Arena mode is a disgrace

[hibachi]

I also call into question to accuracy of the statement that the "top" players today can only win 60% of the time.

To be clear, I'm talking the top players on BGA, most of which are playing Arena. If we setup different game formats that helped de-emphasize risk, such as longer matches, you'd definitely see the best players with higher win-rates because it would help minimize luck.

It's not difficult to confirm this - feel free to go through the list of the highest rated ELO players and top list of the current season and see their win rates. This is exactly what I did.

[tbhp]

No it's actually very commonly 3x in Arena Points. I played the same player back to back last night...lost one (-32 points), and won one (+11 points). For what it's worth I'm not saying he's a bad player...but his ELO was low 100's when we played.

I agree about the 3x. I play an even higher variance card game, and if I were to randomly play arena, I would struggle to win often enough to offset the losses. For this game, the arena winner is determined by who gets lucky at the end of the season.

The points gained and lost depend on Elo. If you lose 3x as much as you gain, it means you are playing against opponents you should beat 75% of the time.

Of course, it doesn't mean the win probabilities are actually correct. You or your opponent may not currently be at an Elo that reflects your true level (which is all the more likely if the game is very luck based and therefore the Elo very volatile). But mathematically it works out. The problem is not that the math is wrong, it's the uncertainty of the game.

[FrankJones]

To be clear, I'm talking the top players on BGA, most of which are playing Arena. If we setup different game formats that helped de-emphasize risk, such as longer matches, you'd definitely see the best players with higher win-rates because it would help minimize luck.

It's not difficult to confirm this - feel free to go through the list of the highest rated ELO players and top list of the current season and see their win rates. This is exactly what I did.

I agree - Gin Rummy is ideally played as a "best of 11" or something like that. Or at least best of 3. Or anything in between. On BGA is it just a single game? That's a terrible way to play. Of course the results are going to be all over the place.

[FrankJones]

The points gained and lost depend on Elo. If you lose 3x as much as you gain, it means you are playing against opponents you should beat 75% of the time.

Of course, it doesn't mean the win probabilities are actually correct. You or your opponent may not currently be at an Elo that reflects your true level (which is all the more likely if the game is very luck based and therefore the Elo very volatile). But mathematically it works out. The problem is not that the math is wrong, it's the uncertainty of the game.

Exactly. Using Open face Chinese poker as an example, I have had plenty of encounters against players who were 180 Elo blow me but were only 80 Elo below me in terms of actual skill. They were just at the bottom end of a bad run of luck, and the nature of the game on BGA (Playing a ludicrously small sample of 6 hands) is such that many players will fluctuate between 140 and 260 Elo, or something like that. As a result ,the Elo ceiling is in the low 300s, and there are capable players in the 100s.

[hibachi]

The points gained and lost depend on Elo. If you lose 3x as much as you gain, it means you are playing against opponents you should beat 75% of the time.

Of course, it doesn't mean the win probabilities are actually correct. You or your opponent may not currently be at an Elo that reflects your true level (which is all the more likely if the game is very luck based and therefore the Elo very volatile). But mathematically it works out. The problem is not that the math is wrong, it's the uncertainty of the game.

Exactly. Using Open fac Chinese poker as an example, I have had plenty of encounters against players who were 180 Elo blow me but were only 80 Elo below me in terms of actual skill. They were just at the bottom end of a bad run of luck, and the nature of the game on BGA (Playing a ludicrously small sample of 6 hands) is such that many players will fluctuate between 140 and 260 Elo, or something like that. As a result ,the Elo ceiling is in the low 300s, and there are capable players in the 100s.

This is EXACTLY what is going on with Gin. Leading to insane volatility. My entire point is it kind of ruins the fun/skill of it, because it artificially inflates luck vs. reality. I doubt it's really 'fun' for anyone to deal with the swings, ultimately discouraging people from even playing. Certainly if someone gets to #1 they aren't going to keep playing, which is indicative of a bad structure.

I looked for a thread that addressed that concern (instead of creating new) and bumped it hoping a solution would be considered. The thread title is terrible however...so that's a negative.

[dschingis27]

If the worst player really can win 40% of the time against the best player, then no system is likely to ensure that the best players win an arena season.

How did Stu Ungar manage to win almost 100% of the time? (At gin Rummy.)

He didn't win anywhere close to 100% lol, unless he was playing total fish every time that were clueless.

He famously beat the top known player 86 games to 0.

Idk I'd say there is something fishy about this story, at least sth we are not told. I mean the point is even if you fully knew your opponents hand after only 3 rounds, that doesn't give you a win rate of 100%, I mean the luck of the card draw is still a thing you can't control.

There are a number of games where it's just theoretically impossible to have a win rate >70% against a pool of reasonable opponents.

[FrankJones]

He didn't win anywhere close to 100% lol, unless he was playing total fish every time that were clueless.

He famously beat the top known player 86 games to 0.

Idk I'd say there is something fishy about this story, at least sth we are not told. I mean the point is even if you fully knew your opponents hand after only 3 rounds, that doesn't give you a win rate of 100%, I mean the luck of the card draw is still a thing you can't control.

There are a number of games where it's just theoretically impossible to have a win rate >70% against a pool of reasonable opponents.

It's likely always going to be impossible to distinguish myth from fact with regard to Stu Ungar. But if the stories are even half true, it's still amazing.

[dschingis27]

I did a little research and the "sth we are not told" about this story is that they played Hollywood Gin which has a very peculiar scoring system. (You can ask AI if you want to know how it works.) Stu did not really win 86-0 single games of Gin Rummy (particularly not in a way how it would work on BGA), he scored 86-0 in the Hollywood Gin scoring system in a marathon of Gin Rummy. Knowing that it's "only" 86-0 in the Hollywood Gin scoring system makes the story 10000% more plausible, even though it is still absurdely astounding that he pulled that of.

And there still seems to be a lot of scepticism and no or only hard to find real verifiable evidence of the story or the existence of the player Stein in the first place. https://www.reddit.com/r/nonmurdermyste ... y_existed/

[FrankJones]

Yes, still impressive.

Regarding Gun Rummy here: I see a game that was recently played by the arena leader was scored 355 to 210. In a match to 100.

What's with the "game bonus" and "line bonus"? I'm not familiar with these, I don't think.