Guild Wars 2

ANET Plz. Why is the current leaderboard and the one before this based on games played? Sure the current one is more based on games won over played, but the ppl who play the most are obviously gunna win the most late night pugging farming solo Qrs lol so it again comes down to whoever plays the most. (Especially when losses don’t even hurt you, they just don’t help you much) It should be quality of wins not quantity like it use to be, but fix the old decay so you cant decay 50 spots after a month of not playing and jump back to top 10 after 1 game win or lose. That was more cosmetic decay than actual decay. Your rank could show 600 but would actually be 6. That was lame but at least that LB was based on quality of wins. Winning 5 matches in a row vs no names barely moved you compared to beating ppl in the top 25. Plz stop making the LBs a grindfest.

Should be Quality, not Quantity. Do we really need to prove who is “best” based on how many wins, or can we simply show who is best by having them beat those who sit at the top?

+1 Old leaderboard was fine, the only thing that was stupid about it was the fact that an old returning player could come back, play 1 game and be where he was before he quit.

Yeah. Exactly this.

SoloQ + TeamQ (full party only) + old LB /w proper decay + seasons + ready check = win?

Prob too hard to let go of all the work they did to revamp the current new system.

100% this.

Address decay and volatility and it would have been like adding oil to a wheel.

Instead we have invented a rectangular one.

Just awkward.

its seems like its Anets philosophy do never revert changes even if they are proven to be worse than before. We wont get the old Leaderboards back. Sadly

My only concern is how would this work exactly. When one of the best players doesn’t play for a while, are you suggesting they get put back into matches with much lower-quality players to grind back up via ratings? Or perhaps keep matching people with the current MMR formula, and have a separate ranking system that allows your to decay and then….how does that person climb back up the leaderboards? How quickly should wins allow them to climb back so that its not a pure numbers game, but not a cheesy “play 1 game after 2 months and now I’m rank 1…lel”?

The old “non-decay” seems closely tied to a leaderboard that only depends on rating, in my mind. Could you try to flesh out exactly how you think people should decay, and how quickly they can grind back up? Should points for winning be weighted higher for games among high MMR players?

The current system could probably be tweaked to perform similar to the old system, but there are important questions that need to be answered to make that happen.

It seems the old system rewarded players too much for not playing (playing when ranked highly was more likely to just hurt your MMR/ranking, while you could consistently gain rating by playing 1 game every 3 days), while the new system is just a numbers game. How do you think a system should work that isn’t either of these?

Not really sure why they wouldn’t just use the elo system, like every single esport does. It’s so simple…

Decay is not necessary, although it helps against high ranked people not playing games to not risk getting their elo lower.

BlackBeard.2873:

My only concern is how would this work exactly. When one of the best players doesn’t play for a while, are you suggesting they get put back into matches with much lower-quality players to grind back up via ratings? Or perhaps keep matching people with the current MMR formula, and have a separate ranking system that allows your to decay and then….how does that person climb back up the leaderboards? How quickly should wins allow them to climb back so that its not a pure numbers game, but not a cheesy “play 1 game after 2 months and now I’m rank 1…lel”?

The old “non-decay” seems closely tied to a leaderboard that only depends on rating, in my mind. Could you try to flesh out exactly how you think people should decay, and how quickly they can grind back up? Should points for winning be weighted higher for games among high MMR players?

The current system could probably be tweaked to perform similar to the old system, but there are important questions that need to be answered to make that happen.

It seems the old system rewarded players too much for not playing (playing when ranked highly was more likely to just hurt your MMR/ranking, while you could consistently gain rating by playing 1 game every 3 days), while the new system is just a numbers game. How do you think a system should work that isn’t either of these?

The old system, if decay had been fixed, would have worked fine. People would have been forced to play to keep their ranking instead of abusing the /afk feature of it.

Indeed, I agree 100% the actual leaderboard is pure trash&grinding so it needs to be fixed (and honestly in 2 months since the last season, i hoped in a more effort&fantasy from Anet devs, basically nothing has changed for lb).

But the old leaderboard had many issues, besides the “i go afk for ages and come back #1” there were:

1). Too much volatility. Basically you went up & down with a lot of positions and it wasn’t clear what was the distance between players

2) Too easy to exploit. And if you have a fixed decay, is not hard for people who get a lucky strike to keep an high position with few games (for example 1 game every 2 days)

3). If you don’t face (for example for different time play) people above your position, you don’t gain positions and you risk to lose more positions. This is a big issue, considering Top50 most of all.

Honestly i’d prefere a score system leaderboard (it is more clear the distance between players) with a number of cap matches….I mean you know, in every sport league competition, teams play the same number of matches and the competions is fair, i don’t see why can’t be the same

My 2 cents.

