Two Games, One System

Nono, nothing gross.  Recently two games have come out into the sparse MMOFPS genre.  Both promise limited team sizes with skill-based matchmaking.  Matchmaking in these games isn’t about making it so you win all the time, or most of the time.  No, it is about pitting you against people of near equal skill.  The goal is, really, to have you win (or lose) about 50% of the time.  That means you are often pitted against people who are at or about your skill level and sometimes you win, sometimes you lose.

However while it generally works in one it almost always does not work in the other.  So here a quick look at why matchmaking in Global Agenda (GA) generally works and why matchmaking in All Points Bulletin (APB) will generally fail.The short answer it set sizes.  However this is a facet most people who lament the poor performance of APB’s matchmaking tend to ignore.  So on to the long answer.  For that let’s first look at a matchmaking system that works then contrast that against APB’s.

Global Agenda, for all its flaws, did have a robust matchmaking system.  Sure some matches were horribly imbalanced but Hi-Rez’s assertion that if your long-term average PvP win percentage is 50% (or very close to it) then the matchmaking system is, over the long term, working.  I put in about 2 months into GA and my personal stats, while not a perfect 50%, are pretty darn close.  I’ve had some blowouts, I have been steamrolled.  But the vast majority of my matches have been fairly even.  For all intents and purposes Global Agenda’s Matchmaking system works.  But, why?

Global Agenda has a single server.  Each match is an instance of 20 people, 10 versus 10.  It pulls from the entire server’s population.  At any given time there are 15-20 PvP matches at play.  That’s 300-400 players in active games with another 20-30 looking for matches.    There are, of course, some limitations.  Players can opt out of certain game types, teams of up to 4 can skew the matching a bit and the set is divided between those levels 1-29 and those level 30+.  So we’re looking at, roughly, 10 to the 300th just for your own team.  Then 10 to the 290th for the other team.  Big, big numbers.

Now, contrast that to APB’s data set.  APB has two servers so already the population is split.  Then each server has several instanced districts.  People can only be matched with others in the same district.  Each district, at the time of playing, is limited to 40v40.  This, too, is another limiting factor since you can only be matched with the 40 people of the opposite side*.  40 people, if you happen to be in a full district.  That’s a far cry smaller than the aggregate combinations that GA has to work with.

Furthermore in GA you have the option of joining the match or refusing to join.  You are not given any information on which side you’re on, what map you’re going to play, who’s going to be on the side with you (outside your 1-3 prefab team mates), who is going to be on the other side.  Accept or refuse.

APB’s matchmaking is different.  One person starts a mission.  During the course of that mission the matchmaking system offers matches to other people.  They are told the type of mission, the number of people on your team, and your rating.  Human nature, being what it is, is going to skew the results away from the hard or even matches onto the easy matches.  Why would I want to take an APB against 2 people above my rating when I can take an APB against someone 3-5 below my rating?  Furthermore why are such APBs offered?

Look back to the data set.  40 people.  Most of whom are probably in missions.  In puttering around on APB and refusing missions offered to me I generally bounce between 2-3 of the same APBs.  Why?  Because that is all APB has to match me against.

APB’s matchmaking system can only do so much with the pool of opposing players.  When that pool boils down to 4-6 people in 2-3 groups of 1-3 people it can only offer those up to me.  Which, naturally, if they are too hard (or too easy, I hate rolling people) I am going to refuse! For APB’s matchmaking system to improve they need to restructure the fundamental basis of the game itself.  They need much, much larger districts so they can dramatically increase the pool of available players that the matchmaker can draw upon.  It’s simply a matter of set sizes.

* This is not strictly true.  Enforcers are indeed limited to 40 criminals.  However criminals in APB can be dispatched against other criminals.  So a criminal’s pool is actually 80 minus their current team size (76-79) players.  However since 1/2 the population is Enforcer I went with those numbers.  In reality, however, even with double the set size the matchmaking is no better.  The above example of the same 2-3 offers coming up was when I was on my criminal.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s