Change to Solo Ranked Rating Calculation

[BalphoWan]

I am sure other threads exist on this topic, but I wanted to chime in with my opinion and request for feedback. I was wondering if others agreed that the rating calculation for solo ranked should be adjusted. Simply using win/loss ratio as the only factor provides a distorted ranking system in my opinion.

 

The problem is that too much of your win/loss ratio (in solo ranked) is determined by the players that are placed in your group. Sometimes, I feel like I am pulling a slot machine when I que up for solo ranked. I have done about 100 solo ranked matches and I am batting a little over 500. And more often than not, I can tell right away if my team is going to win or lose immediately after I enter the WZ.

 

I would suggest a calculation using a combination of:

 

1) win/loss ratio

2) matches played

3) medals earned

4) MVP votes

 

Thoughts???

 

[EDIT]

 

A suggested scoring system that I posted later in the thread:

 

1) Win: +10 points

2) Loss: -20 points

3) MVP Vote: +0.5 point

4) Medal Earned: +1 point

5) Game Played: +2 points

 

Obviously, this is just an example, and those numbers are completely arbitrary, but it could serve as the basis to compare each player’s scores to one another every two hours to establish a real time point ranking system. Also, I would not throw out the system in place now which compares your ranking to the opposing group to adjust your scores accordingly. That differential could be used as a flat ratio to be applied to the above point system.

Edited November 21, 2013 by BalphoWan

[Evolixe]

And how exactly do you propose that would work..

 

You realize you are asking for the entire Elo system to be overhauled completely, right?

[BalphoWan]

I do realize that. I also realize that under a new system, such as the one I am suggesting, players would not be able to get real time updates after each match as to the increase or decrease to their score, because it would need to be calculated in comparison to the rest of the players in game in real time.

 

For example, lets say Player No. 1 plays 4 matches and wins all four, but gets no MVP votes, and only 4 medals. Lets say Player No. 2 plays 8 matches and goes 4 and 4, but gets 8 medals per match (for a total of 32 medals) and 15 MVP votes. Now, under the current system, Player No. 1 would gain approximately 40 points, whereas Player No. 2 would break even.

 

That system, in my opinion, is flawed. I think that when the rankings come out every 2 hours, it should assign arbitrary values to the categories I suggested in my initial post and make comparisons based on those values amongst all of the players in the game. An algorithm for that calulation would not be difficult. I am not saying it would be perfect (nothing would), but it would be less based on RNG than the current system.

Edited November 21, 2013 by BalphoWan

[cashogy_reborn]

Sometimes, I feel like I am pulling a slot machine when I queue up for solo ranked.

 

Thats because you are. Making Solo Queue a ranked game mode, and giving RWZ comms for it, is Bioware latest "biggest mistake".

[Technohic]

Thats because you are. Making Solo Queue a ranked game mode, and giving RWZ comms for it, is Bioware latest "biggest mistake".

 

Yeah, and the Elo is not the slot machine, its the payout. no matter what Elo system would be in place; random PUG is a gamble. They should just make solo ranked into "ranked training" for people to do arenas to get up to speed and then remove arenas from the regular Q.

[BalphoWan]

Yeah, and the Elo is not the slot machine, its the payout. no matter what Elo system would be in place; random PUG is a gamble. They should just make solo ranked into "ranked training" for people to do arenas to get up to speed and then remove arenas from the regular Q.

 

I am not sure I agree with that. I like the idea of players who don't have 3 other good PVP players to que up with on a regular basis being able to get ranked with other lonely PVP players out there. I just think the rating system is not very indicative of player skill. I think it is far too one dimensional (for solo ranked, obviously not grouped).

[Technohic]

I am not sure I agree with that. I like the idea of players who don't have 3 other good PVP players to que up with on a regular basis being able to get ranked with other lonely PVP players out there. I just think the rating system is not very indicative of player skill. I think it is far too one dimensional (for solo ranked, obviously not grouped).

 

I understand, but any system is going to not be a good reflection when you have no control of who you get placed with. It boils down to solo ranked being meaningless anyway and it's already viewed as a gearing mode.

