Saturday, July 22, 2006
Egg Head Likes His Booky Wooks
I almost made it though a whole month without going up to anyone and telling them how great they are. But yesterday I saw Action Dan wandering around looking a bit lost so I went up to him and gave him proper respect for his books. The first book in particular made something click for me and it's a fair bet that if it wasn't for HoH1 I'd still be a wage slave, so I'm sure a quick prop was in order.
It reminded me though that I meant to do a book review of the various tomes I brought with me or bought here. First of all Harrington 3, I didn't tell Dan that I was rather disappointed with this. Maybe I just wasn't in the right frame of mind to read it and I should try again when I'm back home, but it's all about getting into like Ivey's head and Negreanu's head and I'm sorry, if I need a book to help me against these guys it's going to be Kill Phil every time.
While that book underachieved for me, Sklansky/Miller's no limit book was a pleasant surprise. I thought it was very good indeed. There's a lot of stuff in there about sizing bets that's very useful. Another sign of a good book for me is that there's stuff in it that I thought of myself but I haven't read anywhere else. Two points in particular, I've been banging on lately about moving in rather than putting in a quarter of your stack and agonising when you are reraised, and it's right there in this book. Secondly, and I was aware of this before I bought it, he has introduced something he calls the "Sklansky-Chubokov ratings" which is exactly the same concept I set out here. When can you move in with a positive expectation even if your opponent knows your hand and will call with a better hand and fold with a worse one. Obviously the terminology is slightly different but basically it's exactly the same idea. Do you think I should sue ? Anyway as the book says, "Tournament pros know that these 'loose' all-in moves are correct ; in fact, this knowledge is the main reason many of them win money at all playing tournaments". Ouch, he nailed me. I'm not particularly happy that this knowledge is now out there but who's going to buy "No Limit Hold-Em Theory And Practice" by Sklansky when they could buy "How I Lucked Into Millions" by some guy who wears his hat backwards just like I do. Hopefully not very many people.
Off the theory, I enjoyed Nolan Dalla's Stu Unger book which didn't pull any punches. I didn't realise how much other bad stuff happened to the guy, such as his "father figure" died on the night after he first won the World Series, and his adopted son committed suicide. It was quite a sad tale in the end but well worth reading. Less controversial was "Swimming With The Devilfish" which I was given as an unusually relevant birthday present but it was all a bit of a hagiography and while it's just about worth the price for a few interesting snippets the guy just didn't want to offend anyone he was writing about, which I can understand, but that's why it falls a bit short for me.
It reminded me though that I meant to do a book review of the various tomes I brought with me or bought here. First of all Harrington 3, I didn't tell Dan that I was rather disappointed with this. Maybe I just wasn't in the right frame of mind to read it and I should try again when I'm back home, but it's all about getting into like Ivey's head and Negreanu's head and I'm sorry, if I need a book to help me against these guys it's going to be Kill Phil every time.
While that book underachieved for me, Sklansky/Miller's no limit book was a pleasant surprise. I thought it was very good indeed. There's a lot of stuff in there about sizing bets that's very useful. Another sign of a good book for me is that there's stuff in it that I thought of myself but I haven't read anywhere else. Two points in particular, I've been banging on lately about moving in rather than putting in a quarter of your stack and agonising when you are reraised, and it's right there in this book. Secondly, and I was aware of this before I bought it, he has introduced something he calls the "Sklansky-Chubokov ratings" which is exactly the same concept I set out here. When can you move in with a positive expectation even if your opponent knows your hand and will call with a better hand and fold with a worse one. Obviously the terminology is slightly different but basically it's exactly the same idea. Do you think I should sue ? Anyway as the book says, "Tournament pros know that these 'loose' all-in moves are correct ; in fact, this knowledge is the main reason many of them win money at all playing tournaments". Ouch, he nailed me. I'm not particularly happy that this knowledge is now out there but who's going to buy "No Limit Hold-Em Theory And Practice" by Sklansky when they could buy "How I Lucked Into Millions" by some guy who wears his hat backwards just like I do. Hopefully not very many people.
Off the theory, I enjoyed Nolan Dalla's Stu Unger book which didn't pull any punches. I didn't realise how much other bad stuff happened to the guy, such as his "father figure" died on the night after he first won the World Series, and his adopted son committed suicide. It was quite a sad tale in the end but well worth reading. Less controversial was "Swimming With The Devilfish" which I was given as an unusually relevant birthday present but it was all a bit of a hagiography and while it's just about worth the price for a few interesting snippets the guy just didn't want to offend anyone he was writing about, which I can understand, but that's why it falls a bit short for me.
