Flawed Author-Cardplayer Articles

[maurile]

[ QUOTE ]
Ciaffone wrote: "17-to-1 against" and "9-to-1 against"

These are odds. Not chances.

The correct answers are 16:1 (which means 16 to 1 against), and 7.5:1.

See the math in my Ciaffone's Error post.

[/ QUOTE ]
The correct answers are about 16.01-1 against and 7.77-1 against -- or about 1/17.01 and 1/8.77.

So 1/17 and 1/9 are decent approximations.

Writing 17-1 and 9-1 instead of 1/17 and 1/9 is more like a spelling mistake than a reasoning error, in my opinion. It's a sign of carelessness, not incomprehension.

[colgin]

[ QUOTE ]
But he does use two flawed arguments to try to make his points. In the ATs case he says that he would mke the raise in late position and goes on to show a scenario where the late position raise helps you more than it does in the blinds. Granted. So what? How does that imply that the early position raise isn't still profitable?

[/ QUOTE ]

David,

This is what jumped out at me and I was a bit stunned. I just stoppped reading that article at that point.

[skp]

Good grief - you are reading way too much into the Ciaffone error.

I agree with you on Jones

I agree with you on Jim's ATs analysis.

But you are being a bit unfair on the A4s analysis. Brier does allude to the free card aspect albeit not in any great detail. On the other hand, Ed's answer in the book does not say (at least not clearly) that the main reason for the raise is to get a free card as you now state in your post. In fact, Ed's answer (which is set out in Jim's article) specifically refers first to buying two outs and secondly, to getting a free card.

Ed does say that 'unless you are very unlikely to get a free card, you should probably invest the extra bet and reraise" But surely that is not an indication that "making your ace out good was a secondary consideration".

[ QUOTE ]
The play is wrong if your opponents are less than about 50-50 to give you that free card and we say it is "probably" the right move in our book.

[/ QUOTE ]

Well, saying don't reraise if you are "very" unlikely to get a free card is not the same as the 50/50 comment made above. And the word "probably" is right at the end of the explanation in the book. It's easy to miss it. What is easy to see however is "Book Answer: Reraise." which indicates a definitive play.

[binions]

[ QUOTE ]
[ QUOTE ]
Ciaffone wrote: "17-to-1 against" and "9-to-1 against"

These are odds. Not chances.

The correct answers are 16:1 (which means 16 to 1 against), and 7.5:1.

See the math in my Ciaffone's Error post.

[/ QUOTE ]
The correct answers are about 16.01-1 against and 7.77-1 against -- or about 1/17.01 and 1/8.77-1.

So 1/17 and 1/9 are decent approximations.

Writing 17-1 and 9-1 instead of 1/17 and 1/9 is more like a spelling mistake than a reasoning error, in my opinion. It's a sign of carelessness, not incomprehension.

[/ QUOTE ]

Ciaffone did not write 17-1 or 9-1. He wrote

"The odds of running into a pocket pair in the big blind are about 17-to-1 against. The odds on either one of the blinds having a pocket pair are about 9-to-1 against"

[maurile]

[ QUOTE ]
5.878% x 2 players = 11.756% chance 2 players might have at least one pair between them = 1 chance in 8.5 that either player will have pocket pair = 7.5:1 against.

[/ QUOTE ]
This mistake is worse than Ciaffone's since it is conceptual.

(If each player has a 5.8776% chance of holding a pair, the chance that at least one holds a pair is not 5.8776% * 2. It is 1 - (1 - 5.8776%)^2 = 11.4096%, or about 7.76-1 against -- which is still an oversimplification because the third player has a slightly worse than 5.8776% chance of holding a pair given that the first two players don't. So the true answer is more like 7.77-1 against.)

[Rudbaeck]

[ QUOTE ]
Good grief - you are reading way too much into the Ciaffone error.

[/ QUOTE ]

No, he's not. Coming from an academic background these mistakes really grate on me, especially when coming from experts who make big bucks on selling their wisdom. When you get paid to teach you better make damn sure you're meterial is impeccable.