[Kerensk]

I wish they would some additional factors into the ranking. I haven't been on my game recently but alot of other players play amazing and deserve some bonus thrown their way. Or a loss where 1 player still dominated but eventually fell because half the team got creamed don't deserve a full "loss".

 

The problem is coming up with some metrics that are fair and people can somewhat agree on. Personally I'd like to see some hard to play classes get some multipliers and the faceroll classes get some handicaps, to encourage people to play more diverse classes and not just FOTM. But it's hard to come to a consensus of who is OP, plus it changes by update.

[Capt_Beers]

 

1) win/loss ratio

2) matches played

3) medals earned

4) MVP votes

 

 

Medals and MVP votes? Are you kidding? Talk about meaningless stats. Plus I see no way MVP votes contributing to your rating being abused.

 

If you don't like the way that solo queue works queue for group.

[Seengularity]

They should make it so that only people in stock conqueror (minimum) and 60~70+ valor would be able to queue up for solo ranked. You're still subjected to that slot machine solo ranked is, but it'd be more forgiving... Edited November 21, 2013 by Seengularity

[Capt_Beers]

This thread makes me sad.

[Technohic]

This thread makes me sad.

 

LOL Why's that? It is a bad system. Problem is there is no great solution. Least for solo. Very hard to make an unbiased system determine how well you did in a loss.

[BalphoWan]

Medals and MVP votes? Are you kidding? Talk about meaningless stats. Plus I see no way MVP votes contributing to your rating being abused.

 

If you don't like the way that solo queue works queue for group.

 

I understand what you're saying. And I expected that repsonse. You are correct. But like I said in a previous post, I don't think it's asking too much for PVP players who don't have a regular group of other players to que with to be able to compete in a fair ranking system.

 

Also, I understand your concern about MVP votes and medal counts being used to calculate your ranking. That is why I only suggest that they be calcualted as one component of your ranking. And there is always going to be impetus for abuse in any scoring system. It's just the nature of the beast. I don't think that is reason not to try.

 

But I am not married to those categories. It's just a suggestion. My biggest concern, for instance, is a situation in which a new player happens to win 8 out of his first 10 matches and is ranked higher than someone with a record of 50 and 40. I just don't think that is fair at all, and under the current ranking system the former player would be ranked higher than the latter.

Edited November 21, 2013 by BalphoWan

[BalphoWan]

They should make it so that only people in stock conqueror (minimum) and 60~70+ valor would be able to queue up for solo ranked. You're still subjected to that slot machine solo ranked is, but it'd be more forgiving...

 

I like that idea as well. A barrier to entry for those who are not properly geared I think would decrease the instances of good players getting reductions to their ranking due to RNG.

[BalphoWan]

LOL Why's that? It is a bad system. Problem is there is no great solution. Least for solo. Very hard to make an unbiased system determine how well you did in a loss.

 

Agreed. That is the heart of the problem. It is a difficult one to solve. But I do think the current system can be improved without too much of a software overhaul.

[Capt_Beers]

LOL Why's that? It is a bad system. Problem is there is no great solution. Least for solo. Very hard to make an unbiased system determine how well you did in a loss.

 

I'm just tired of reading threads where everybody tries to re-engineer a system that has been working since the 60's. The man who developed the Elo system was a physics professor who developed it to be used to rank world class chess players. His system works and has been working. Everyone seems to think if they lose a match it's not because the other team was better than them it's because they are great and their teammates are bad.

 

Everyone says the same exact thing "Why should my rating suffer because of stupid teammates?" The funny thing is all your "stupid teammates" are saying the same thing about you. If you don't want to rely on random people get a group together.

 

The ranking system seems fine to me because when I look at the top solo rankings on POT5 I see all the names I expected to see on top. I can recognize most of the names as people I have played against and are damn good. And they win with random teammates.

[HaLeX]

I'm just tired of reading threads where everybody tries to re-engineer a system that has been working since the 60's. The man who developed the Elo system was a physics professor who developed it to be used to rank world class chess players. His system works and has been working. Everyone seems to think if they lose a match it's not because the other team was better than them it's because they are great and their teammates are bad.