Anti.9156:

its seems like its Anets philosophy do never revert changes even if they are proven to be worse than before. We wont get the old Leaderboards back. Sadly

Agree. Really sad for the game but o well. Other games and hobbies out there

I 100% agree with you Quilja.

BlackBeard.2873:

It seems the old system rewarded players too much for not playing (playing when ranked highly was more likely to just hurt your MMR/ranking, while you could consistently gain rating by playing 1 game every 3 days), while the new system is just a numbers game. How do you think a system should work that isn’t either of these?

There’s a solution that has already been porposed several times.
You use whatever “grind” based point system (like the current one, for example), but you also introduce an additional multiplier which value decreases the closer to the MMR the total score is.
If the amount of games required to converge onto the MMR is large enough, the deviation/volatility of the MMR itself should be minimum. Also, since the amount of points per game is really small when close to the MMR, it should be quite hard to capitalize on a MMR boost caused by a lucky win streak, even harder if some kind of decay would be applied to the LB score as long as it changes faster than the MMR deviation (In fact, under this kind of seasonal scoring, the deviation increase caused by inactivity isn’t even necessary).

In any case, prior to any LB refinement, I would trash the current MMR Glicko2 system and try to develop an specific system for GW2 and its particularities.

Bottom line is that I want a leaderboard to reflect skill.

Right now it does not.

Back then it kinda did.

It wasn’t perfect but I knew that the people at the top were really good players.

I prefer THAT to this. No matter how flawed it was.

Yep. The only thing Anet had to change about the old leaderboard was the silly non-permanent decay. Instead, the leaderboards are worse than ever. It’s ridiculous.

Anet just likes trying out terrible ideas.

There was actually nothing wrong with the previous leaderboard besides the “imaginary decay” which allowed someone to afk 3 months then jump back to top 10, as was seen when rewards were given for top 20 soloq? and a bunch of afkers just jumped to the top while people playing all week and earned a top 20 spot got deranked to 30+ on reward day lel.

100% agree with OP.

We have a whole year until Justin deems the current system a failure, according to one of his earlier posts. We have to endure and hope for the best.

The current Leaderboard doesn’t make sense.

You dun goofed, anet. Get rid of the farmboards, don’t try to devise your own clever little weighting system to reward casuals, use ELO. There’s a reason every other esport uses it.

-Someone who played 2.5k+ games with your previous MMR-based leaderboards and can’t find the motivation to queue up since the implementation of farmboards.

As someone who played extensively through the old leaderboard, I really think it was nearly perfect, except for the decay. I rose and fell in the old boards (in solo and team) and could observe that the difficulty of the matches was greater as I went higher up in the board, and the players ranked higher tended to play better and try harder (of course there were outliers, but it worked as well as I can imagine for this sort of thing).

-Currently at 5,259 ranked games played and prefer everything pre-December update (aside from the minor changes like match ready sound and light-up when tabbed out). Been PvPing in Guild Wars for the better part of 10 years, and I’m not sure my enthusiasm has ever been lower.

Saiyan.1704:

We have a whole year until Justin deems the current system a failure, according to one of his earlier posts. We have to endure and hope for the best.

I’ll give Justin credit. There’s been times where I felt matchmaking was nearly perfect where matches were close and challenging.

Though the point system itself really makes no sense for pvp. pvp should be about promoting competition and not who can generate the most points by volumes of game played.

+1

It’s like you took the words right out of my mouth Quilja.

+1
This current leaderboard makes no sense. Those who have 10 hrs to spam Q per day will claim the top spots despite their lackluster skill. Quality > quantity.

Quilja.9425:

Winning 5 matches in a row vs no names barely moved you compared to beating ppl in the top 25.

Had to make an extra reply for this. I was all for the new changes (making a sound for match timers <- #1 complaint in soloq 4v5s). The problems with the leaderboards now is that there is no differentiation from beating “good” and “bad” players . The system is meant to compensate the massive drop in rating which would happen in the old system if you are matched vs overwhelming odds (in this case premade) and nudge you to play harder. The system itself won’t work (imo), unranked should be scrapped for soloq and let regular queue account for duos, trios, quads? (the only problem is that anets system cannot mm teams using duo trio or quad which means you will wait 1 hr+ queue times, they went for the easiest solution, which is merge). If 3 cannot find 2, just 1+1, if 2 cannot find three just 2+1 or 1+1+1 OR if it’s really active 1+1+1+1+1 vs 1+1+1+1+, etc.

