Anyone with ~2100 rating in 5vs5 can fully gear out in a season. Anyone higher will gear out faster and be a better team.

Honestly the 2vs2 and 3vs3 dont even seem worth it if you plan to be the best. I think 2v2 and 3v3 will become mercy groups with the unguilded dumbass from work to help him get a few pieces of gear.

Does all this assume there will be new sets of arena gear each season?

This seems intended, with a 2man group you can essentially have groups you are unable to kill at all, and while you can form you team around being able to always draw a balanced team from 4, that’s hard without 4 hybrids. 3v3 seems to be just an longer version of 2v2, one more person to kill, one more team-mate. Unless your 3v3 or 2v2 team has 2+ healers, dps will still focus the healer down then cleanup.

5v5 is totally different, you can form a group that covers all the bases and is redundant enough that you could win against any team. Things like the standard MMORPG Trinity, and presumably a hybrid or few that can cover multiple roles will become more important. In almost every WSG custom I run with we totally stack dps, it helps us wipe PuG's quickly and unorganized / unprepared customs, but when we fight customs that have better balance then us, we have to work a lot harder and often loose. 1 Loss out of 20 games isn’t bad now, but with the arena system this loss becomes huge.

Good point. Lets assume its every 2 seasons that gets new gear and each season resets your rating.

With a true strength of 1720 you can get all the gear in 26 weeks or 260 games. 37% of all participants should have a true stregnth equal or higher than 1720.




Edit: I did a calculation of a team with a true strength of 0. Seems the minimum number of arena points you can get from playing a full season is 2700. After 17 weeks(resets after 13) you can buy a 2 hander.

Your maths kinda flew over my head (it's been twelve years since my university maths days), but do you really think this system will enable 18% of all teams to get a full set of Gladiator gear in one season?? That sounds remarkably high?

First of Bel - a season is 13 weeks, those teams wouldnt get it overnight. Second Blizzard realizes that the PvP system they had in place before 2.0 was flawed and rewarded those who had the most time, rather than the most skill (sure there are plenty of skilled HWL, but time was the deciding factor) As far as the math goes - that result is misleading. 18% of the ranking scores are above the cut off, that does not mean 18% of the teams will be above that cut off. IE there could be 15% of the teams in top 18% of possible ratings.

However as far as this post goes - for a thread entitled Optimizing Reward/Work Ratio there sure is a ton of work being done in this thread :-P

I'm pretty sure that the derivative of the probability curve I gave shows the density curve of the expected distribution of ratings. The whole point of using a logistic curve to map out the rating distribution is that it fit the data observered in the early implementations of the ELO system. In other words, I think what I said was not misleading and infact shows the chance of getting a specific rating. The proof of it is beyond my understanding of discreet mathematics, but while reading about the ELO system, logic curves and their applications, it was mentioned.

I'm told by people much better at math than I am that there is a coefficient that will determine how widely our Arena scores will swing. I am also told that right now we don't know what that coefficient is, but we'll be able to figure it out if we have the Elo scores of the teams before the fight, who won and who lost and the resulting rating change.

Knowing how much our scores change will allow us to model how long it should take for the aggregate to reach steady state. It will also answer questions raised here and another thread as to how many games is truly ideal.

Because the final scoreboard doesn't show either team's rating, finding out the other team's score before the fight will need to be done outside of the game.

With the Arenas starting tomorrow, if you're able to do this, please post your data here so we can determine a) what the coefficient is and b) if it differs from bracket to bracket.