New Formula for Distributing Ranked PvP Rating

[Lyons_Crest]

Given how our current Ranked Rating System works and how it encourages many of it's victors to quit while they are ahead, I am going to suggest an idea that might improve things for everyone and encourage more people to Que and continue to Que for ranked.

 

The Goal here is to keep people playing - get more people into the mix - expand the pvp community across all servers - limit the long duration between Ques so they pop more regularly - Reduce the amount of rage towards less knowledgeable players or geared - and possibly reduce the amount of trolling in ranked pvp matches.

 

To achieve all of this we must change how rating is gained :

 

a) Ranked players would continue to win X rating based on their wins (same as now)

 

b) Ranked players that lose their matches would not gain or lose any rating.

 

* The change in line " B " is major as it would provide many benefits.

- Eliminates any risk of losing rating

- Produces more Competition within the ranks to continue Queuing up.

- Provides an opportunity to expand and encourage more people to Que in ranked

- Reduces the amount of time many people have to wait in order to get a ranked que pop

 

This formula is simple - as it eliminates risk and lessens the blow for losses in ranked PvP. This would encourage more people to focus on their " Main toons" if they want to stay on top in the leader boards. It does not reduce competition in anyway be cause we will be expanding the playing field. By removing the " Risk " factor - we allow more people to join Ranked PvP without having negative consequences, this in turn takes pressure off of those that have already achieved a high rating and those who have not. The Quality of play does not change either as their is NO rating gained if a team losses, so no rewards for baddies. If some exceptional player joins the season in it's 3rd or 5th week they too can rise up through the ranks and keep the other exceptional high rated players on their toes as well.

[Evolixe]

Also creates plays most = wins system.

 

Fundamentally flawed. But nice try anyhow.

[Raansu]

Why does this suggestion continue to come up? Not losing any rating is THE WORST idea ever. At that point why have a rating at all?

[Tellenn]

You will never get enough people to queue ranked as long as you give them the option of regs. While a small amount want harder fights the vast majority dont. Even pretty decent players and guilds. They will always take the easier option.

[Raansu]

You will never get enough people to queue ranked as long as you give them the option of regs. While a small amount want harder fights the vast majority dont. Even pretty decent players and guilds. They will always take the easier option.

 

You cant force force ranked onto people. A lot of people don't even like arenas. Lets face it, bringing arenas into this game and removing 8v8 ranked was probably the biggest mistake BW has ever made when it comes to pvp in this game. They should have just focused on improving 8v8 ranked or cross server queuing. I like Arenas, but BW made a huge mistake bringing them into this game. A game that was never designed or meant to have 4v4.

[Capt_Beers]

Why does everyone on this forum think they can come up with a better rating system in between queues than the physicist who came up the Elo rating system?

[foxmob]

Why does everyone on this forum think they can come up with a better rating system in between queues than the physicist who came up the Elo rating system?

 

because you need approximately a 3:1 ratio to gain elo from ~1200 on up. and that's a pretty crazy winning percentage. I'm also not sure how the system actually factors opponent/teammate rating, cuz my gain/loss is always predictably within 3 points no matter who's on what team.

Edited September 21, 2014 by foxmob

[Raansu]

Why does everyone on this forum think they can come up with a better rating system in between queues than the physicist who came up the Elo rating system?

 

Because the current rating system is a joke? With that said, the OP's idea is terrible.

[Capt_Beers]

because you need approximately a 3:1 ratio to gain elo from ~1200 on up. and that's a pretty crazy winning percentage. I'm also not sure how the system actually factors opponent/teammate rating, cuz my gain/loss is always predictably within 3 points no matter who's on what team.

 

My guess is that's because of the population. When you get to the 1500 range you are at the top for the most part. The people you are playing against are far below you in terms of rating so you are always expected to win. Thus if you lose you lose a lot of rating. If we ahd the population where we had a healthy amount of consistent queuers the numbers in each range of rating would be healthier so you would be more likely to to play people with a similar rating.

[foxmob]

My guess is that's because of the population. When you get to the 1500 range you are at the top for the most part. The people you are playing against are far below you in terms of rating so you are always expected to win. Thus if you lose you lose a lot of rating. If we ahd the population where we had a healthy amount of consistent queuers the numbers in each range of rating would be healthier so you would be more likely to to play people with a similar rating.

 

I don't know, so I can't refute that or agree. but I will say that I am often grp'd with ppl pretty well above and pretty well below me (200-400 in both directions). so how are we basing this? on the average team elo? lol individual elo? but the individuals are like 2k and 1200 on the same team. I think, at this point, a simple win/loss would be better, cuz trying to incorporate teammate's rating and opponent's rating is a bit of a joke when the various teammates' ratings are so disparate.

