Luck vs Skill in Arena

[Vandaley]

I prefer arena formats that are more fun rather than have minimum luck, so for example sushi Go with 4 players not 2

When the argument is applied to games that are designed for kids, or just generally designed for fun it is the most frustrating for me. I really love the game Happy City, it wen to 2 player and the main reasoning as that there was too much variability in multiplayer. So I still play it, but I much prefer playing 3 and 4 player.

[jonob]

I prefer arena formats that are more fun rather than have minimum luck, so for example sushi Go with 4 players not 2

When the argument is applied to games that are designed for kids, or just generally designed for fun it is the most frustrating for me. I really love the game Happy City, it wen to 2 player and the main reasoning as that there was too much variability in multiplayer. So I still play it, but I much prefer playing 3 and 4 player.

I completely agree. Happy City is probably the most frustrating example of this trend. In addition, as mentioned in other threads, the players who play 2 player exclusively get more games and are a higher percentage of gurus, who then get to pick the next arena format.

I proposed a rotation for Happy City for next season (I'm currently 2nd in arena but still prefer 3 or 4 player). Hawaii and Fleet are both doing 2-3-2-4 rotations which seems like a great balance between the players who like 2 player and the players who want to try other formats. If you're a guru go vote for it!

[Lumin_S]

If you want to come up with a better ELO calculation, or a better system, then let's hear it.

TrueSkill has been around for 10+ years. There's now a TrueSkill2 also. https://en.wikipedia.org/wiki/TrueSkill

The ratings created by Juho Snellman for *another website* are also excellent: https://www.snellman.net/blog/archive/2 ... yer-games/

*moderator edit: please remember that it is against our terms of service to post about other gaming websites

edit: sorry mod, should have remembered that one and worded it better. apologies!

[RicardoRix]

Not really sure if this is going off on a tangent with regard to the original topic, but yeah, you maybe onto something in general, that looks pretty cool from a casual quick read. Maybe ELO is not quite up to the task for multiple player games.
I'd thought that previously (around Arena conception time) BGA did make some changes with the ELO calculation and it got some review from phd types like Sprockitz that looked over it.

[msclark]

I think Arena is great - better opposition (usually), fewer abandonments, consistency of settings, easier to start games - but if you think it's some sort of World Championship, carefully designed to determine the strongest players, then I'm afraid you're mistaken. The season winners are usually pretty good, to be sure, but it massively rewards volume over skill and, regardless of any game's luck factor, there's a lot of luck in who you get matched up with and what your current run of wins is like when the season ends (or you park your score). The pool of players is largely bifurcated between real-time and turn-based without a huge amount of crossover, meaning that some contenders will never play each other, and there are similar issues with time zones and preferred speed settings. And though it's a minor thing that I doubt many people abuse, the whole red-thumb thing really bugs me - it should not be possible to filter out your potential opponents.

In short, Arena really isn't a serious competition that in any way adequately determines who the best players are, so we should stop treating it as such. It's just a fun casual competition that's a welcome step up from the chaos of simple matchups. And that's fine by me.

That doesn't mean that I wouldn't love to see another level, though. A genuine season-long tournament that more satisfactorily gets to the best players. A fixed number of games with all-play-all, perhaps, or an extended knockout structure. But there are difficulties in organising such a thing, and it's probably not possible without it being officially supported within the software. Or maybe cheating will render it meaningless.

[RicardoRix]

That doesn't mean that I wouldn't love to see another level, though. A genuine season-long tournament that more satisfactorily gets to the best players. A fixed number of games with all-play-all, perhaps, or an extended knockout structure.

Yes, a huge swiss based qualification round, and then phase 2 where everyone is separated into leagues (of 8 players?) with a round-robin all v all. The top league taking all the prizes, and everyone getting to play the same level of opponents
There would need to be turn-based and real-time variants.
Previously, there have been some very large tournaments organised that have been excellent.

