BETA: Arena mode (Season 0 - BETA)

[sourisdudesert]

Will it stil be possible to use "Advanced mode" and ignore Arena Mode.

Sure! Automatic lobby will be mandatory for Arena mode, but for a regular game you'll have the choice.

[sprockitz]

I'm concerned that a large percentage of these games would never get to the highest levels because they aren't popular enough.

Take jump gate for example, it has been around for 3 years since the game was introduced and with 23 wins I'm in the top 20 for wins. 3 years is 12 seasons, so I'm averaging 2 wins per season...and that puts me in the top 20 for wins! This means a top 20 active player wouldn't even get out of the 1st level! Let alone have any chance of getting close to level 6.

And based on the competition page that orders games by popularity, there are 40 games behind jump gate!

So there is no way this will work in ~75% of games...

To have any chance of success there needs to be a different number of levels based on the popularity of a game. Probably 3 levels for the bottom half of games, then could go in tiers up to 4, 5, and 6 based on increasing popularity. I'd likely go further and say a different season length (6 mos. for less popular games). And of course prestige payouts should be different since the amount of competition is so different...winning 'world champion' in a less popular (3 level) game, should only be worth a small fraction of an immensely popular (6 level) game.

Now if you don't relegate or only relegate 1 level, less popular games could eventually get to level 6...but it would take many seasons. Depending how you do EAS, just resetting that would be more than enough to force players to play a lot in order to have a chance to win...I mean I've played many games 50+ times and my Elo in those games hasn't stabilized yet. I could say relegate 1 level then rely on EAS taking enough games to give you a chance would be more than enough for each season.

[sourisdudesert]

So there is no way this will work in ~75% of games...

This is difficult to say without trying, so we're going to try, learn and adjust.

For example: we may reduce the number of leagues for less popular games.

However:
_ situation is better that you described (we expect to have nice and challenging ompetitions for more than 25% of the games).
_ for less popular games, there are no real good solutions since there is not enough players to have a real "ladder style" competition.

[sprockitz]

I am looking forward to it and hope I don't come across too negative, I'm just trying to point out potential issues beforehand. I hope it works for more than 25%, that was just a guesstimate. I think less levels would be a good compromise for less popular games.

I'm most excited for the games that are better with more than 2 players where comparing Elo between people who play 4p and people who play 2p isn't fair b/c 4 player is going to have a much higher variance and be more difficult to get a top rank in the current system. Particularly excited for Terra Mystica and Clans of Caledonia...hope they'll both be 4 player

[Ahla]

I think this idea is very interesting and well worth a try. There are a few problems I see though:

If you play turn based, 3 months is really short. You need to play a very high number of games at the same time to get a chance to climb enough, and timing the end of these will be an important "skill". You want to finish a bunch of winning games in a row, just to promote yourself, and then finish a lot of losing games now when you can't be relegated, to improve the percentage of winning games you have going etc. Also, stalling to get a game outside the 3 month wall to avoid a loss is a great tactic. And if one or more players in the game have this tactic, games starting one month before the season end will not finish in time. You will basically have to start a ton of games first 30-45 days of a season to give yourself a chance.

As others have mentioned, some games are probably not popular enough to gain enough players for this to work. Maybe have a rotating schedule between somewhat similar games, or just hold a long tournament to decide the champion there, using some sort of Swiss-like pairing system.

[Curling]

Spring is coming.

Will you be ready for spring 2019 as first season ?

[eoc]

For example: we may reduce the number of leagues for less popular games.

Starting at the top, I hope, so that all games would have an Elite league but not all games would have Bronze (or lower) leagues? Going in the reverse direction as mentioned would presumably discourage players from choosing the less popular game even more than usual (harder to find opponents in the first place), but having an easier way to the top (because you'd start in a higher league than usual) could actually become a motivation to try some of them and increase popularity!

I would also like to second the points Ahla raised about turn-based tables and timing them – when exactly is the point difference calculated? Between players at the start of the table, or when it's ending? The FAQ wording and replies here seem to indicate the former, since otherwise if a table spans across seasons, you might very well end up finishing a game with league difference of larger than 2, which apparently isn't supposed to happen.

I'm not sure of the exact implications, but I feel like introducing an additional third mode of play isn't the best solution to what you're trying to achieve. Were other mechanisms discussed, such as (for example) opt-in or opt-out from your profile and just counting all non-training games for leagues? This could allow special handling of the case where one of the players likes to compete while others don't, which in the proposed system basically is a table that won't ever exist and potentially reduces my number of possible matchups in 3+ player games by quite a bit – if one subset of possible opponents / teammates prefers league play and another set doesn't. For 2-player games it's less of a problem of course

[sourisdudesert]

Starting at the top, I hope, so that all games would have an Elite league but not all games would have Bronze (or lower) leagues?

Good point. Yes, it would be better to do it this way. Thanks for the idea.

turn-based tables and timing them – when exactly is the point difference calculated?

It will be calculated at the end of the game, like it is for ELO right now. I agree this is not ideal, but it brings other problems if it is done the other way.

To be honest, we don't know if Arena mode is well suited for turn based (and the opposite). It may be well suited, it may not. We'll see. What is sure is that in the future we'll work on better tournaments and THAT will be a major improvement for turn based competitive play on BGA.

I feel like introducing an additional third mode of play isn't the best solution to what you're trying to achieve

Yes, we discuss MANY possibilities. However, Arena mode is very specific (automatic matchmaking is mandatory, options and number of players are the same for everyone), so this is quite difficult to mix this mode with other modes. This is true that this may reduces the matchups, but this specific problem is a priority for us and will introduce a solution to avoid this.

[xxerox]

I suppose you are planning on removing ELO and the whole point system and change it for Arena?

I was thinking, if you they gonna do that, arent you gonna split the gamebase even more?

And if not.. What makes you think some players arent competitive because they want high ELO and arena mode won't change anything?

So basically -

ELO removed: Competitive players will stop playing normal games and games will have less players and will fill less.

ELO remains: Competitive players will still play competitive in normal games despite Arena mode which makes it no different than how it is now.

[distortion]

relegation is needed.