Though I'd happily argue that the fault isn't so much Ciaffone's as the editor of Cardplayer. This should never have made it past the editor desk without being corrected.

The two other flaws are worse, for sure. More peer review is needed, and I think both DS speaking out on it here as well as Brier's articles are a big step in the right direction. Brier is atleast willing to discuss things in detail. I sure hope more 'academic peer review' finds itself into poker, instead of the penis measuring contest that so frequently takes place.

[Kellon]

[ QUOTE ]
Coming from an academic background these mistakes really grate on me, especially when coming from experts who make big bucks on selling their wisdom. When you get paid to teach you better make damn sure you're meterial [sic] is impeccable.

Though I'd happily argue that the fault isn't so much Ciaffone's as the editor of Cardplayer. This should never have made it past the editor desk without being corrected.

[/ QUOTE ]

Better fire your editor. [img]/images/graemlins/blush.gif[/img]

[Rudbaeck]

[ QUOTE ]
Better fire your editor. [img]/images/graemlins/blush.gif[/img]

[/ QUOTE ]

I don't recall being paid for that post, but when the cheque arrives I'll talk to my editor.

[Stork]

[ QUOTE ]
The fact of the matter is that not raising here is a terrible play. It seems so obvious to me that it astounds me that Brier would think differently.

[/ QUOTE ]

The Jury has spoken [img]/images/graemlins/shocked.gif[/img].

[binions]

[ QUOTE ]
[ QUOTE ]
5.878% x 2 players = 11.756% chance 2 players might have at least one pair between them = 1 chance in 8.5 that either player will have pocket pair = 7.5:1 against.

[/ QUOTE ]
This mistake is worse than Ciaffone's since it is conceptual.

(If each player has a 5.8776% chance of holding a pair, the chance that at least one holds a pair is not 5.8776% * 2. It is 1 - (1 - 5.8776%)^2 = 11.4096%, or about 7.76-1 against -- which is still an oversimplification because the third player has a slightly worse than 5.8776% chance of holding a pair given that the first two players don't. So the true answer is more like 7.77-1 against.)

[/ QUOTE ]

You are introducing more information into the equation.

From the point of the button, which is Ciaffone's assumption, there are 2 known cards and 50 unknown. The two known cards are not paired.

Leaving 72 possible pairs left out of 1225 possible starting hands.

In the button's mind, the SB and BB EACH have an independent 72/1225 chance of having a pair.

Now, if you want to introduce more info, like what the SB has and how that influences what the BB might have, then the calculation is much more detailed.

For example, if the SB has a pocket pair, is it one of the ranks shared by the button's unpaired hand? If so, there are 69 possible pairs left out of 1128 possible hands or 6.12%. If the SB has a pocket pair of a rank different than the button's unpaired hand, then there are 67 possible pairs of 1128 possible hands or 5.94%.

There are 6 possible pairs that the SB could have that share a rank with the button, and 66 that don't.

((6*69 + 66*67)/72))/1128 = 5.954% chance that the BB will have a pocket pair if the SB has a pocket pair.

So, 5.878% of SB having a pair x 5.954% of BB having a pair when SB does = 0.350% chance they both have pairs.

What about if the SB does not have a pair? Well, it depends on whether he shares ranks with the button.

If he shares no ranks, there are 66 possible pairs out of 1128 hands. There are 962 hands out of 1225 that share no rank with the button.

If he shares one rank, there is 67 possible pairs out of 1128 hands. There are 264 hands out of 1225 that share 1 rank with the button.

If he shares both ranks, there is 68 possible pairs of 1128 hands. There are 9 hands out of 1225 that share both ranks with the button.

((962*66 + 264*67 + 9*68)/1225)/1128 = 5.920% chance that the BB will have a pocket pair if the SB does not have one.

So,

5.878% chance the SB has a pair. If he does,
5.954% chance the BB does to, If the SB does not,
5.920% chance the BB has a pair

5.878 + 0.350 + 5.920 = 12.148% chance one or both will have a pair.

Odds are 7.23:1, not 7.77:1