# posted by Andy_Ward @ 7:49 PM 
Comments:
<< Home
Surprisingly I remember the Sklanskys rating concept from the original 2+2 thread. http://archiveserver.twoplustwo.com/showthreaded.php?Cat=0&Number=382772&page=
In compact format:
http://www.decf.berkeley.edu/~chubukov/rankings.html
Is it the same one in the book?
cheers
Aksu
In compact format:
http://www.decf.berkeley.edu/~chubukov/rankings.html
Is it the same one in the book?
cheers
Aksu
# posted by 
: 8:27 PM
: 8:27 PM
Yep, looks the same. I note that this thread is dated 2003, so of course I was only joking and no one should sue anyone just because one pre-dates the other :-)
I'd also like to thank Aksu for reassuring me that someone is reading this, I was starting to wonder ...
Andy.
I'd also like to thank Aksu for reassuring me that someone is reading this, I was starting to wonder ...
Andy.
Roy Brindley's autobiography is coming out soon.
Hope you give us a review of that as soon as it becomes available.
Hope you give us a review of that as soon as it becomes available.
# posted by : 5:05 AM
: 5:05 AM
I've had these chubukov rankings as a link in my "Poker" category for years, and I've never before known what to do with them. I mean, what is anyone meant to do with this explanation?
"N_call and N_fold are the number of hands that will call/fold given that you move in with the maximum stack that you will do so with. P|call is the probability of winning given that you are called (plus 1/2 the probability of tieing). The last number is the original question."
Well, finally I know what the numbers mean.
Now, Andy, for the slow of thought such as myself, could you tell me what I have to do with those numbers in the last column of http://www.decf.berkeley.edu/~chubukov/rankings.html to give me a "maximum stack required" when I am on the button, rather than in the small blind?
So, for example, according to these numbers, with Q5s and the blinds at $1-$2, I should go all-in from the small blind (when passed round to me) if I have $20 or less. Correct?
What mathematical process (in words that a five-year old could understand) would I have to apply to the numbers in columns two and three (which give the number of hands you are favourite against and the number of hands that you are a dog against) to come up with a maximum dollar figure when I am on the button and it has been passed round to me? In other words, when I am facing two hands rather than one?
We assume for mathematical purposes that both players play perfectly and both have complete information (i.e., they can see each other's cards, as well as yours). This gives us the "profitable even in the worst-case" scenario.
According to the chubokov columns, you have a 711/1225 chance of being favourite against a single random hand with the Q5s.
So, if it is two hands, your chance of being the favourite is (roughly) half that - yes? Which gives us 350/1225.
If I move down the table, that equates to something like 96s. which is 10 times the small blind. So, with Q5s, i would go all-in on the button when down to 10 x the small blind or less. Yes?
And so we move found the table, to the cut-off and so on.
Do I (once again, roughly) make it one-third the value of the heads-up situation when I have three opponents? That would reduce Q5s to 240/1225, which my table gives as the equivalent of 87off heads up, and 8x the small blind.
Like I say, I'm not looking for rigid mathematical accuracy here (although if you could explain how to obtain that in simple terms, that would be nice :-) ) But just a general rule of thumb on what I should do with the chubukov ratings as supplied.
P.S. If the answer to this question is available in the Sklansky/Miller book, just point me to the page, and I'll let you know if I understand it. I'm still working through chapter 2. :-)
PJ
"N_call and N_fold are the number of hands that will call/fold given that you move in with the maximum stack that you will do so with. P|call is the probability of winning given that you are called (plus 1/2 the probability of tieing). The last number is the original question."
Well, finally I know what the numbers mean.
Now, Andy, for the slow of thought such as myself, could you tell me what I have to do with those numbers in the last column of http://www.decf.berkeley.edu/~chubukov/rankings.html to give me a "maximum stack required" when I am on the button, rather than in the small blind?
So, for example, according to these numbers, with Q5s and the blinds at $1-$2, I should go all-in from the small blind (when passed round to me) if I have $20 or less. Correct?
What mathematical process (in words that a five-year old could understand) would I have to apply to the numbers in columns two and three (which give the number of hands you are favourite against and the number of hands that you are a dog against) to come up with a maximum dollar figure when I am on the button and it has been passed round to me? In other words, when I am facing two hands rather than one?
We assume for mathematical purposes that both players play perfectly and both have complete information (i.e., they can see each other's cards, as well as yours). This gives us the "profitable even in the worst-case" scenario.
According to the chubokov columns, you have a 711/1225 chance of being favourite against a single random hand with the Q5s.
So, if it is two hands, your chance of being the favourite is (roughly) half that - yes? Which gives us 350/1225.
If I move down the table, that equates to something like 96s. which is 10 times the small blind. So, with Q5s, i would go all-in on the button when down to 10 x the small blind or less. Yes?
And so we move found the table, to the cut-off and so on.
