Player rankings , what is needed for each level ? Apprentice , Average, Good Expert etc.

[FSKFSK]

My profile for this game says "39 times played, 28 victories, 36% wins". That makes no sense at all. Isn't that a 71% winrate? Is it counting wins as ties? (71 divided by 2 is approximately 36)?

[tarongot]

My profile for this game says "39 times played, 28 victories, 36% wins". That makes no sense at all. Isn't that a 71% winrate? Is it counting wins as ties? (71 divided by 2 is approximately 36)?

That topic might interest you https://boardgamearena.com/forum/viewto ... 15#p256815.

[FSKFSK]

I already knew about that. It's possible to get a winrate over 100% for multiplayer games if you win more than double your share of games, because winrate is adjusted by player count. They have to do that to make winrate meaningful for games with a variable number of players.

For this game, there should be no adjustment and it should be straight winrate. Actually, winning games with more players is harder, but it doesn't seem to adjust for that.

[tarongot]

I already knew about that. It's possible to get a winrate over 100% for multiplayer games if you win more than double your share of games, because winrate is adjusted by player count. They have to do that to make winrate meaningful for games with a variable number of players.

For this game, there should be no adjustment and it should be straight winrate. Actually, winning games with more players is harder, but it doesn't seem to adjust for that.

In linked topic is my post about coop games. Winrate in those seems to be divided by half for no reason, except one that BGA hates coop games and treat them as something worse.

[FSKFSK]

It's probably half for coop game because the algorithm is scoring it as a tie (half win). I think you also get a half win for regular games if it's a tie. They need a coded exception for coop games to just display the flat winrate.

[FluffyWuffy1]

I moved from Average to Good after a winning game. My stats right after that game were:

- 125 games
- 79 wins

The next game I played was a loss. I went back to Average with stats, pretty obviously, of:

- 126 games
- 79 wins

I then won the next, and moved back to Good again. So this at least confirms you do lose elo for a loss.

[FSKFSK]

Yes, I confirmed you can lose elo in "The Gang". I played only 1 game and lost and my elo chart went down.

It takes forever to get over 100 elo, and your elo won't go down until you reach 100, which is a rule that applies to all BGA games.

It also looks like there's a defect in the elo formula, where it's easier to gain elo in 3-4 player games. I assume that's the reason there are so many 3-4 player games. I'd rather play at 6 instead of playing advanced mode and having to deal with blackout.

[PseudoPserious]

I'm not sure if I like being able to lose elo for a cooperative game. I enjoy playing with new players and helping them learn the game, and I guess I'm now doing that at my own detriment? Obviously, the fake number on the board game site doesn't really matter, but it seems like a perverse incentive to me.

[tarongot]

I'm not sure if I like being able to lose elo for a cooperative game. I enjoy playing with new players and helping them learn the game, and I guess I'm now doing that at my own detriment? Obviously, the fake number on the board game site doesn't really matter, but it seems like a perverse incentive to me.

I'm sure I don't like cooperative games not having ranking at all. No way to determine on what level my team mates are. How well he/she know game. Can he/she handle harder games variant? Or is just waste of time. Before I can gather team of players I believe can handle harder challenges I have to play with them a lot. All that info ranking could give me.
And you can't have ranking without losing points for lost games.

If you want teach new players you can always play no ranking games, or yes accept playing on your own detriment.

[PseudoPserious]

Eh, it seems like it introduces "table selection" as a component of the "skill" of playing the game which, while very true to its poker roots, doesn't tell you much about if they're a good player at the table or not.

If you contend it's a more accurate measure of a player's skill, I think my response would be that it can't be all that much better than a point per win, and whatever miniscule increase in accuracy you get is outweighed by the negative impact on the playing community overall.

But that's just my subjective opinion; obviously I don't have any data to back that up.

----

Coincidentally, we played together this morning and won, so kudos to us.