Update...
So you are correct about the win streak thing. I’ve argued how arena is flawed in many threads & actually showed examples of players using your examples. You can play tougher competition, have a better win % even playing tougher opponents, & still lose to someone who puts together a good streak at the end of their season and stops playing.remembermyn4me wrote: ↑16 July 2025, 10:36 The ranking system makes no sense because it rewards only and only winning streaks. It does not reward the win ratio. You could lose 300 games then win 20 in a row to be 1st with only 20 games won out of 320, totally ridiculous. Just like the opposite, win 300 games then lose a few after, and not be 1st with 300 wins out of 300 and some games. Statistically the chance of stringing together enough wins to move to 1st will inevitably come one day or another, so it's just a question of playing time, where the luckiest will just take less time than the others. For the ranking to be fair, the win ratio would have to be taken into account but also that points are lost and that the player goes down in the ranking when not playing. It's too easy to stabilize your score by not playing, and it's doable even with 1 account so that's not the issue. Stopping playing to keep your place is obviously not fair play, it's unacknowledged cheating. A season should be played from start to finish. But as long as the ranking system doesn't change, the puppets who believe they're better than the others by stopping playing for 2 months won't change. That's why they must be deprived of victory.
Overall win rate is a joke. It says more about how bad your opponents are then how good you are. Not to mention obvious issues with account history and general skill changes over time. Ideally, the win rates of all players should be roughly 50%. A change in skill level should not lead to a significant change in your win rate, but rather a change in the level of your opposition that eventually smooths out your temporary increase in success.remembermyn4me wrote: ↑16 July 2025, 10:36 The ranking system makes no sense because it rewards only and only winning streaks. It does not reward the win ratio. You could lose 300 games then win 20 in a row to be 1st with only 20 games won out of 320, totally ridiculous. Just like the opposite, win 300 games then lose a few after, and not be 1st with 300 wins out of 300 and some games. Statistically the chance of stringing together enough wins to move to 1st will inevitably come one day or another, so it's just a question of playing time, where the luckiest will just take less time than the others. For the ranking to be fair, the win ratio would have to be taken into account but also that points are lost and that the player goes down in the ranking when not playing. It's too easy to stabilize your score by not playing, and it's doable even with 1 account so that's not the issue. Stopping playing to keep your place is obviously not fair play, it's unacknowledged cheating. A season should be played from start to finish. But as long as the ranking system doesn't change, the puppets who believe they're better than the others by stopping playing for 2 months won't change. That's why they must be deprived of victory.
Saying win rate means nothing in Bang! is the real joke. Over thousands of games, luck and teammate/opponent quality balance out. Everyone plays with the same distribution. What’s left is how you played. That’s why overall win rate is the only meaningful indicator of long-term skill in this gameCeaseless wrote: ↑16 July 2025, 20:28Overall win rate is a joke. It says more about how bad your opponents are then how good you are. Not to mention obvious issues with account history and general skill changes over time. Ideally, the win rates of all players should be roughly 50%. A change in skill level should not lead to a significant change in your win rate, but rather a change in the level of your opposition that eventually smooths out your temporary increase in success.remembermyn4me wrote: ↑16 July 2025, 10:36 The ranking system makes no sense because it rewards only and only winning streaks. It does not reward the win ratio. You could lose 300 games then win 20 in a row to be 1st with only 20 games won out of 320, totally ridiculous. Just like the opposite, win 300 games then lose a few after, and not be 1st with 300 wins out of 300 and some games. Statistically the chance of stringing together enough wins to move to 1st will inevitably come one day or another, so it's just a question of playing time, where the luckiest will just take less time than the others. For the ranking to be fair, the win ratio would have to be taken into account but also that points are lost and that the player goes down in the ranking when not playing. It's too easy to stabilize your score by not playing, and it's doable even with 1 account so that's not the issue. Stopping playing to keep your place is obviously not fair play, it's unacknowledged cheating. A season should be played from start to finish. But as long as the ranking system doesn't change, the puppets who believe they're better than the others by stopping playing for 2 months won't change. That's why they must be deprived of victory.
BGA populations don't generally support this very well, so you end up with unbalanced pairings more often. While I like elo for measuring skill in general, I think it gets less accurate about the degree of difference as the rating gap expands, underestimating the gap in games with low variance and overestimating it in games with high variance.
Regardless, Bang isn't really a great game for making sense of skill. It's high luck, hidden information, requires more than 2 players in a format where they interact with each other's resources. It's an impressive number of ways to maximize reducing the impact skill has on the outcome. You're not going to find the "best" player in Bang because skill isn't the most relevant factor in the outcome.
