Review: The Science of Poker

[grapes]

This book is terrible - total nonsense. Amazon auto-recommended it to me, and I was curious, so I bought it.

Basically, it seems like someone who has never played poker bought Turbo Texas Hold'em, ran a bunch of simulations, and then printed them in a book, complete with terrible advice based on these statistics.

This is In fact exactly what happened, as he says on the first page of the introduction that his results come from simulations run on Wilson software, and then adds that some probabilities come from books like "Super Systems by Doyle Brunson." He also claims to have taken stats from " 'Formula Won' by Michael J Barry" which doesn't exist as far as I can tell.

For example, he says that if you have QTs and someone that you know has AKs raises in front of you, you have to call and at least see the flop. You should probably fold Q7s, but if you are against the AKs and three others then it is profitable. Q7s against AKs and three random hands wins 21% of showdowns according to Turbo TH; therefore you should call because 21% is more than one-fifth, and there's 5 total players. (hmm... is this where the nickname "computer hand" for Q7 came from?)

He refers to Omaha hands as "X-X-X-X(s)" and "X-X-X-X(o)", for suited and unsuited. Apparently, it doesn't matter which cards are suited in "A-9-8-7(s)".

He covers hold'em, omaha, and stud, each in limit and pot-limit. There's no mention of no-limit. Actually, that's probably for the best, that he covers games like pot-limit stud that aren't really played. That way his nonsense can't do any damage.

The author is an Iraqi native, according to the back cover, but you only need to read a few sentences to know this isn't the writing of a native English speaker. It's filled with grammatical mistakes, typos, and awkward phrases - at one point he calls another player's hand his "attainment." The organization isn't any better.

He divides hold'em hands into "pairs" and "non-pairs". The sub-categories under non-pairs are "ace-high hands," king-high, all the way down to 9-high and finally "8 high and lower hands." His insight on each type starts with "you will be dealt an X about 15% of the time", where X is whatever high-card hand he's talking about. Gee, thanks... I didn't notice a pattern forming after the ace, king, queen and jack were all "15% of the time."

The numerous statistics and charts in the book might still be helpful despite the nonsense around it, except that those are often useless, and other times not even fully specified (there's a chart in the stud section giving percentages of how 6c-7c-8c-9c on fourth street does against a pair, two pair, and trips - without saying what ranks those are). I ran simulations to verify four of his charts, and found most of the numbers off by 1-3%.

Returning books on amazon is a hassle, and I usually wouldn't bother for $20, but this one is going back on principle.

Total nonsense. Avoid this book.

[theRealMacoy]

grapes,

thanks for the excellent review. i have been curious about this book for awhile now. it is on my wish list, as it is still unavailable to your canadian little bro's. now armed with this new info, i can safely eliminate it, so any lingering curosity doesn't prevail when we finally get access (the forbidden fruit always seems to taste sweeter.....well not in this case, in fact the apple was rotten).

thanks again,
The Real Macoy

ps. absolutely return that stinkin' book!

[BluffTHIS!]

Although much has been written about the relative reliability of hot & cold simulations, I think Dr. Mahmoud is correct in some of the reasoning (I have the book). You are looking at this only from a best hand perspective pre-flop and not from an expected value one. He is merely stating that although you would not want to take such a hand as QTs against AKs headsup to try to outflop it, if you are getting enough odds on your money, and assuming you play correctly from the flop on then playing such a hand is correct. By your reasoning you would never call 1 raise in no-limit with a medium pocket pair to try to flop a set on aces when the implied odds stack-wise are there (this of course assumes that you know opponent has aces but he doesn't know your hand, and that he will go broke or lose a lot with them).

As far as your comments about pot-limit not being played, I play pot-limit omaha and holdem almost every day, and although not played much in the U.S., pot-limit stud is played in London where Dr. Mahmoud lives. His simulations on omaha should be of considerable use to omaha players who often overvalue 4 in-a-row suited hands like 8765 and play them in early position, not realizing how many opponents they really need to be profitable, similar to HPFAP advice on not usually playing medium suited connectors or small pairs except in late position with several other players in an unraised pot.

So I do think there is some valuable information in this book if you use it properly as part of your game plan. And for what it's worth, Dr. Mahmoud is a well-respected semi-professional player, whom I actually believe plays very tight despite your conclusions from his simulations.

[Michael Davis]

I don't really see a compelling reason to fold any hand if you know your opponent has AK.

I think it's hilarious that a book like this could get published. Mad props to the author.

-Michael

[BluffTHIS!]

[ QUOTE ]
I think it's hilarious that a book like this could get published. Mad props to the author.


[/ QUOTE ]

You did not state that you actually have read this book, but it seems likely not given your opinion. You seem to assume that it is merely a collection of simulation printouts when it is nothing of the sort. Dr. Mahmoud uses the results of exhaustive simulations to form strategy advice, which although possibly bordering on weak-tight, is nonetheless well written and though out. Although I do not consider it a complete exposition on the various games discussed, I still believe it to be a valuable contribution. And I am giving opinions based upon actually having read the book.

[grapes]

You make a couple good points, but I stand by my review. The book is garbage, IMO, and following its advice will cost a lot of money.

