Elo adjustment

[Polymorphism]

Hi everyone, would like to hear your opinion on the best way to score ELO in "The way to Juliet".

The goal is to make the ELO system for the game as good as possible to represent the next ranks:
ELO ranking allows to classify players in 7 distinct skills levels: Beginner (0 point), Apprentice (1-99 points), Average (100-199), Good (200-299), Strong (300-499), Expert (500-699), Master (700+).
More info about Elo: https://boardgamearena.com/faq?anchor=faq_account_elo

What should be considered?
- In BGA co-op games the elo does not decrease, only grow up (at least what I hear from similar topics).

The current ELO system:
Lose - 0 points
Regular mode - 10 point
Advance mode (100 turns) - 15 points
Advance mode (75 turns) - 20 points
Advance mode (50 turns) - 25 points

Which looks like:
Regular: 10 wins * 6 min = 1 hour (average game time) victory experience for the next Elo rank (100 Elo points)
Advance (50 turns): 4 wins * 6 min (average game time) = 24 min victory experience for the next Elo rank (100 Elo points)

But this calculation is a bit wrong as the average game time for won will be a bit higher as you can lose even in one move (most likely it will be 8-10 min).

Which basis for Regular mode in your opinion will better fit, 15 points (7 wins to the next rank), 20 points (5 wins to the next rank), or another?
And are 5 point steps for advanced mode good or better use other values (or even a score system)

[miklomike]

Hopefully others will weigh in as well. But I think I like the 20 or even 25 for regular, since the player count is a little low it will help differentiate between a couple games won and 10-15 games won.

As for the advanced mode I think 5 is too little, as someone will need to be good enough at regular games before they can win at advanced a lot. With that in mind I'm ok if people who can play advanced have a higher ELO, as it will signal that either they have played a LOT of regular games, or a good amount of advanced games. Both of these will give a good indication of a player's experience. As you mentioned it's a co-op game, ELO is just a showing of experience in the game, not skill vs others. But this is what we have to visually see quickly to get a sense of our partner.

Thanks for the game and engagement with player base.

[ufm]

I think current values are OK, considering the win ratio.

[Polymorphism]

I think the root issue is the different player experience in real-time and turn base modes.

The current ELO scoring system a designed for real-time mode. For example I can play 3 game in a row and score 30 points in 15 minutes in regular mode. The game required many turns (in average 53), but they are really quick usually 5-7 seconds which create feeling of fast game.

On the other side, for turn-based mode is more correct to base player experience on the average turns number rather than the average game duration. In general, if each player makes one turn in a day the game will take an average of 27 days, and getting back in the context of the game takes more time than making a turn. Plus as the game has some technical moves - moving around a column, the feeling of the game is more longer than it actually is.

So we need a proper solution to better cover player experience for both modes.

So far I see two options:

• Make a middle ground between these modes
• Or, create a separate score system for each

I will ask more players to join our discussion, so we will find the best solution for all.

P.S. Also think there should be a bigger gap for Advance mode with 50 turns, as there was only 3 games in this mode and not sure if any successful

[ufm]

Actually it doesn't matter as multiple turn-based games can be started and then be played at the same time.

[miklomike]

You aren't going to be able to keep up with too many turn based games at once and remember what is happening. If you have to delve into the play log just to make your next move it's taking much longer to play.

Plus it's easy to play 2 or 3 games a day over the course of a month. I'm not going to play 60 to 90 turn based games at once over that same period.

I'm in favor of higher ELO for lower player base games. Let's see what other people who play the game say.

[ufm]

You aren't also going to be able to play real-time games as you expect due to the matchmaking issues.
If the player pool is small, even getting a single real-time matchmaking becomes really hard.
Turn-based games suffer less from such things obviously.
As a matter of fact, I've never seen turn-based players arguing they deserve more rating compared to real-time players in co-op yet.

Also, the player pool size has little (or nothing imo) to do with the amount of rating gained.
Needlessly inflating it just deteriorates player level distinction. You should consider the rating does not drop on losses.

Plus it's easy to play 2 or 3 games a day over the course of a month. I'm not going to play 60 to 90 turn based games at once over that same period.

You haven't seen some co-op guys playing plenty of tables then.
Playing >20 tables at the same time is definitely possible and it actually happens.
Playing even >5 is relatively common as well (assuming you can find enough players).

[Polymorphism]

I agree that we can not give extra scores just based on mode as it will create possibilities to abuse the system by playing real-time games in turn-based mode. Which can make it even harder to find a partner for real-time mode. Definitely not an option.

So probably our core discussion is whether should we a bit increase ELO, for example, make it 15 points to regular mode, to make it a bit less tough for players that played a game a month. The second option is to give some point bonus if the game is played at least for example 10 days. I don't think a few points will motivate someone to hold the game for 10 days on purpose.

I have no time to ask other players' opinions but will try to do it this week.