In the many games, especially ones where you could actually say skill is a real factor in the first place, player quality is higher for higher rated players due to pairings. BGA's not the greatest when it comes to this, but generally higher rated players get paired with higher rated opposition, even more so if you use the rating filters, which I generally do. Granted, looking at Bang, the rating ceiling is so low that might not even exist. It's bad enough that I've detected cheaters just by checking for accounts rated over 300 to see if it was some lucky streak or if they were boosting using multiple accounts. Even the best players in Bang frequently get ratings so low they're getting close to hitting 100, where BGA stops you from going any lower. So Bang at the very least has lessened the problems of that particular factor of win percentage. Though if your comments are about the rating system beyond just Bang, which I've assumed them to be, this is an issue you'd have to confront.slav4ella wrote: ↑17 July 2025, 23:20Saying win rate means nothing in Bang! is the real joke. Over thousands of games, luck and teammate/opponent quality balance out. Everyone plays with the same distribution. What’s left is how you played. That’s why overall win rate is the only meaningful indicator of long-term skill in this gameCeaseless wrote: ↑16 July 2025, 20:28Overall win rate is a joke. It says more about how bad your opponents are then how good you are. Not to mention obvious issues with account history and general skill changes over time. Ideally, the win rates of all players should be roughly 50%. A change in skill level should not lead to a significant change in your win rate, but rather a change in the level of your opposition that eventually smooths out your temporary increase in success.remembermyn4me wrote: ↑16 July 2025, 10:36 The ranking system makes no sense because it rewards only and only winning streaks. It does not reward the win ratio. You could lose 300 games then win 20 in a row to be 1st with only 20 games won out of 320, totally ridiculous. Just like the opposite, win 300 games then lose a few after, and not be 1st with 300 wins out of 300 and some games. Statistically the chance of stringing together enough wins to move to 1st will inevitably come one day or another, so it's just a question of playing time, where the luckiest will just take less time than the others. For the ranking to be fair, the win ratio would have to be taken into account but also that points are lost and that the player goes down in the ranking when not playing. It's too easy to stabilize your score by not playing, and it's doable even with 1 account so that's not the issue. Stopping playing to keep your place is obviously not fair play, it's unacknowledged cheating. A season should be played from start to finish. But as long as the ranking system doesn't change, the puppets who believe they're better than the others by stopping playing for 2 months won't change. That's why they must be deprived of victory.
BGA populations don't generally support this very well, so you end up with unbalanced pairings more often. While I like elo for measuring skill in general, I think it gets less accurate about the degree of difference as the rating gap expands, underestimating the gap in games with low variance and overestimating it in games with high variance.
Regardless, Bang isn't really a great game for making sense of skill. It's high luck, hidden information, requires more than 2 players in a format where they interact with each other's resources. It's an impressive number of ways to maximize reducing the impact skill has on the outcome. You're not going to find the "best" player in Bang because skill isn't the most relevant factor in the outcome.
First of all, you're not even talking about win rate. On BGA, what you're referring to is not a classic win rate — it's what's called game success, and it's a strange metric.Ceaseless wrote: ↑18 July 2025, 02:23In the many games, especially ones where you could actually say skill is a real factor in the first place, player quality is higher for higher rated players due to pairings. BGA's not the greatest when it comes to this, but generally higher rated players get paired with higher rated opposition, even more so if you use the rating filters, which I generally do. Granted, looking at Bang, the rating ceiling is so low that might not even exist. It's bad enough that I've detected cheaters just by checking for accounts rated over 300 to see if it was some lucky streak or if they were boosting using multiple accounts. Even the best players in Bang frequently get ratings so low they're getting close to hitting 100, where BGA stops you from going any lower. So Bang at the very least has lessened the problems of that particular factor of win percentage. Though if your comments are about the rating system beyond just Bang, which I've assumed them to be, this is an issue you'd have to confront.slav4ella wrote: ↑17 July 2025, 23:20Saying win rate means nothing in Bang! is the real joke. Over thousands of games, luck and teammate/opponent quality balance out. Everyone plays with the same distribution. What’s left is how you played. That’s why overall win rate is the only meaningful indicator of long-term skill in this gameCeaseless wrote: ↑16 July 2025, 20:28
Overall win rate is a joke. It says more about how bad your opponents are then how good you are. Not to mention obvious issues with account history and general skill changes over time. Ideally, the win rates of all players should be roughly 50%. A change in skill level should not lead to a significant change in your win rate, but rather a change in the level of your opposition that eventually smooths out your temporary increase in success.
BGA populations don't generally support this very well, so you end up with unbalanced pairings more often. While I like elo for measuring skill in general, I think it gets less accurate about the degree of difference as the rating gap expands, underestimating the gap in games with low variance and overestimating it in games with high variance.
Regardless, Bang isn't really a great game for making sense of skill. It's high luck, hidden information, requires more than 2 players in a format where they interact with each other's resources. It's an impressive number of ways to maximize reducing the impact skill has on the outcome. You're not going to find the "best" player in Bang because skill isn't the most relevant factor in the outcome.
