Dragonflight Season 1 Rated PvP Rating Discussion

Clarification on Ratings and Titles for PvP Dragonflight Season 1

We’ve seen a lot of discussion about ratings in the 2v2, 3v3, Rated Battleground, and Solo Shuffle brackets over the past few weeks and we wanted to clarify a few points, especially about “rating inflation.” We have alluded to a “rating inflation” system in rated PvP before, but we’ve never explained what it is, how it works and why we like it.

Rating inflation allows us to increase the “mean rating” over time. What is the mean rating? Well, it’s probably easiest to say that a player with exactly average skill will have a “mean rating”, due to how the math of rating calculations works. Players above the mean rating have above-average skill, while players below the mean rating deserve our love and support. Under the hood, rating inflation increases that mean rating over time, meaning that while at the start of the season an average player will quickly gravitate toward a rating of 1500, towards the end of the season their rating may rise to 1800. Originally, this system was developed to counterbalance the effect of players who earned a high rating early in the season then opting out of playing the rest of the season.

We feel like every PvP season has three stages:

  • Early season - When most players are earning Conquest on a weekly basis to acquire gear.
  • Mid-season – When most players are finishing their gear sets, gearing up alts, and trying out different classes.
  • End of season – When a lot of players are pushing for seasonal milestones or titles.

Rating inflation was meant to ensure that the real action for titles happened at the end of the season. However, it also benefited other players, who could feel progress over the course of a season as they pushed their rating. One downside of this system is that when the season rolls over, rating inflation resets, and players feel unable to achieve a rating that they had only the week before. Another variable here is the rate at which we inflate the rating. We’ve tried a lot of different numbers over the years. What we’ve learned is that increasing the mean rating at a rate of less than 15 points per week doesn’t allow people to overcome those early season ladder squatters. But increasing inflation at more than 20 points per week results in feeling like no rating is safe. It also can lead to super-inflated ratings at the end of the season.

Something we decided to do for Dragonflight Season 1 was to delay the start of inflation until week 10. We didn’t feel like inflation was necessary during the early season when players are playing to earn Conquest to complete their gear set. Chasing titles and end of season rewards isn’t entirely relevant yet. But now at week 10, players will be completing their Conquest sets, so inflation will kick in at 20 rating per week to help dislodge people from the top of the ladder and reach whatever their seasonal goal might be. Since we’re starting it later in the season at this aggressive rate, we won’t reach ratings that are highly overinflated. In retrospect, we think this delay was too long and we’re considering starting it earlier in the future, perhaps at a reduced rate.

While we have you here, we also want to address faction-specific titles. Originally, these were added to encourage players to play both factions and this was largely successful. Prior to cross-faction groups in Arena, social dynamics were the primary driver behind which faction got the most player in organized PvP. Players decided which faction had the racials they preferred and flocked to that side. If you wanted to play with those players, you needed to join that faction. Now, players can play whichever faction they prefer and cross-faction titles are no longer necessary. We’ve decided that end-of-season titles will once again be faction agnostic in Dragonflight Season 1.

Thank you and we hope this provides answers to some of the feedback you have delivered regarding this PvP season.

3 Likes

woah yeah exponential functions are cooler than linear and step functions. Try a Cumulative distribution function

1 Like

I would have my Conquest gear, if it was finaly uncap like you promise it 2 weeks ago

2 Likes

Correct me If I am wrong but they announced that conquest will be uncapped 14/15.02.2023 a while ago which has happend today for EU servers, it was never stated that it will be uncapped before.

To the topic I think some kind of inflation is necessary for the sake of progress, majority of WoW’s players aim to push rating and If they stagnate it is frustrating. A problem I see with a high inflation rate is that for a lot of people competing at high rating, the start/mid of the season feels meaningless, I know plenty of people which don’t bother to play 3s because the rating will be most inflated in the last week.

The result of this is they dont feel any urge to queue in the early/mid season and just push their titles in the last 2 Weeks which results of queues being often dead early on.

I feel like it is hard to create a balance becasue if you remove inflation this might solve the issue of people only queueing towards end of the season but allows people to push rating at the start and then sit this whole season and cashout the title at the end of the season which should also not be the goal.

1 Like

Correct me If I am wrong but they announced that conquest will be uncapped 14/15.02.2023 a while ago which has happend today for EU servers, it was never stated that it will be uncapped before.

To the topic I think some kind of inflation is necessary for the sake of progress, majority of WoW’s players aim to push rating and If they stagnate it is frustrating. A problem I see with a high inflation rate is that for a lot of people competing at high rating, the start/mid of the season feels meaningless, I know plenty of people which don’t bother to play 3s because the rating will be most inflated in the last week.

The result of this is they dont feel any urge to queue in the early/mid season and just push their titles in the last 2 Weeks which results of queues being often dead early on.

I feel like it is hard to create a balance becasue if you remove inflation this might solve the issue of people only queueing towards end of the season but allows people to push rating at the start and then sit this whole season and cashout the title at the end of the season which should also not be the goal.

Removing faction based cutoff is obviously a good decision now that arena is available via cross faction.

Edit: i was wrong Blizzard say its on the week of the 15.Feb wenn conquest is uncaped

How can read got a clear advantage

1 Like

The only thing you have done and accomplished is to force people to play solo shuffle for absolutely no reward pushing them away even further by removing actual rewards from 2s and zoning them completely from 3s but I guess you enjoy seeing the same people and their alts in the 3s ladder.

RIP Blizzard 1991-2008

2 Likes

No?

12 days ago:

With weekly resets the week of February 15…

You don’t understand why there are more people playing Solo Shuffle than competing in teams. Suppose that everyone has only 2 hours of games after work or school every day. What he wants most is to open the game and click to enter the arena, rather than spend half an hour looking for teammates. The current Solo Shuffle mode needs to play six games. In high rank, the energy consumption of six consecutive games is huge. Solo Shuffle is a more modern game mode, and can be fully explained. It is suggested to add higher rewards, equipment appearance or mounts

1 Like

This content is more appropriate to use the logarithmic function. (Inversa of the exponential)

Maybe you should google what i wrote; ”cumulative distribution function”

So u want good players to play for the whole season (which happens anyway, they just make an alt), and u force me to quit the game since im hardstuck 1300 playing against multiglads and boosters all day long, tired of losing games against people i have no chance against. cool.

2 Likes

I guess thats a l2p issue if you are hardstuck at 1300

I do not need to google, I use this daily. You are the one who needs to google, the cumulative distribution function will depend on what distribution is relevant to you and how much you accumulate by the value you choose of the independent value. To say to use a CDF in general, this is not defined. The only thing in common is it starts at 0 ends at 1 for a probability measure.

For your purpose, it is better served to use some sigmoid or hyberbolic tangent, this is what you really mean if you say to me that I need to “google CDF” which is going to give results of normal distribution because google only knows PageRank not math. You google CDF of the Delta function, it is obviously not what you mean.

no it is not what i meant :smiley: also dont pm me please am not interested in your highschool bs :smiley: grats on graduating babbys first classification course, but if you want a function to inflate rating you dont use classification functions like sigmoid for it, dumba$s :smiley:

You think 1/(1+e^-x) can only be used for the classification, skip your rudeness. Don’t care about you, mathematics heathen.

CDF 1/(1+e^{-(x-1500-20t)/500}) where t is number of weeks. That you think cumulative distribution function has shape without a given distribution shows your weakness in mathematics.

Toxic and dumb, you don’t PM me :smiley:

Can we get an Update on PVP Talents cause we realy need the dispell protection on our freedom.

Why is this in the General not the arena section

Because rbg is rated pvp too