Guild Wars 2

Hello guys, Poliator here.

We are all frustrated with the new Leaderboard. We think it’s bad, it’s a farm, et cetera… and I hope it changes. (We are still getting Test Seasons, so changes are always possible) But this post isn’t to whine about it. I want to ask you, players, what is your “ideal” of a Leaderboard? What should it mean to you?

In my opinion, a Leaderboard has three different purposes:

• Showcase the most skilled players at the top of the leaderboard, so the better players get recognition and, maybe even more important, top teams can scout out the better players for their team if they have a vacant place.

• To give the player progression. I don’t like at all the old MMR-based Leaderboard because we didn’t have any kind of metric to see progression. We’d just see our rank go up and others’ go down. I want to see some kind of number that indicates me how far I am from X player or what I need to do to get to any rank.

• Also, this “metric” rewards consistency; playing the game often. For me, it doesn’t matter if a player is the very bestest and wonnered against TCG (Twitch chat ftw, I’m sorry). If a player is good, he has to demonstrate he is good every day. It’s not fair that a player, just because his MMR is high, sits on the top of the Ladder, without even bothering to play more than one match per week. There should be a balance: play often to maintain your high position vs. raw skill.

• To distribute rewards among the competitive players. This can be understood better in a “league-based” Leaderboard. At the end of the season, the highest league will get the best rewards and in that higher league there will be few selected people (think of LoL’s challenger). This way, people have to only aim at a certain interval of ranks (or a certain league) to get better rewards: “ohh, shiny! I need to get to rank 800 at least to get X rewards!” It gives the player an incentive to improve.

My preferred type of Leaderboard would be one based on points and Leagues (easier aggregation of players), like League of Legends has, but I’m not necessarily against going back to an MMR-based Ladder, as long as it accomplishes what I think a Leaderboard should mean. I’d like to add too that I’m hopeful we will get leagues with this point system in a fancy, ArenaNet-style way.

TL;DR: In my opinion, a Leaderboard should have these purposes:

- Showcase the most skilled players.
- Give the players progression.
- Reward both skill and consistency (playing often).
- Be a way to distribute rewards.

What is your ideal of Leaderboard, guys? Do you prefer MMR-based or a League system? Discuss! I’d love to see everyone’s ideas.

We know you are Poliator. It is written in red next to your comment. Duh.

I actually like the idea of having leagues. I suggested that once long time ago.

But I also liked the MMR Leaderboard.. Could we have both?

I wonder..

Terrorsquad.2349:

I actually like the idea of having leagues. I suggested that once long time ago.

But I also liked the MMR Leaderboard.. Could we have both?

I wonder..

I would love a league system. However, I fear the player base is a little too small for that. Maybe after HoT?

Sorel.4870:

Terrorsquad.2349:

I actually like the idea of having leagues. I suggested that once long time ago.

But I also liked the MMR Leaderboard.. Could we have both?

I wonder..

I would love a league system. However, I fear the player base is a little too small for that. Maybe after HoT?

Anet knows exactly how many people pvp, couldn’t they just scale the size of these leagues to the size of the player base? Just make each one a percentage of the player base. I think of Anet did divide up the player base in that fashion they probably would actually end up having more players pvping because the matches would be more fair and fun.

The old MMR-based leaderboard mostly worked, but there were a few problems with it. All were fixable.

1. Initial rating could reach the top too quickly. We saw many 11 game players reach the top 100, which shouldn’t be happening. This is likely an issue with the initial rating deviation parameter in Glicko/Glicko2 being set too high for how it’s used in GW2.
2. Unclear rating gain/loss. It’s still not clear how rating increases and decreases are calculated for an individual on a team. Glicko/Glicko2 only describe calculations for fixed units (inviduals or fixed team rosters), so it’s unclear how players of disparate ratings on your team are affecting your rating. Is the clueless guy causing your rating to go down, or is the rating calculation accounting for that? Is it bad to play with your PvP-challenged friends from time-to-time?
3. Phantom decay didn’t work. The players at the top could sit on their rating and play one game at the end of the season to erase all phantom decay. That single game wasn’t enough for Glicko/Glicko2’s increase in rating deviation to kick in. When rewards are attached to a leaderboard, this becomes a relevant problem. For general use, the phantom decay caused the rating of players to fluctuate by hundreds of ranks over a couple days well into the season simply because other players were slipping off the boards or erasing their phantom decay. It seemed too volatile.

Leagues
Forum comments on leagues are often express them as a hard division of players. It should be thought of and is often implemented as a masked division of players based on percentile of a player’s rating. If you’re in the middle of a block, you’ll almost always be paired with or against people in the same block. If you’re near the start or end of a block, you’ll see a lot of mixing between the two bordering blocks. Because a league is a mask of rating percentile, changing one’s mask could have some kind of rating buffer. Having a league label, especially if it’s in-game, is a way for players to better compare against each other.

Decay
A major problem with Elo/Glicko/Glicko2 in video games is that there is no incentive to play constantly. Players can power to the top of a leaderboard in the first few weeks while ratings are less stable and then sit on their rating. Because a large amount of players above a certain rating threshold don’t play often (or at all), it becomes difficult for the rest of the player base to break into that upper echelon later in the season. Finding a balancing point between decay forcing players to play often, but not being overly burdensome is difficult.