Cribbage ARENA should be best of three

[Gamesterz]

Why in heavens name did they change it to just a best of one game. it needs to be best of three

[Quizmaster]

I agree. I was disappointed when I saw it had changed. While a best of three rewards skill, a single game can flip on just one hand of play. Hopefully they will review this for the next season; until then, we just have to roll with it.

[markpie]

Just returned to playing, was surprised and disappointed there was only one game. Hope they change it back to three next season.

[Stu_1977_SEmelb]

Agreed.

[weinsist]

The odds of winning as a pone in best of one is quite exactly the same as the odds of winning as the first pone in best of three. So, yeah, best of one makes a lot more sense as more games played reduces the variance.

So, perhaps somewhat paradoxically, best of three encourages gambling over actual strategy.

If not convinced yet, some math with reasonable simplifications for clarity:

Assume we have two identical players with dealer victory distribution N( 0.6 ; 0.4 ). The number of dealer's wins in Bo3 is then distributed with roughly N( 1.6 ; 0.7). For a comparison, number of victories per player in two games, once pone once dealer, is distributed with N( 1 ; 0.56). The best way to beat the statistical insignificance of actual skill in Bo3 is to build volatile hands and peg recklessly. Thus gambling is encouraged.

If not happy with Bo1, I'd suggest two games, once as a pone once as a dealer with a draw as a possible outcome. This way the luck factor would be minimised and gambling strongly discouraged.

ETA: I've just realised that that loser becomes the dealer. Ah, it's even worse than I thought. So in the current Bo3 the odds of winning are in favour of players having lower winning rates. Which is insane. It literally makes the match outcome almost a coin toss, and encourages even more rampant volatility.

[Parritae]

Here an example how "great" the current Bo3 works:

Three games, Player 1 starts as a dealer:

Hands total
Player 1: 136 points
Player 2: 165 points
Cribs total:
Player 1: 50 points
Player 2: 71 points
Pegging total:
Player 1: 63 points
Player 2: 64 points

Guess who won.

[Thor]

The odds of winning as a pone in best of one is quite exactly the same as the odds of winning as the first pone in best of three. So, yeah, best of one makes a lot more sense as more games played reduces the variance.

So, perhaps somewhat paradoxically, best of three encourages gambling over actual strategy.

If not convinced yet, some math with reasonable simplifications for clarity:

Assume we have two identical players with dealer victory distribution N( 0.6 ; 0.4 ). The number of dealer's wins in Bo3 is then distributed with roughly N( 1.6 ; 0.7). For a comparison, number of victories per player in two games, once pone once dealer, is distributed with N( 1 ; 0.56). The best way to beat the statistical insignificance of actual skill in Bo3 is to build volatile hands and peg recklessly. Thus gambling is encouraged.

If not happy with Bo1, I'd suggest two games, once as a pone once as a dealer with a draw as a possible outcome. This way the luck factor would be minimised and gambling strongly discouraged.

ETA: I've just realised that that loser becomes the dealer. Ah, it's even worse than I thought. So in the current Bo3 the odds of winning are in favour of players having lower winning rates. Which is insane. It literally makes the match outcome almost a coin toss, and encourages even more rampant volatility.

This analysis is very flawed. If the players are the exact same caliber of course in a game of 90% chance it will be decided most of the time by who’s dealing. However the whole point of tournament play is to ascertain who is better, not only match two equal opponents. Over the course of multiple games, skill and odds calculation will show who is better. Will it always show? Heavens no. Will it show up in 3 games more often than in just one game, absolutely.