That said, the problem with overall win percentage is that it lessens the impact of improvement, even just looking at Bang. It doesn't measure how well you're currently playing, but instead uses data from your full history, which may include periods where you weren't playing at the level you presently are. Many players would be rated better simply by discarding their old account and starting over so that their previous performance no longer weighs them down.
On another separate note, win percentages in Bang are... something. Since it's more than 2 player, the win rates go over 100%, I'm not actually sure what you'd even call a perfect percentage in Bang. Not only do players regularly go over 2 players and throw off the standard 100%, but they aren't even always the same, the basic range being from 4-7. I haven't looked much beyond the base game to see if they have even more variance. Some players will play any player count, though others might be picky and stick with a specific number. In addition to messing with the percentages, each player count is a different game and changes the player balance. It makes it chaos to compare players in general, with a mean curveball to percentages on top of that. How relevant is each measured percent in Bang? 1% in 100% is likely worth more than the 1% of 200% (in which case that "1%" is .5% of the total 200%). What's the percentage gap where it's not mostly chance?
Trying to get all of that sounds like a lot of work, not to mention the natural variance in what losing the abandoned games is. And all of that for an approximation that still has issues. That the elo system is so messed up for Bang shows us that skill isn't really a big factor in Bang compared to other games, and not worth putting that much work into trying to sort out. Sure, there's skill in the game, with some decisions being better than others, but the degree of effect they have on the game's outcome is pretty low when compared to other games.slav4ella wrote: ↑18 July 2025, 15:06First of all, you're not even talking about win rate. On BGA, what you're referring to is not a classic win rate — it's what's called game success, and it's a strange metric.Ceaseless wrote: ↑18 July 2025, 02:23In the many games, especially ones where you could actually say skill is a real factor in the first place, player quality is higher for higher rated players due to pairings. BGA's not the greatest when it comes to this, but generally higher rated players get paired with higher rated opposition, even more so if you use the rating filters, which I generally do. Granted, looking at Bang, the rating ceiling is so low that might not even exist. It's bad enough that I've detected cheaters just by checking for accounts rated over 300 to see if it was some lucky streak or if they were boosting using multiple accounts. Even the best players in Bang frequently get ratings so low they're getting close to hitting 100, where BGA stops you from going any lower. So Bang at the very least has lessened the problems of that particular factor of win percentage. Though if your comments are about the rating system beyond just Bang, which I've assumed them to be, this is an issue you'd have to confront.slav4ella wrote: ↑17 July 2025, 23:20
Saying win rate means nothing in Bang! is the real joke. Over thousands of games, luck and teammate/opponent quality balance out. Everyone plays with the same distribution. What’s left is how you played. That’s why overall win rate is the only meaningful indicator of long-term skill in this game
That said, the problem with overall win percentage is that it lessens the impact of improvement, even just looking at Bang. It doesn't measure how well you're currently playing, but instead uses data from your full history, which may include periods where you weren't playing at the level you presently are. Many players would be rated better simply by discarding their old account and starting over so that their previous performance no longer weighs them down.
On another separate note, win percentages in Bang are... something. Since it's more than 2 player, the win rates go over 100%, I'm not actually sure what you'd even call a perfect percentage in Bang. Not only do players regularly go over 2 players and throw off the standard 100%, but they aren't even always the same, the basic range being from 4-7. I haven't looked much beyond the base game to see if they have even more variance. Some players will play any player count, though others might be picky and stick with a specific number. In addition to messing with the percentages, each player count is a different game and changes the player balance. It makes it chaos to compare players in general, with a mean curveball to percentages on top of that. How relevant is each measured percent in Bang? 1% in 100% is likely worth more than the 1% of 200% (in which case that "1%" is .5% of the total 200%). What's the percentage gap where it's not mostly chance?
By win rate, I mean the number of wins divided by the number of games played.
But I also adjust for abandoned games.
Roughly 10% of Bang games are abandoned — and that inflates win rates by about 5%.
I usually lower this extra percentage to 4%, assuming that a small portion of those abandoned games were our own
So to get a real win rate in Bang, you have to calculate it manually, not rely on game success shown on BGA.
I agree that even this corrected win rate doesn’t reflect your current level — but it does represent your overall level across the full distance you’ve played.
It shows a trend. You won’t see a clear progression, but if I see a player with a true, clean win rate of 46–48%, or even close to 50% (which is almost impossible in Bang), I immediately know I’m looking at a very strong player.
At the very least, they know how to play their roles properly.
Because on a long enough timeline — even if you were learning during the first 1000 games, and then played 4000 more with solid understanding — if you’ve maintained that win rate across the total sample, I can draw conclusions about your skill.
We have players here with 10,000+ games and a clean win rate of 42% — what does that tell you? That they’ve stayed average throughout those 10,000 games.
As for ELO, which has a supposed ceiling of 300 — it’s completely meaningless in Bang.
That number can be inflated by a single 10-game win streak, and lost again by a 3-game losing streak.
There’s really nothing to analyze in that rating — it’s not representative of skill at all.