 

Everyone says the same exact thing "Why should my rating suffer because of stupid teammates?" The funny thing is all your "stupid teammates" are saying the same thing about you. If you don't want to rely on random people get a group together.

 

The ranking system seems fine to me because when I look at the top solo rankings on POT5 I see all the names I expected to see on top. I can recognize most of the names as people I have played against and are damn good. And they win with random teammates.

 

Thanks for the uneeded history lesson. If you would read your own post you would realize elo was designed for 1 vs 1 or one team vs one team not 4 random vs 4 random. Do you see the difference? ELO does not work in 4 random vs 4 random and EA needed to come up with a new system. They weighted the first 10 games heavily then just reverted back to chess elo for the rest of the games. Bozo the clown can do better than that really.

[Capt_Beers]

Thanks for the uneeded history lesson. If you would read your own post you would realize elo was designed for 1 vs 1 or one team vs one team not 4 random vs 4 random. Do you see the difference? ELO does not work in 4 random vs 4 random and EA needed to come up with a new system. They weighted the first 10 games heavily then just reverted back to chess elo for the rest of the games. Bozo the clown can do better than that really.

 

So what's your system?

[Kerensk]

The elo system also relies on a wide enough player pool to create ranked groups. When you only have about 8 people queued the best players still get teamed up against the worst players and the rankings get thrown further off. Without additional metrics the lower level players can't break out of the downward spiral. The system only sees loss after loss even if they play better than most because they're repeatedly matched against the top players

 

Add to that having 6 of the 8 players on the same guild and you have a further imbalance where you have 4 people on voice chat against 4 people whose rely on text (and 2 disgruntled guildies knowing loss is probably inevitable and hoping to be matched with their team next round). It's the effect of having solo players up against a team and the elo system is further broken.

 

Now nobody wants to play the guild and the player pool shrinks more, further breaking the elo system. Soon ranked PvP is dead (again).

Edited November 21, 2013 by Kerensk

[BalphoWan]

I'm just tired of reading threads where everybody tries to re-engineer a system that has been working since the 60's. The man who developed the Elo system was a physics professor who developed it to be used to rank world class chess players. His system works and has been working. Everyone seems to think if they lose a match it's not because the other team was better than them it's because they are great and their teammates are bad.

 

Everyone says the same exact thing "Why should my rating suffer because of stupid teammates?" The funny thing is all your "stupid teammates" are saying the same thing about you. If you don't want to rely on random people get a group together.

 

The ranking system seems fine to me because when I look at the top solo rankings on POT5 I see all the names I expected to see on top. I can recognize most of the names as people I have played against and are damn good. And they win with random teammates.

 

That is a good point. The names at the top of the list on my server are also, for the most part, the names I expect to see there. But I also see alot of names that I don't recognize at all, with only 10 wins to their name. That is the imbalance I am addressing. And look, this isn't chess. Chess is not a game where you have to rely on others to win. If it was one on one, the Elo system would be perfect.

[JP_Legatus]

That is a good point. The names at the top of the list on my server are also, for the most part, the names I expect to see there. But I also see alot of names that I don't recognize at all, with only 10 wins to their name. That is the imbalance I am addressing. And look, this isn't chess. Chess is not a game where you have to rely on others to win. If it was one on one, the Elo system would be perfect.

 

There's nothing wrong with the solo rank calculation. If anything, it's the one rating that functions properly in swtor.

 

Think about it: You've got 4 guys in an arena match, lets say I put some random baddie into a team with the three best players in the universe. He probably will drag the team down a bit, but they will also be holding him up a lot. And since there are only three bad other teams in the queue, he's going to win all his games because the other three guys know they can win anyways. His rating will be way higher than it should be, and it won't affect the good players' ratings at all because they are still winning.

 

Now think about rated solo queue: you've got the system, which is admittedly ******, putting all the players in swtor into, eventually, an equal number of ****** games as good games. All the good players get mixed in with the bad ones and play each other and with each other a random number of times. Where are the three good players going to end up? Probably at the top somewhere. Where's the baddie going to end up? At the bottom.

