🌿ELO System Updates

[juggernaughts11]

Yeah, it seems that BGA still doesn't understand the biggest issue with their power-leveling gimmick. It distorts (penalizes) the ELO of people who are strong and active, in pretty much every game, except for those with a massive player base. This new tweak where the lower ranked player won't lose ELO if playing a stronger player that they've lost to several times fixes the obvious issue of overall ELO decay for both players for being active, but of course it opens up a new incentive for lower ranked players to try to match with strong players with zero risk. If they lose, they lose 0 ELO, if they win, they get 100% of the ELO (so they are playing with free money after 5 losses), and it's the opposite for the strong players - they can only lose. This could be justifiable if you are playing the same player with no other players in between (to cut down on point farming from a secondary account in a single session), but the fact that you can play 10 other players in between and still get this penalty is just terrible (especially in arena, with unknown opponents).

Here's an idea - if they really want to try to indicate potential cheaters, for anyone in the top 10 of a game, they could indicate what percentage of their games are against their most frequent opponent. When compared with others in the list, that could give an indication if that person is farming from a secondary account, especially if it's their more recent games that are against that account.

At any rate, the new ELO rules make ELO significantly worse at knowing a person's true strength, which should be the principal goal of a rating system... Otherwise, we have to come back to assume that the true goal is what the new system actually does - discourages active players from playing to reduce server load.

[Stroom]

Here's an idea - if they really want to try to indicate potential cheaters, for anyone in the top 10 of a game, they could indicate what percentage of their games are against their most frequent opponent. When compared with others in the list, that could give an indication if that person is farming from a secondary account, especially if it's their more recent games that are against that account.

Then you just create 10-20 different users to farm with. Any kind of system will be abusable. The general problem with BGA developers is that they always fall into the sunk cost fallacy. They have tried to make changes so for them the only ay to go forward is to force ANY kind of change, rather than going back to the old way. Sometimes the old way is better and reverting is the best way to go forward.

[KenFin1]

Then you just create 10-20 different users to farm with. Any kind of system will be abusable. The general problem with BGA developers is that they always fall into the sunk cost fallacy. They have tried to make changes so for them the only ay to go forward is to force ANY kind of change, rather than going back to the old way. Sometimes the old way is better and reverting is the best way to go forward.

Agree on all points.

[ChiefPointThief]

1. Avoiding ELO Loss for Consecutive Losses: [/b]
Your ELO is a measure of your skill on a game against the WHOLE community, and you should not receive any benefits from repeatedly playing against the same player(s).

This has been explained to bga many times in great detail. Example. I am a caravan player. I am #1 in caravan and there are only around 30 players in the community for arena. I'm an active player who can play triple digit amount of games in an arena season. Meaning that I can literally play the WHOLE community multiple times and I would be penalized simply for being good.

I met a player this season and they told me, "What is going on with the whole elo thing is terrible. I completely understand if you eventually have to block me". The account wasn't even recognizable to me so I didn't understand. Then I realized I had 3 games with this person and they probably assumed they were going to lose because of my ranking. Is this a thing you want on your site? People being blocked to avoid your terrible rating system.

3. Red Thumbs Change for next Arena Season: [/b]

It appears you went back on this change due to the testimony of people like agricola players. Is it possible to add an option for games to be continued when there is a quitter with the quitter coming in last? That way 2 hour games of agricola and others that can continue don't go to waste?

[ChiefPointThief]

In what way are you considering it a good thing for the elo penalty for repeatedly playing the same opponent? I've got to assume the BGA team sees it as a win because it reduces reports you have to review of people abusing the rating system, and so from an admins daily perspective, it may be a win. But from the point of view of the user, the game player, it's still a total mess. You're penalizing strong players for winning a lot, you penalize players in games with small player pools, you penalize players for playing more often, and you disincentive players from making friends and repeatedly playing the same opponent(s) (even with many games played in between).

I am guilty of playing the "find a good excuse for bga" game as well. Partly because I was in denial. Honestly we all need to stop doing this. There appears to be a target on strong players for whatever reason right now and if there is some sort of reason like the one you stated then I agree with h_illes. Bga could and should just be transparent and speak for themselves instead of us giving them outs.

[6element]

BTW, changing how thread views are counted is truly pathetic.

[Janetta]

BTW, changing how thread views are counted is truly pathetic.

6element! What do you mean?
Has BGA changed how thread views are counted?
If yes . . . that is indeed disturbing!
Tell us more . . . please!

[6element]

BTW, changing how thread views are counted is truly pathetic.

6element! What do you mean?
Has BGA changed how thread views are counted?
If yes . . . that is indeed disturbing!
Tell us more . . . please!

In January this thread was getting ~50K daily views, by the end of February it naturally declined to ~20K daily views or so - there was no response for 2 months, and sufficient amount of feedback has been provided, so people stopped following it. This is how it got to 2.5M views in 90 days.

Last 4 days, after they told us what they told us we've "got" ~4K views. In four days. So, something changed? I assume they only count one view per user per day or something.

[torvaldur_makan]

Upcoming Changes:

1. Avoiding ELO Loss for Consecutive Losses:
First, the introduction of an ELO reduction for consecutive wins is a positive change, from our point of view and analysis in the long run– it has become more difficult to abuse the system. Your ELO is a measure of your skill on a game against the whole community, and you should not receive any benefits from repeatedly playing against the same player(s).

However, we recognize that losing consecutive games to the same opponent can be frustrating and can lead to an unfair ELO drop. To address this, we will be implementing a mechanism to prevent ELO loss for consecutive losses against the same opponent, like how consecutive wins are currently handled. This will help maintain a more balanced ELO distribution.

What???

I don't think anyone is arguing that it's unfair to lose Elo when you lose consecutive times against the same player.

What we're saying is that it is unfair that you stop gaining points when you win consecutive times against the same player. Since player pools are small (if you consider time of day, game mode etc.) you don't really have a choice but to play the same players many times. And if you are one of the strongest, you will win often, and get penalized.

[miklomike]

Many good points already, but i would like to add that this is not a sufficient response, especially after so long. Even the suggested k factor lowering doesn't seem right to me, if you win more your elo goes up, and so the difference between you and the loser grows, thus making you win less ELO the next time. This already happens, I don't see a need to lower it artificially even more.

Not reverting the ELO changes, but added ways for people who are bad at the game to not be able to lose is a step in the wrong direction. This is seemingly the opposite result of what you mention, that ELO is only for better match making. If the ELO isn't reflecting skill anymore then it's not good for match making.

Why have achievements in the site at all if you don't want us to strive for things? You must see that adding these goals increases user play time and drives more site interaction. So why does it seem like you want people to engage with the site and other players less often? Personally I have played games more than I would have without achievement and ELO because I like trying to reach a certain point. Without these things I would probably stop playing a lot of games after the first or second game. This can't be a good thing for your game publishers, who are hoping that people will play their game more, start to like it, and purchase it. I have definitely purchased games I played here, and only started liking the game when trying to go for higher ELO.

You said we should just try to enjoy playing and getting better, but how can I know if I'm improving if my ELO is hidden for co-op games? Please, listen to your players. It is ok to admit something was not the best idea. There is way too much sink cost fallacy thinking at BGA and a mentality of "bga has never been wrong, we must only move forward". As a kindergarten teacher said, "mistakes + correction = learning".