[Tellenn]

ELO can only work if theres a significant number of players not a handful and if people dont sit on their ratings. Edited September 21, 2014 by Tellenn

[t-darko]

ELO can only work if theres a significant number of players not a handful and if people dont sit on their ratings.

 

Quite right . . . And the only thing that would attract more players is!? Well looking at the past soon to be 3 years. I'd say the only thing that would make more players do ranked, is to guarantee that no one will lose!!! Just to sum up the thoughts of the Quitters, Whiners and 'The rest of my team sucked' threads ...

[Haivon]

Regardless on whether you lose or gain rating when you lose, people still enjoy tending their ego's. Better to reach the required rating and stop queueing so you don't risk a loss to tarnish your reputation than keep queuing and risk some losses which might make people believe you to be inferior. Generally the ranked-active players on each server is rather known and are generally either

A) Considered good among the plebians, yet most likely has an ego

B) An ******e

Someone who is considered good and knows it would not risk a loss for fear that people wouldn't consider him or her or it to be as good. Someone who is an ******e would not like to lose in fear of his haters having leverage against him.

 

Only way to truly increase ranked is either remove ego or get the egotistical people to stop queuing so plebs can begin. Regardless of wins or losses, even in regs, people would rather win than lose and whenever a newer ranked group begins queuing it's almost a guaranteed loss due to the amount of elites that'll probably be queuing as well for easy wins. They wouldn't gain much and it would probably take a merciless number of Ranked Warzones before they finally gain rating from their opposition.

Edited September 21, 2014 by Haivon

[Metthew]

I might be (as of now) still **** at playing my sniper, but this I can answer.

 

The nice property the ELO rating in chess has is:

 

1/ It's a predictor of talent, i.e. if you match people with equal ELO ratings, you'd expect them to win with %50 probability, and someone with higher ELO will more probably win against someone with lower ELO.

2/ It is convergent with infinite games between the players, if no player is so good he never loses. It is because of the self-correcting property.

 

If you make people lose no ranking, then you open the possibility of exploiting the system such that neither property* no longer holds, unless you set everybody's ranking to a constant and also award no points (because over the whole population, you obviously expect %50 wins and %50 loses, in case of even numbered teams for any game). If you award constant ratings per win, you allow people to grind up their rating instead of having the most rating by being the best. I think it's pretty obvious why this is so, though I can show you why if you don't see it immediately.

 

However, if you give diminishing points based on the opponents' comparative ELO, then you don't end up fixing how the system can be gamed. Assume two equally bad teams, and a great team in the system (wherein we ignore personal ELO and say there's a team ELO, for simplicity in demonstration). If the bad teams play all week for 5 hours a day inbetween themselves, they will drive their ELO up at a constant rate - their ELO will always be about equal and each wins half the games. Now the great team might win even all their games, but if they aren't just playing enough, they'll never catch up with the bad teams' rating. This implies that this system also leads to a grind.

 

The ELO system is fine. What we're seeing is a problem of incentive to do ranked, because there is little to no reason people should keep doing ranked PvP. The availability of ranked comms does not incentivize playing ranked over unranked, and the season rewards are mostly irrelevant for incentive concerns until it's sufficiently near the end of the season. Furthermore, if a player believes he is assigned a higher rating than he deserves, he will not want to keep doing ranked PvP. Therefore the "sustained rewards" you can get from doing ranked PvP instead of unranked PvP isn't good enough, the wait times are too long to justify picking ranked over unranked, and season rewards are too far off and too far ahead to be of any worth.

 

My next post will deal with how one could actually solve this.

Edited September 21, 2014 by Metthew

[Metthew]

Alright, first of all, to clear something: the only way to "lose rightfully earned rating" because your team is "full of bads", is if there's a visible skill difference among factions - you need to be much more skilled than the rest of your faction members, who are on average worse than the other faction's players. More correctly, what you want to be equal is something that weighs skills of players, and activity rates per skill level. It's also assumed that the number of ranked PvP'ers is "sufficiently large".

 

What this means is, if there's to be leaderboards with meaningful ratings:

 

1/ they should be seperated by server AND faction.

2/ the games should be cross-server and leaderboards should be regional (Korea/SC shows clearly that regional skill differences are very real. Furthermore, due to latency issues any cross-server matching would need to be still regional).

3/ the games should be mixed faction and leaderboards should be separated by server.

