Page 1 of 1
Is ELO rating/matchmaking well-suited for this game?
Posted: 28 May 2021, 12:42
by Farsen
I love 7 Wonders duel, and I´ve played 200+ games on BGA. But I find the ELO rating/matchmaking to be quite bad as a skill-rating for this game. It seems that when you a good base understanding of the game, and knows which cards to prioritize (especially in the wonders selection phase), the luck factor rises significantly.
Lets sayt hat a 400 elo player gets matched agains a 100 elo player which has a good understanding of the game. The 100 elo player is going first and ends up with 3 double turns wonders. Then the 400 elo player (with 1 double turn wonder) is gonna have a hard time winning if the card layouts favors both players equally.
This is fine and is the nature of this kind of game. But the implemented Elo rating system is giving the 400 elo player a high elo point reduction if he looses, and a very small elo point increase if he wins. A high risk/low reward game - and a low risk/high reward for the 100 elo player.
This would not be a problem if the 400 elo player had the option to setup a game with a more well-suited matchmaking. But as it stands today, he can only hope for a similar ranked player. Alternatively he can create a live game, and then cancel/refuse to play with lower ranked players when the game is about to start.
I guess that this is a general "issue" with games who has a medium+ luck-factor? Or am I wrong in this matter?
Re: Is ELO rating/matchmaking well-suited for this game?
Posted: 28 May 2021, 13:49
by MoiMagnus
Farsen wrote: ↑28 May 2021, 12:42
I guess that this is a general "issue" with games who has a medium+ luck-factor? Or am I wrong in this matter?
I'm not sure I've found a single game on BGA where ELO was satisfying.
Between
(a) cooperative games where ELO make no sense and encourage toxic behaviours,
(b) team games where where ELO is allegedly even more broken (though I've not followed the details on that point)
(c) high player count games where tables are usually quite diverse in ELO rank, and past a certain point you lose ELO for second place, while there is no realistic way a to ensure first place especially if the other players play against you (which is reasonable if you are the strongest player at the table),
(d) games with so much variety of options/expansions/player count that it is basically 10 games in one hence the ELO is meaningless for anybody who frequently change such options.
(e) games where a significant part of the community use programs or browser expansions to "help" them to play, while another part doesn't
(f) games that are full of underrated players [either because they mostly play outside of BGA, or that they have multiple accounts on BGA]
I'd say 7W Duel is one of the games on BGA where ELO is the
least broken. And that's after fully agreeing with your remarks on how it is not well-suited for 7W Duel either.
Re: Is ELO rating/matchmaking well-suited for this game?
Posted: 28 May 2021, 14:45
by steve5332
ELO is perfect for two player games. It doesn't matter what luck there is in a game, it works. Don't forget that your ELO is your average over time, so the odd speed bump doesn't matter.
I'd say that 7WD has medium-low amounts of luck, but that's just my opinion. I see that you have played far more games than I have so perhaps your views are more accurate!
You do realise that you can create tables and restrict access by ELO? There's no need to boot players, just set your minimum ELO rating when you create the table if you want to avoid lower ranked players. Hope that helps!
Re: Is ELO rating/matchmaking well-suited for this game?
Posted: 28 May 2021, 16:05
by Wonderful Plays
MoiMagnus wrote: ↑28 May 2021, 13:49
Farsen wrote: ↑28 May 2021, 12:42
I guess that this is a general "issue" with games who has a medium+ luck-factor? Or am I wrong in this matter?
I'm not sure I've found a single game on BGA where ELO was satisfying.
Between
(a) cooperative games where ELO make no sense and encourage toxic behaviours,
(b) team games where where ELO is allegedly even more broken (though I've not followed the details on that point)
(c) high player count games where tables are usually quite diverse in ELO rank, and past a certain point you lose ELO for second place, while there is no realistic way a to ensure first place especially if the other players play against you (which is reasonable if you are the strongest player at the table),
(d) games with so much variety of options/expansions/player count that it is basically 10 games in one hence the ELO is meaningless for anybody who frequently change such options.
(e) games where a significant part of the community use programs or browser expansions to "help" them to play, while another part doesn't
(f) games that are full of underrated players [either because they mostly play outside of BGA, or that they have multiple accounts on BGA]
I'd say 7W Duel is one of the games on BGA where ELO is the
least broken. And that's after fully agreeing with your remarks on how it is not well-suited for 7W Duel either.
I would agree with this one. But 7WD for me has a higher luck factor than I would like for a 2P game.
Re: Is ELO rating/matchmaking well-suited for this game?
Posted: 28 May 2021, 23:26
by Le007n
In my experience, the majority of 100-200 players in this game play at this level and its much less often that luck really comes into it over the course of the game. I think elo works pretty well in this instance as an indicator of a players skill but there is definitely room for improvement on multi player games
A good player playing 10 games against a low to mid player is probably going to drop one, possibly two based on the nature of the game
There are occasional game where you are three to one down on go again wonders. This can be hard but by no means impossible to turn around. Often you have to look for ways to force them to use these wonders and find the best route from there
There are also games where a single card can change the flow of the game. If you are not on the right side of this it can be very frustrating especially when it takes four or five more games to get that elo back
That said, for each game you are on the wrong side of the draw there are probably other where you are on the right side. The elo comes in based on how well you can turn around the harder draws vs the possibly easy games I think which evens out over time
It is only the odd game where I think I was lucky or my opponent was lucky. Restrict elo to better players and you get a slightly more even drop and increase and learn more playing better players. I've watched a few of your videos wonderful plays so I know you like to watch the top players as I do
Once pantheon is released this stops some of the scripted feel you can get by adding more options to benefit from card turns or prevent turning and we'll be a great addition (though often the forcing players down a path is not by chance