There was mention that there is a big announcement on tournaments from Sourisdudesert in his youtube appearance, but nothing has materialised. I was kinda hoping for something like this.

[dschingis27]

I'd thought that previously (around Arena conception time) BGA did make some changes with the ELO calculation and it got some review from phd types like Sprockitz that looked over it.

Yes, there were comments and ideas by Sprockitz. No, BGA never made changes. The scoring for Arena is the same as of day 0 of Arena and EAS calculation is basically equal to Elo.

[Vandaley]

I think Arena is great - better opposition (usually), fewer abandonments, consistency of settings, easier to start games - but if you think it's some sort of World Championship, carefully designed to determine the strongest players, then I'm afraid you're mistaken. The season winners are usually pretty good, to be sure, but it massively rewards volume over skill and, regardless of any game's luck factor, there's a lot of luck in who you get matched up with and what your current run of wins is like when the season ends (or you park your score). The pool of players is largely bifurcated between real-time and turn-based without a huge amount of crossover, meaning that some contenders will never play each other, and there are similar issues with time zones and preferred speed settings. And though it's a minor thing that I doubt many people abuse, the whole red-thumb thing really bugs me - it should not be possible to filter out your potential opponents.

In short, Arena really isn't a serious competition that in any way adequately determines who the best players are, so we should stop treating it as such. It's just a fun casual competition that's a welcome step up from the chaos of simple matchups. And that's fine by me.

That doesn't mean that I wouldn't love to see another level, though. A genuine season-long tournament that more satisfactorily gets to the best players. A fixed number of games with all-play-all, perhaps, or an extended knockout structure. But there are difficulties in organising such a thing, and it's probably not possible without it being officially supported within the software. Or maybe cheating will render it meaningless.

Really great contribution to the thread I think.

There are some games that have attempted tournaments that do this with multi stages. I did a 2021 Thurn and Taxis Championship where you qualified by placing in a tournament, placing in Arena in any season, or having an Expert+ ELO. I invited all that met those criteria and stages a multi stage tournament with differing games (2 vs 3 player for example). It was a huge hit, but was a lot of work. I know Welcome To also did this as I participated in it. If you are wondering User mikim was the T&T champ in our tournament.

[sprockitz]

I'd thought that previously (around Arena conception time) BGA did make some changes with the ELO calculation and it got some review from phd types like Sprockitz that looked over it.

Yes, there were comments and ideas by Sprockitz. No, BGA never made changes. The scoring for Arena is the same as of day 0 of Arena and EAS calculation is basically equal to Elo.

Yes I had a lot of suggestions, but none of them were ever considered (to my knowledge) and certainly nothing was implemented.

A strong tournament structure with big official tournaments is one way to go, and as of 2 years ago that was the plan but I haven't seen any progress on that front besides a few updates to tournaments back in 2020. If this ever gets implemented it would provide nice opportunities.

Turn-based lends itself well to tournaments, but solving the issue of dropouts in multi-player games is key (few options here...disallowing players to kick (then on time-outs all players with positive share win, all players out of time get loss), auto-kicking for games that support playing with a missing player and then keeping results from game instead of 1-1-1-0 default). Still time settings too fast will prevent some from playing, too slow and large events would take most of (or more than) a year to declare a winner. The 2 stage ability is there, so not a ton of work to automate this (N stages and multi-player simultaneous games would make things more effective)

Real-Time has potential but would require commitment from players, you can run qualifiers, 2-3 stages of it if needed for very large fields, and qualifiers can vary time of day when they take place. But final round players would need to commit to a certain time in advance...which is doable but also means some regions of the world would have to play in the middle of the night. This would need some work to automate.

[snorkel83]

I think this is on topic, but a little off the side of where the discussion is heading.

I play casually, but like other have written here, Arena is good for matching up with player who not just kidding around, stop playing etc.

Anyway could it be a possibility in the most popular games, like Carcassonne, Azul, Splendor etc, to have 2 seasons going on the same time. One head to head that are optimize skill over luck and one multiplayer?