P.S. If somebody prefers not to express their opinion publicly, just write it to me and it will stay anonymous.

[ufm]

So probably our core discussion is whether should we a bit increase ELO, for example, make it 15 points to regular mode, to make it a bit less tough for players that played a game a month. The second option is to give some point bonus if the game is played at least for example 10 days. I don't think a few points will motivate someone to hold the game for 10 days on purpose.

As the second option incentivizes a mode over another, I'd rather opt to the fairer first option.
Slightly reducing the bar from 10 wins to 8 wins won't do much I guess.

[SimplicatusGames]

So probably our core discussion is whether should we a bit increase ELO, for example, make it 15 points to regular mode, to make it a bit less tough for players that played a game a month.

The large differences in ELO "rewards" between co-op games is frustrating, and I completely understand that it is difficult for a new BGA game developer (you are among the few that have designed, illustrated and implemented your own game on BGA, well done!) to decide what model to choose.

Having said (read: written) that, I was surprised that I got as much as 10 ELO for a single win. My personal opinion is that this is too much for such a short game. Another co-op game that also requires many small moves (so it is slow on turn-based mode, but quick to play in real-time) is Bandido. It also involves laying tiles and can have a similar number of moves. However, sometimes you can be lucky and solve Bandido very fast, so on average it is perhaps a faster game than The way to Juliet.

Bandido awards 1 ELO per win, regardless of how long the game takes. That seemed unreasonable to me at first, but there are other games that take longer to play that also give just 1 ELO. On the other hand, there are games that give quite a lot of ELO. A couple of other co-op games I know well are None Shall Pass and Solar Storm. The former gives 1-5 ELO per game, depending on how well you play, and the latter gives from 0 to 10 ELO (more when you have less than 100 ELO), depending on the difficulty level, but also on the number of games played - Bandido and None Shall Pass do not follow this common adjustment factor for ELO.

The benefit of using this factor is that new players get more ELO and therefore are motivated to play more, while those who play hundreds of times don't end up with many thousand ELO, as I've seen for some co-op games. In Solar Storm it took me 6 months to go from 200 ELO to 700 ELO at a rate of 2-3 ELO per game (on the most difficult setting, the easy setting awarded zero ELO), and each game took at 20-30 minutes. A game of None Shall Pass also takes 20-30 minutes, often more with less experienced players, and then it also might not yield as much as 5 ELO, as that requires practice and optimal play.

So for this game, I would suggest two options (if you want to change it again):

A) Lower the ELO earned to e.g. 1 or 2 per game on basic settings, and 4 or 5 on the most difficult setting (minimum number of moves on advanced)

B) Use the current 10 ELO for normal and perhaps 15 for advanced (not sure if you already do that today as I haven't tried yet), but include the adjustment factor based on number of games played, so the ELO earned ends up being around 2-3 per game on advanced setting and zero on basic setting after a significant number of games. This works really well for Solar Storm IMHO.

Not sure if B is a lot more work to implement than A, even if it seems more complicated. In any case, I think both of these are better than leaving it the way it is now. People who enjoy a co-op game here on BGA will play it hundreds of times, and if the ELO awarded per game is too high, they all end up with 8700 ELO or whatever the system max is (that example is from Regicide, where it is easier to win with 4 players but you get 48 ELO for such a win, while you only get 24 ELO for winning with 2 players and 36 ELO with 3 players). It's hard to win Regicide so it is very nice to get a lot of ELO when you manage to win, but once you get good at it, you will win quite regularly and then you "easily" get a very high ELO.

This game is not hard to win on the basic setting once you figure out how to do it, but the advanced setting with a limited number of moves is very interesting, and I look forward to exploring that more. Therefore, for me personally, option B above is the best choice, since an experienced player (who has at least 100-200 ELO) will win every time and therefore should be awarded zero ELO for it. This encourages experienced players to play the advanced mode, which is more interesting once you have "figured out" the game. Compare this with the more complex co-op game Hanabi where you depend on some luck with the cards and a lot of communication and coordination in order to win, especially on advanced setting. There you also get zero ELO for a win on the easy setting.

This is a family game, and casual players can be discouraged if they don't get any ELO for winning (unless they are so casual that they don't care about ELO and might even play in friendly/training mode). This is another argument for option B, where you get more ELO to begin with so you can reach 100 ELO quite fast, and then the gains slow down. However, as mentioned I don't know if this is a lot of work for you.

PS: I don't think awarding more ELO is going to attract more players. People play a game because they like it, and they won't recommend a game to others because they were easily able to get 1000 ELO, but because they enjoyed it. IMHO, for co-op games where you don't need to reach 100 ELO in order to participate in tournaments or Arena, ELO is only useful for seeing how many times somebody has played a game, and when the scales vary wildly between games, it is difficult to know at first. I prefer a lower amount per game because I'm among those that have played at least one co-op game over a hundred times.