)
To summarise, like most card games, if you get the wrong deal it can feel one sided but I think 7wd has a lower amount of luck than a lot of games once you dig into it, understand the options available, how to navigate the cards and the probabilities involved. If you look at the top players they tend to stay around the same level with some deviation but ultimately, a steady elo level
Re: Is ELO rating/matchmaking well-suited for this game?
Posted: 31 May 2021, 10:27
by Farsen
Thanks alot for your feedback.
I haven´t played that much on BGA, but I can understand that ELO is a pretty bad measurement when it comes to most of the games

Guess it´s pretty hard to find a solution that fits all game.
steve5332 wrote: ↑28 May 2021, 14:45
You do realise that you can create tables and restrict access by ELO? There's no need to boot players, just set your minimum ELO rating when you create the table if you want to avoid lower ranked players. Hope that helps!
Doh!! I´ve started alot of games, and never seen the "Restrict table access" link in the right top corner! Guess I have been focusing on the game options section to the left. Thanks!
Le007n wrote: ↑28 May 2021, 23:26
In my experience, the majority of 100-200 players in this game play at this level and its much less often that luck really comes into it over the course of the game. I think elo works pretty well in this instance as an indicator of a players skill but there is definitely room for improvement on multi player games
A good player playing 10 games against a low to mid player is probably going to drop one, possibly two based on the nature of the game
There are occasional game where you are three to one down on go again wonders. This can be hard but by no means impossible to turn around. Often you have to look for ways to force them to use these wonders and find the best route from there
There are also games where a single card can change the flow of the game. If you are not on the right side of this it can be very frustrating especially when it takes four or five more games to get that elo back
That said, for each game you are on the wrong side of the draw there are probably other where you are on the right side. The elo comes in based on how well you can turn around the harder draws vs the possibly easy games I think which evens out over time
It is only the odd game where I think I was lucky or my opponent was lucky. Restrict elo to better players and you get a slightly more even drop and increase and learn more playing better players. I've watched a few of your videos wonderful plays so I know you like to watch the top players as I do
Once pantheon is released this stops some of the scripted feel you can get by adding more options to benefit from card turns or prevent turning and we'll be a great addition (though often the forcing players down a path is not by chance

)
To summarise, like most card games, if you get the wrong deal it can feel one sided but I think 7wd has a lower amount of luck than a lot of games once you dig into it, understand the options available, how to navigate the cards and the probabilities involved. If you look at the top players they tend to stay around the same level with some deviation but ultimately, a steady elo level
Thanks for your insight. I mostly agree with you. I´ve just had many games where the win could be on either side, and where it came down to a single certain card being flipped. It´s not common for me to see mistakes by the 0-200 elo players.