Blood Sport: Arena and the (old) Elo ranking system

Want to crush your enemies, see them driven before you and hear the lamentation of their women? Blood Sport investigates the entirety of all things arena for gladiators and challengers alike. C. Christian Moore, multiple rank 1 gladiator, examines the latest arena strategy, trends, compositions and more in WoW.com's arena column.

Listening music Young MC's Bust A Move. Because I can.

Last week We talked about griefers in arena and some of the tools they can use to make WoW a much more miserable place.

This week We'll be talking about the old arena system in The Burning Crusade. From season 1 to season 4, we used a completely different system to gauge where we stood among everyone else on the arena ladders. The shift from an Elo rating system to a Gaussian Density Filter caused more changes than most people realize. We'll get into that a bit later -- don't worry, it's not all mumbo-jumbo.

Arena representation dropped 65% from season 4 to season 5. One of the reasons players decided to no longer participate in rated arenas was the change in the arena rating system.

The Elo rating system

At the beginning of The Burning Crusade, developers wanted to implement a fair way to rank players in a competitive-style PvP deathmatch environment (i.e., arena). When arena was introduced, it came with an Elo rating system attached. If you didn't already know, the Elo rating system is the way chess players (grand masters included) are rated. Chess ratings are actually very similar to arena ratings. Someone with a 1,500 chess rating is an average club player, while grandmasters are 2,500 or higher. The number-one-rated player in the world, Magnus Carlsen (a 19-year-old prodigy) is currently rated 2,813.

Many other competitive games use the Elo rating system as a way to gauge official rankings. College football, national Scrabble organizations and Magic: The Gathering all use the Elo rating system as a way to change rankings based on the result of matches played between participants. To say that the Elo rating system is well established is an understatement.

The general idea is that the best players will have the highest ratings while lesser players will not. This system works best when a large aggregate amount of games are played. It's much easier to tell who is the best participant when all participants have played 160 games rather than 16. Organizations that have more games played in a short amount of time (such as chess, Magic and WoW) are much more accurate than ones who do not (such as college football). I would argue that WoW is the most accurate of any Elo system, as games take the shortest amount of time and can be played with little downtime in between each match.

Elo is very easy to explain in WoW terms. Everyone starts at 1,500. If Team A (who is at 1,500) fights team B (who is also 1,500), there are 16 points at stake. The winning team will go up to 1,516, while the losing team will drop to 1,484. If they play again, Team B can win more points off of Team A because they are rated lower. The cap on points is around 30 -- so if a 2,800 team goes up against a 1,800 team and get a disconnect, they won't lose 100+ points.

This is very unlike the system we have in today's arena. In today's system, everyone starts at 0 (or 1,000). Instead of there being one rating that matters (team), there are three -- team, personal and matchmaking. Matchmaking rating is really the most important rating because the other two ratings will adjust themselves to the matchmaking rating after a certain amount of games are played, but I digress.

Season 1/Season 2

In season 1, players delved into arena for the first time. There were no personal ratings, no matchmaking ratings. The only thing that defined you as a player were two numbers -- your team rating and your resilience. Chat messages looked something like this:

[2. Trade] [Progladiator]: 1900 3v3 team LF 150+ resilience mage, pst. We're paladin/hunter, need a mage.
[Nubberz]: I only have 100 resilience, I can leave my 1700 team if you want.
Progladiator]: You wanna do some games?
[Nubberz]: Okay.
[Progladiator]: We're just going to play five or six games, if you want to get points off of us you can, we're just trying out a lot of mages.
[Nubberz]: Oh, cool. Yeah, I already did 14 or 15 games with my other team.
[Progladiator]: That's cool, we have a 2100 team we're saving up for points this week, if you're good we can play with you on that one, you get something like 1000 arena points a week on it.
Nubberz]: Sweet!

Because the only rating that mattered was team rating, team-hopping was very common. I remember playing 3v3 with seven or eight different teams in the span of a single week in season 2 and still ending up with around 1,000 points when it was all over. If I did that today, I'd have to play 40 games with the same team to get that many points. That's because my personal rating would be so much lower than any team I joined, even if my matchmaking rating were vastly higher than the teams I'd be playing on.

The old system encouraged better players to "try out" new players on their teams. Today, if you don't have some kind of proven track record of arena prowess, you're probably not worth the risk for a more experienced player. That gladiator might need to drop his personal rating and matchmaking rating, just to see if you're any good. He'd rather play with other "proven" people.

The old system also encouraged team-hopping -- playing with different people on different kinds of team compositions is fun. It's very fun. Not just for the top tier of the arena ladders, either. If new players want to get anywhere they have to grind out games against teams that are vastly better than them. New players can't just leave their team -- they need to bite the bullet and take losses again and again rather than try to play with other new people.

I enjoyed the old Elo system much more than the current implementation. I think most people did, too. In fact, I know most people did. After all, we've had a giant drop in arena representation from season 4 to season 5 and it just hasn't recovered since.

Next week

Although this was a bit of awesome nostalgia, we'll discuss why the ranking system was changed in the first place and what benefits the new system offers us. It will be a good bit dissimilar to this article, where we have more opinion and less facts. Think of this article as a warm-up for next week. We're setting the ground work right now to talk about what's really going on.

Want to ascend the arena ladders faster than a fireman playing Donkey Kong? Check out WoW.com's articles on arena, successful arena PvPers, PvP and our arena column, Blood Sport.