 

Now that we cleared this, let's talk incentive. As stated before, here are some major reasons why someone might not want to queue ranked PvP:

 

1/ If you just want to have a game, unranked queue pops much more frequently.

2/ Unless you really want to grind up a Brutalizer (which is a pretty brutal grind, moreso than the PvE gear I think), ranked PvP match rewards are not that much.

3/ You might believe your current rating is greater than what it actually should be.

4/ After grinding your Obroan in unranked, you're bored to death from the same old warzones, and the prospect of arenas intimidates you even more.

 

The possible solutions to the problem of not enough ranked PvP are pretty simple in concept, though I don't know which ones are feasible to code in and which ones aren't.

 

1/ Combine the unranked queue and the ranked queue: Here's how it will work - if you queue unranked, you get bolster and unranked rewards. If you queue ranked, you don't get bolster (or you get a different bolster), and you get ranked rewards. I only included the possibility of a different bolster because as of now, even Obroan is bolstered a bit in unranked PvP (your armor rating will go up). The rating you get from this will still be indicative of your skills, because everyone who tries to ranked queue will play alongside the same unranked+ranked PvP'er population.

 

2/ Increase the sustained rewards for doing ranked PvP: By sustained rewards, I mean incentives. This isn't necessarily just more ranked comms and more cash, though it could work. In fact, if you allowed the money to "pool" (the more time passes when nobody completes a ranked game, the more the prize money will be next game), you would have a "sliding high incentive" that would net yield in some equilibrium match rate, and you could then slide it higher or lower depending on how much ranked PvP you want.

 

However, you can simply divide the season into smaller blocks, and give awards biweekly or monthly. With a seperate "in-season" rating (i.e. you need to play each season to keep your rating, only your in-season rating is transfered forward), this would force people to play all the time to reap the rewards. The end-season rewards can be allocated either by giving points per mini-seasons, or by checking who survived with what rating at the end.

 

Obviously these aren't the only two ways. The important point is that the rewards for doing more ranked PvP would be better somehow. A small detail is, more credits would mean more ranked PvP'er which would mean an even greater credit input to the economy - so inflation. But this is a PvP board, not a game economy board, and to solve the inflation problem the game would need to be more EVE-like anyways: a separate discussion!

 

3/ Make rankings decay in time: Something like %5 decrease every day (or a greater percent per week) would not change the nice properties of the ELO rating, but it would disallow people from sitting on their ratings. Simple and straightforward.

 

4/ More warzones, with different formats, for ranked: Including maybe even 16v16. If ranked was somehow more fun, people would go more for ranked.

 

One or multiple of these solutions could be used for great effect. Personally I think 1+3 would be enough by themselves to move the geared part of unranked population to ranked without depopulating unranked (because we aren't generating more PvP'ers out of thin air). 4 is eventually crucial in sustaining the PvP population, because if people get bored of the game they eventually realise they can just stop playing. I am indifferent to 2.

[DarrelK]

^ This guy knows the deal.

[foxmob]

rating decay means absolutely nothing (to me at least) because highest rating earned is the only thing that determines success (season rewards). also...I play 4 active toons. I know a lot of ppl play 8+. you need to be very careful about rating decay. I think it works better for registered grp comps. but outside of that, it's either a bad idea or the decay would have to be so minimal as not to be effective.

[Metthew]

rating decay means absolutely nothing (to me at least) because highest rating earned is the only thing that determines success (season rewards). also...I play 4 active toons. I know a lot of ppl play 8+. you need to be very careful about rating decay. I think it works better for registered grp comps. but outside of that, it's either a bad idea or the decay would have to be so minimal as not to be effective.

Well, your motivation for disagreeing with a rating decay seems selfish - though a too harsh rating decay would indeed disallow a truly skilled person from getting 8 characters in the leaderboards (more than one of each class is excess, you must admit). It depends on what Bioware wants for the leaderboards I guess.

 

A monthly larger decay could solve the problem though. The problem with continuous decay is, you wouldn't have enough time to get your characters up again. Seperating up the decay time would remedy that.

[Raansu]

Honestly without cross server no matter what BW does nothing will really worked.

 

Team ranked is dead because the good players get bored and the average players get tired of playing the same 3 teams and losing. So people just stop queuing as there just isn't enough variety in a single server queue so the ELO system can never function correctly.

 

BW failed in not anticipating this. I don't care if they expected before launch that this game would have millions of people playing. Even WoW has cross server queues. BW's arrogance and lack of forethought has caused a lot of problems throughout the years though.