I did not say that pot-limit is not played - I said that in hold'em pot-limit is much less popular than no-limit these days; and that in stud, pot-limit is rare. I've talked to some European players and was told it (PL stud) is played but that it would be hard to find a regular game anywhere. Regardless, this "edition" or whatever you want to call it would have been updated to at least mention no-limit hold'em, if the author had a clue or cared.

Do you have any reliable information on the author? The back cover says he is a semi-pro player. Even if that's true, that hardly qualifies him to write a book on poker, and presume himself enough of an expert to write a book that he calls definitive, and to cover six different games , no less. Lots of full-time pros not talking out of their asses have written bad books.

[ QUOTE ]
He is merely stating that although you would not want to take such a hand as QTs against AKs headsup to try to outflop it, if you are getting enough odds on your money, and assuming you play correctly from the flop on then playing such a hand is correct. By your reasoning you would never call 1 raise in no-limit with a medium pocket pair to try to flop a set on aces when the implied odds stack-wise are there

[/ QUOTE ]

It is nothing like that. First of all, the context is limit, not no-limit. Second, implied odds are not nearly as good with QXs as with a pocket pair. With the pocket pair, you can flop a set and be pretty confident you have the best hand. With Q7s against AKs, what are you looking to flop, that will get a lot of action? Further, the assumption is that you know he has AKs, but you never really know that. What if the raise is just from a tight player, meaning he could have AA,KK,QQ,AK,AQs. Still like cold-calling raises with a random queen?

Besides all that, he is saying that if the AKs raises you can call if 3 other people call too. He bases this on simulations that include three other random hands. If 3 other people cold-call the raise too, I think their hands will tend to be a little better than completely random. Oh, and I just reread it - he says you should call with QTs against AKs - not sometimes, but you should call, and the questionable situation is with Q7s.

You defend his Omaha section, which I think is just as terrible as the rest of the book. For example - he says that with A-Q-J-T , you want 3 opponents, no more, no less; and with A-J-T-9, you want lots of opponents - based again on hot and cold sims, I think. He doesn't explain why. And what about him referring to hands as "A-J-9-7(s)"??

I agree with you that hot & cold simulations can be very useful, if applied in the proper context. I bought this book knowing it would make extensive use of numbers like that. The problem is, he doesn't put anything in a useful context. He justifies playing that Q7s against AK and 3 random hands because it wins 21% of the time in showdown sims. I can't find any mention of "implied odds" or "expected value" in the book, or anything that would make me believe he even understands the concepts.

[BluffTHIS!]

Regarding his credentials, I've read comments by Ciaffone and Reuben and David Spanier talking well of him and I have seen one of them refer to his research. Regarding the QTs vs AKs, the fact that other hands are random does not matter as it is merely a matter of odds on your money knowing you have a potentially dominated hand but getting odds to see a flop assuming you can play correctly from there, which means not being a calling station with either top pair weak kicker or middle pair. And regarding his selection of games and it not including no-limit, he is merely talking about games regularly played at the Victoria Grosvenor Casino in London where he plays in the big game regularly. I don't know about the validity of some of the plo stuff regarding the hands you mentioned like AJT9, but I do believe it correct regarding the number of opponents to play something like 9876ds, which like I said is similar to 76s in holdem. The book is by no means perfect, but as I said it does contain some valuable information you can add to your game if you don't rely on it entirely.

[maurile]

Here is Rolf Slotboom's review, for what it's worth.

[SheridanCat]

I'm not defending this book. I own it but have yet to plow through it. Also, I think hot and cold sims are not particularly useful. However, I had to run the QTs vs AKs and see how it played out. I put the other players in with hands that could potentially call the AKs player's raise. I gave one player a random hand assuming he was the big blind and came along cheap. PokerStove says:

<font class="small">Code:</font><hr /><pre>
equity (%) win (%) / tie (%)

Hand 1: 30.5261 % [ 00.30 00.00 ] { AcKc }
Hand 2: 21.4211 % [ 00.21 00.01 ] { QhTh }
Hand 3: 19.7995 % [ 00.19 00.01 ] { TdTc }
Hand 4: 16.8530 % [ 00.17 00.00 ] { 7d7c }
Hand 5: 11.4003 % [ 00.11 00.00 ] { random }

</pre><hr />

QTs actually does have about the proper equity to be in this hand. Not that I'd do it.

Regards,

T

[droidboy]

[ QUOTE ]
Besides all that, he is saying that if the AKs raises you can call if 3 other people call too. He bases this on simulations that include three other random hands. If 3 other people cold-call the raise too, I think their hands will tend to be a little better than completely random.

[/ QUOTE ]

For what it's worth, grapes is exactly right. People don't cold call raises with random hands. Their hands tend to look like:

<font class="small">Code:</font><hr /><pre>
5678888777766
75##875443332
876##64322111
8657775321111
744285432111.
52221#54211..
3111..#3321..
3......#321..
2.......83211
2........7211
2... ...611
2... ...51
2... ....5
</pre><hr />

The matrix is the canonical preflop matrix with AA at the upper left, and A2s at the upper right. Each hand is weighted from 0-10, with . representing less than 1, and # representing 10. People cold call most often with medium pairs like 77 and 88, as well as with big suited broadway cards.

I can only imagine how bad a book would be which relies on hot/cold simulations, especially versus random hands.

- Andrew