finish the dungeon, but only after repeated wipes
  • Group finishes dungeon without wipes, by using all available means including crowd control
  • Group finishes dungeon without crowd control
  • Group uses AoE on trash
  • Group uses AoE on boss
  • Group finishes dungeon in record time
Now personally I would say that at least for a first attempt on a new dungeon, the middle range of this scale is the most fun, where you use all the available possibilities of your character and your group, and have to figure out how things work. But obviously the further you proceed on that scale, the faster it goes, so with people trying to get the most emblems per minute of repeated heroic runs the AoE end of the scale is preferred by most players. The more powerful the group you bring, the faster you get your rewards.

Now that miraculous group power score addon doesn't exist, and can't exist. However there are addons like Gearscore who at least try to measure the power of an individual character. Take 5 characters with a high gearscore, apply some common sense on group composition, and you end up with a high group power score. This is why Spinks thinks that Gearscore is necessary to evaluate PuGs. The same thinking goes into various other requirements people advertise when putting together a pickup group, whether that is achievements or gear checks.

There are two major problems with the approach: One is that with players preferring faster rewards, the highest score is always the best. The other is that whatever existing score you use, measuring skill correctly is impossible, so group leaders tend to overestimate the need for gear and achievements to compensate for the possibility of low skill. In the extreme somebody organizing a Naxxramas pickup group will only invite people who can demonstrate that there is absolutely nothing in Naxxramas that would still be an upgrade for them. That is completely counterproductive, and excludes players trying to gear up from such pickup groups.

Funnily enough the Dungeon Finder looking-for-group tool introduced into World of Warcraft with patch 3.3 uses a basic form of Gearscore, whose exact workings are secret to prevent players from manipulating it. But what you can do is open the Dungeon Finder (or its hidden brother, the Raid Dungeon Finder with /lfr) and go to the page where it allows you to select a specific dungeon to go to. If you are undergeared, you will find a little lock symbol on the harder dungeons, that is neither you nor a preformed group containing you will be randomly assigned these locked dungeons due to you not having good enough gear for it.

But there is an important difference between the Dungeon Finder and a PuG leader using Gearscore: The Dungeon Finder isn't all that worried about the possibility of somebody having very low skill, nor is being able to finish the dungeon in record time a requirement. Skill in playing World of Warcraft, like many naturally occuring things, is distributed over a bell curve, a Gaussian distribution. That means the chance that somebody is exceptionally good or exceptionally bad at something is very, very small. A random group will nearly always find their skills grouped around the average (in spite of how skilled the individuals think they are). And it will be symmetrical, with good and bad players balancing each other out. In one extreme case I visited at level 79 a normal dungeon where the damage meters at the end revealed my mage in the middle of the 3 dps, with one level 80 full epic hunter at 4.5k dps at the top, and one level 79 death knight with a measly 0.5k dps at the bottom. Due to at least the dps part of a group being a shared responsability, that group did just fine, even with me doing AoE and not being too slow finishing the dungeon.

In short, the Dungeon Finder is balanced enough to put a group together which even in the less ideal cases will finish the dungeon in good time and with few or no wipes. And due to putting the requirements less high, there is even a good chance that the people involved can use some of the loot that drops, which after all used to be the purpose of a dungeon before we got emblem rewards. In spite of putting a group together that few PuG group leaders using Gearscore would have invited, the Dungeon Finder has a very high probability of finding a group able to finish the dungeon successfully, often even in good time.

Now, while I do see the added difficulty when going from 5-man leveling or heroic dungeons to 10-man or even 25-man raid dungeons, both from a point of having a balanced composition and the issue of raid lockouts, I do wish that Blizzard would extend the Dungeon Finder to a fully operational Raid Finder, and not just the hidden and unused fragment they have up to now. There is currently a big hole in the end game content with the Dungeon Finder having filled up heroics, and the guilds mostly doing the higher up raid dungeons, with few or no people visiting Naxxramas. Some further work on the Dungeon Finder tool could make finding a raid as easy and fun as finding a heroics group. And then we would find that in fact you don't need full T9 gear to do a wing of Naxxramas, as the PuG raid leader with his Gearscore addon demands.

No comments:

Post a Comment