There needs to be a direct connection between a player’s MMR and their placing in the leaderboards. I suggest dividing the current leaderboard into different sections based on MMR. In this way you need a certain MMR to be ranked at a certain division of the leaderboard. This way we won’t have people with low MMR on top of the leaderboard.

Thoughts?

edit for clarification:
This means that player’s who play the most at their MMR-segment will be the highest on their MMR-segment.

Penguin.7906:

There needs to be a direct connection between a player’s MMR and their placing in the leaderboards. I suggest dividing the current leaderboard into different sections based on MMR. In this way you need a certain MMR to be ranked at a certain division of the leaderboard. This way we won’t have people with low MMR on top of the leaderboard.

Thoughts?

Indeed, this is the concept of a League division. But i’d add in every case a cap of number matches (for LB) or you have always the same problem in each league, who plays more, will win the division (in a scoreboard lb).

It’s like Soccer league, If Bayern M. plays 30 matches and Wolfsburg plays 80 matches, Wolfsburg will be the champion 100%

MarkPhilips.5169:

Penguin.7906:

There needs to be a direct connection between a player’s MMR and their placing in the leaderboards. I suggest dividing the current leaderboard into different sections based on MMR. In this way you need a certain MMR to be ranked at a certain division of the leaderboard. This way we won’t have people with low MMR on top of the leaderboard.

Thoughts?

Indeed, this is the concept of a League division. But i’d add in every case a cap of number matches (for LB) or you have always the same problem in each league, who plays more, will win the division (in a scoreboard lb).

It’s like Soccer league, If Bayern M. plays 30 matches and Wolfsburg plays 80 matches, Wolfsburg will be the champion 100%

Or you could retweak the points rewarded for wins and points lost on losses. You should always lose points on a defeat, how many could still be dictated by chance of victory.

Penguin.7906:

MarkPhilips.5169:

Penguin.7906:

There needs to be a direct connection between a player’s MMR and their placing in the leaderboards. I suggest dividing the current leaderboard into different sections based on MMR. In this way you need a certain MMR to be ranked at a certain division of the leaderboard. This way we won’t have people with low MMR on top of the leaderboard.

Thoughts?

Indeed, this is the concept of a League division. But i’d add in every case a cap of number matches (for LB) or you have always the same problem in each league, who plays more, will win the division (in a scoreboard lb).

It’s like Soccer league, If Bayern M. plays 30 matches and Wolfsburg plays 80 matches, Wolfsburg will be the champion 100%

Or you could retweak the points rewarded for wins and points lost on losses. You should always lose points on a defeat, how many could still be dictated by chance of victory.

This would go to advantage a lot full teams instead of solo/duo comp.

Theoretically, in the actual lb, there should be a system like this (according with the news) but honestly nothing seems to be changed and without any details it’s really difficult to understand.

I mean, of course a full 5 premade (if it’s not terrible) has a better win/ratio instead of people who go solo/duo/, if you want to keep an unique queue because spvp population is not so big, i think cap the number of matches is the more balanced solution (with a scoreboard system).

Indeed you can continue to play, but only X matches are valid for LB Season.

MarkPhilips.5169:

Indeed you can continue to play, but only X matches are valid for LB Season.

I don’t think capping valid matches is the answere. Encouraging playing more is a good thing, as hopefully it will lead to more players actually playing. If the system is poor though, you would just get the opposite effect.

Although my suggestion might favor premades, this will atleast be a good compromise between promoting playing and playing good.

Quilja.9425:

Should be Quality, not Quantity. Do we really need to prove who is “best” based on how many wins, or can we simply show who is best by having them beat those who sit at the top?

Agreed.

chibbi.3706:

Not really sure why they wouldn’t just use the elo system, like every single esport does. It’s so simple…

GW2 uses Glicko2 for match-making and for the leaderboards prior to December 2014. Glicko2 is an evolution of Elo.

Decay is not necessary, although it helps against high ranked people not playing games to not risk getting their elo lower.

I would argue that you do need a decay mechanic (though better than the old one). The problem with Elo-based systems is that when people at the top stop playing, it’s very hard for anyone to overtake them. Top players can play a lot in the first few weeks and then sit on their rank for the remainder of the season.

Just wanted to point out ANET, this is one of the few posts where everyone is in agreement. If that doesn’t count for something lol.

+1 agreed

If you are doing badly, the new MMR doesn’t match you with people who are good. It matches you with people who play bad as well. Before, it punishes those who play good by making them play with lower rank players, or players lower percentage of wins, to “help” them, to teach them, make them better. The only good thing it does is help farmers earn more points, since if you play badly, chances are, the old MMR will take “mercy” with you, and pair you with people who have a good record.