 

TLDR it has problems but at the end of the day it's a working system. If anything you should be looking at the old system with skepticism.

Edited November 21, 2013 by JP_Legatus

[-Messiah]

 

I would suggest a calculation using a combination of:

 

1) win/loss ratio

2) matches played

3) medals earned

4) MVP votes

 

Thoughts???

 

Matches can't be based on these things...

 

1) you could keep versing the same crap team over and over again to boost your win/loss. This is why you get less rating for some wins than others

 

2) any retard could queue all day and your rating would be based on who plays the most rather than skill.

 

3) medals is just a fun system bioware added to warzones lol this doesn't indicate anything! And should not even be looked at by a player to judge their performance in an arena.

 

4) this may be good, maybe give an extra +1 rating per vote. Something a little extra because your team recognizes how good you were playing.

 

What I think needs to be fixed is the match making system... High end players want to have some good games at their highest level. This is basically never happening and it's causing threads like this to come up because some players know they are doing good but keep losing because bioware are putting ******* with good players.

[BalphoWan]

So what's your system?

 

Like I previously stated, I would assign pre-determined arbitrary point values for certain “achievements,” like:

 

1) Win: +10 points

2) Loss: -20 points

3) MVP Vote: +0.5 point

4) Medal Earned: +1 point

5) Game Played +2 points

 

Obviously, this is just an example, but it could serve as the basis for an algorithm to compare each player’s scores to one another every two hours to establish a point ranking system.

Edited November 21, 2013 by BalphoWan

[BalphoWan]

Matches can't be based on these things...

 

1) you could keep versing the same crap team over and over again to boost your win/loss. This is why you get less rating for some wins than others

 

2) any retard could queue all day and your rating would be based on who plays the most rather than skill.

 

3) medals is just a fun system bioware added to warzones lol this doesn't indicate anything! And should not even be looked at by a player to judge their performance in an arena.

 

4) this may be good, maybe give an extra +1 rating per vote. Something a little extra because your team recognizes how good you were playing.

 

What I think needs to be fixed is the match making system... High end players want to have some good games at their highest level. This is basically never happening and it's causing threads like this to come up because some players know they are doing good but keep losing because bioware are putting ******* with good players.

 

1) Technically yes, but you could always keep the differential in ranking between the two teams playing as a factor to be calculated in the new scoring system. And you can take advantage of the same crappy team requing under the current system, so I don’t see your point here, unless I am confused.

 

2) I am not sure that this is a bad thing. Obviously, games played should not be weighted nearly as heavily as games won, but in a “season” with no predetermined amount of games, those who play 150 matches versus those that play 20 should be granted some kind of reflection of that in their ranking.

 

3) Fair point. But I don’t necessarily think it would be a bad idea to use player opinion to some small extent to calculate the rankings (admittedly probably very small). But I feel it would make the MVP vote much more meaningful, which could be a good thing. And just because something is ripe for abuse, doesn’t mean it should never be done. This is an MMO, and we are part of the internet. Everything is ripe for abuse.

 

4) Now you’re getting it.

[-Messiah]

I'm just tired of reading threads where everybody tries to re-engineer a system that has been working since the 60's. The man who developed the Elo system was a physics professor who developed it to be used to rank world class chess players. His system works and has been working. Everyone seems to think if they lose a match it's not because the other team was better than them it's because they are great and their teammates are bad.

 

Everyone says the same exact thing "Why should my rating suffer because of stupid teammates?" The funny thing is all your "stupid teammates" are saying the same thing about you. If you don't want to rely on random people get a group together.

 

The ranking system seems fine to me because when I look at the top solo rankings on POT5 I see all the names I expected to see on top. I can recognize most of the names as people I have played against and are damn good. And they win with random teammates.

 

Does this include the match making system of how the teams are paired? Because that's my problem. This isn't a game where 1 player can carry to win such as DOTA where literally 1 player can carry when the other 4 players are doing bad. This is why it isn't working. Maybe if a really good player could 1v3 or even 1v2 but that doesn't happen unless the 2 or 3 players are retarded(really really retarded)