 

Solo ranked is just a complete joke and should just be removed completely. That entire format is completely pointless as only classes that can carry or if you queue sync will you ever have a high rating. Case in point, I know for a fact I'm way better on my marauder than my juggernaut, but in solo rating my marauder is like 980 something (mostly thanks to queue syncers...) but I started queuing on my juggernaut and I'm almost 1500 rating on him only because the class can carry and can live through the dps race.

[sanchito]

nice posts Metthew, thanks :-)

i wish someone at bioware would actually read them...

Edited September 21, 2014 by sanchito

[foxmob]

Well, your motivation for disagreeing with a rating decay seems selfish - though a too harsh rating decay would indeed disallow a truly skilled person from getting 8 characters in the leaderboards (more than one of each class is excess, you must admit). It depends on what Bioware wants for the leaderboards I guess.

 

A monthly larger decay could solve the problem though. The problem with continuous decay is, you wouldn't have enough time to get your characters up again. Seperating up the decay time would remedy that.

I fail to see how it's selfish. the season rewards are predicated on highest earned. decay only prevents you from parking your toon at a legit rating and doing something else, then coming back to the first toon and trying to go higher. it has zero affect on your rewards/highest earned. if I want to, selfilshly, stop Qing after hitting 2k on the VG, decay isn't going to do a thing to my placement at the end of the season. it would only affect what the leader board reflects for "current" rating...which is kind of useless. I don't quite understand why the LB is structured like that.

[EarnestBunbury]

nice posts Metthew, thanks :-)

i wish someone at bioware would actually read them...

 

If you put the freaking word CARTEL and BioWare can make more MONEY, then these ***** would read it. I hate pvp until there is a close game, like maybe once a week. I don't give a dam if I win or lose, I just want to have the ability to do something cept have 3 people dot me up with "uncleansable" dots and run away, or stealth up to me while i have the USELESS 30 stacks of "ready for anything" ( if by "anything" they mean "to freaking die") and take my health down to 1/4 before I can activate anything. Been PvP'ing for a while but it is just the worst selection process, 4 PUB dps, vs 2 PUB heal and 2 PUB dps? Even a retarded person can even that out.

 

Running sm TFB for the 200th time makes me keep looking for another game but the mechanics here are very good, it's the rest of the junk like devs and cartel that ruins this game

[Metthew]

I fail to see how it's selfish. the season rewards are predicated on highest earned. decay only prevents you from parking your toon at a legit rating and doing something else, then coming back to the first toon and trying to go higher. it has zero affect on your rewards/highest earned. if I want to, selfilshly, stop Qing after hitting 2k on the VG, decay isn't going to do a thing to my placement at the end of the season. it would only affect what the leader board reflects for "current" rating...which is kind of useless. I don't quite understand why the LB is structured like that.

Ah, I would also give rewards on the most recent standings. Or pull points through mini-seasons, like F1 does.

 

Also I apologize for the provocative language, I didn't intend it - you can read "selfish" as "self-motivated". Because someone who only PvP's with one toon consistently may not care at all about the decay.

 

But yeah, with considering only highest rating earned, no decay would be needed, nor would it be useful except for adding a small grind factor.

[Capt_Beers]

I don't know, so I can't refute that or agree. but I will say that I am often grp'd with ppl pretty well above and pretty well below me (200-400 in both directions). so how are we basing this? on the average team elo? lol individual elo? but the individuals are like 2k and 1200 on the same team. I think, at this point, a simple win/loss would be better, cuz trying to incorporate teammate's rating and opponent's rating is a bit of a joke when the various teammates' ratings are so disparate.

 

I have no clue how BW decides what the "team rating" is. My best guess based on what I have heard form people is that it takes the average of the 2 teams and makes that the team rating. Once the match is over it compares your rating to the opposing teams rating and adds/deducts points accordingly. So if you are a 1500 rated player and lose to a team that has a 2k, 1400, 1200 and a 900, average rating 1375 (we can see a problem here already as 2k is well over that) you would stand to lose a good chunk of rating because comparing your 1500 to 1375, of course you should have won. That's just my guess but it seems to fit with what is actually happening.

 

They would have to answer this themselves though (hint hint Eric). It's a more than fair question and I think we deserve an answer, dare I say, we are entitled to one.

[KumbayaGOD]

i agree with OP , lose no point and win get some point, we need to change the current system of ranked , you cant expect the same result by using the same elo system that is implemented now , it just wont work ( see season 1 and 2)

 

rating decay also no, i mean this game is not only pvp some people may do pvp and pve and by doing so you expect people to do just pvp and nothing else , a big no no , and bioware wont do it