Do I (once again, roughly) make it one-third the value of the heads-up situation when I have three opponents? That would reduce Q5s to 240/1225, which my table gives as the equivalent of 87off heads up, and 8x the small blind.
Like I say, I'm not looking for rigid mathematical accuracy here (although if you could explain how to obtain that in simple terms, that would be nice :-) ) But just a general rule of thumb on what I should do with the chubukov ratings as supplied.
P.S. If the answer to this question is available in the Sklansky/Miller book, just point me to the page, and I'll let you know if I understand it. I'm still working through chapter 2. :-)
PJ
# posted by : 9:36 AM
: 9:36 AM
Pete,
The second and third columns don't add much. Take the basic number and divide by 3 for the "M factor" required, where M = your stack divided by (blinds + antes). This is the maximum M to play the hand in the small blind. For the button, divide by 2. For the cut off 3. More than 3 it probably loses some accuracy.
It's true that the book skips over this quite quickly.
Fred,
It is a simple game. Me like cards, all in :-)
Andy.
The second and third columns don't add much. Take the basic number and divide by 3 for the "M factor" required, where M = your stack divided by (blinds + antes). This is the maximum M to play the hand in the small blind. For the button, divide by 2. For the cut off 3. More than 3 it probably loses some accuracy.
It's true that the book skips over this quite quickly.
Fred,
It is a simple game. Me like cards, all in :-)
Andy.
Hi Andy:
Actually, I did a bit of thinking about the whole thing later in (my) day. Although the dividing by two and dividing by three (for the button and the Cut Off) works well when you have hands that have some strength, the calculation falls apart fairly quickly when you are likely to be a dog.
The chance of "being behind to at least one opponent" is of course calculated by taking the chance of being ahead of both opponents, and subtracting that number from 1.
So, taking one of your favourite examples (A7s), you have an 88% chance of being ahead of a single random hand. So your chance of being ahead of three random hands is 0.88 ^ 3, which is 60%. I just move down the chubukov table until I find the magic 60%, and it's at the 17xSB mark (or where, in Action Dan terms, M=6).
Those numbers kind of feel right to me, so I may compile my own "complete" table for reference when playing online.
I printed out your original table when you started writing this, and after a while I kind of had a feel for it in my head. However, it looks to me that, with manic callers of overbets around in the online game, you can move to higher levels of M (above, say 7) and still get callers of your all-in bets.
BTW. Check out Terrence Chan's latest post. He, like me, got there in then end. "I have become a strong believer in the linear value of chips".
PJ
Actually, I did a bit of thinking about the whole thing later in (my) day. Although the dividing by two and dividing by three (for the button and the Cut Off) works well when you have hands that have some strength, the calculation falls apart fairly quickly when you are likely to be a dog.
The chance of "being behind to at least one opponent" is of course calculated by taking the chance of being ahead of both opponents, and subtracting that number from 1.
So, taking one of your favourite examples (A7s), you have an 88% chance of being ahead of a single random hand. So your chance of being ahead of three random hands is 0.88 ^ 3, which is 60%. I just move down the chubukov table until I find the magic 60%, and it's at the 17xSB mark (or where, in Action Dan terms, M=6).
Those numbers kind of feel right to me, so I may compile my own "complete" table for reference when playing online.
I printed out your original table when you started writing this, and after a while I kind of had a feel for it in my head. However, it looks to me that, with manic callers of overbets around in the online game, you can move to higher levels of M (above, say 7) and still get callers of your all-in bets.
BTW. Check out Terrence Chan's latest post. He, like me, got there in then end. "I have become a strong believer in the linear value of chips".
PJ
# posted by : 9:08 PM
: 9:08 PM
Yes, I think it does fall down with weaker hands and more players behind you. I can trust it with Ace-rag and 4 players behind me but beyond that you just have to figure that most people will call you with a reasonable range like AT/77, or just fall back on good old-fashioned judgement based on experience.
It's quite a comfort to read bloggers like Chan and Matros who are playing in a very similar style to me in short/mid stack scenarios [1]. Intelligent guys will get there but it's amazing how long it actually takes for these concepts to sink in. I think one definition of genius is the ability to realise these things very quickly. Few people have it.
Andy.
[1] The trouble is they play a big stack WAY better than I do, which is what I have to work on.
It's quite a comfort to read bloggers like Chan and Matros who are playing in a very similar style to me in short/mid stack scenarios [1]. Intelligent guys will get there but it's amazing how long it actually takes for these concepts to sink in. I think one definition of genius is the ability to realise these things very quickly. Few people have it.
Andy.
[1] The trouble is they play a big stack WAY better than I do, which is what I have to work on.
I should point out that when I say "few have it" I'm not implying that I do ; it's taken me forever to grind my way towards knowing these things.
Andy.
Post a Comment
Andy.
<< Home
