On Fri, 13 Jul 2007 15:16:05 -0700, bob wrote:

On Jul 11, 1:01 am, Patrick Volk wrote:


In chess, no matter how much you study, the odds of being able to hold
your own against Kramnik or Kasparov in their primes is next to zero.
Mastery in this sense is not going to happen, and if it does, you
would be one of a dozen or so people who can make a decent "living"
from chess.

The disparity is less in poker, because of the variability. Where you
against Kramnik might leave you at 1-199 if you played him 200 games,
it would be more like 10-190 if you played Phil Ivey or someone.


I think your 5% estimate is not even close. In 2 person play Frank
Frigo argues that someone
who knows the rules (even a 10 year old child) could be taught within
a few seconds how to win about 25% vs. a
World class player in poker. His argument is good. Read
http://www.gammonlife.com/writers/06frigo2.htm

Bob Koca

I think the methodology is suspect. He tries to make poker a 2-player
equivalent game, which it isn't. He then goes on to say the dice
winning moves can cause swings of 100% towards the later game... If
that isn't pure 'probability' (a.k.a luck) I don't know what is!

As far as the all-in, all the time strategy, if it's known, it
changes. If Ivey waits for a 75% sure thing (he mentions 10-10, which
is the 7th best hand out of 169 different ones... But your chances of
getting that or better are less than 5%!).
Informally thinking about it, if the blinds are 10% of the stack,
and my 75% hand comes 1 out of 4 times (A-x, K-x Q-x, and any PP). Of
course, I could get a hand worse than the all-in guy, but just as
easily I could get one that has more of a 90% chance. I should hit
before I go less than half of his stack.

Let me interject a few more things:

Every decision in Backgammon is a skillful decision

In poker, it probably doesn't have to be, but it should be. You should
be aware of the game you're playing, and the effect position has.

With the exception of forced moves, every choice requires a skillful assessment of alternatives

Very true in poker. Alternatives offer different levels of risk and
reward.

Some choices carry more weight than others but all are affected by some amount of skill

Very true in poker as well.

A raw beginner (someone who has just learned the basic rules and a few general strategies) would have almost no chance of defeating a world-class expert in a 25-point match
The expert would almost certainly be greater than a 95% favorite

This is a bit of apples and oranges. Poker isn't a 2-player game,
but I think you'd have to say a RAW beginner (keep in mind, the 'new
faces' you see in poker have been playing online for years, and for
money) would have about the same chance in poker.

There is great parity among top players
Among experts, it is rare to win more than 60% of open level matches over the long haul

Very true in poker.

There is extreme volatility as a result of the dice
Single game winning chances can swing 5 to 10% on the opening moves and up to 100% in late game situations

Game winning chances swing 100%? But skill is still involved? Huh?

Skill is based on more than just the fundamental knowledge of the game

Very true in poker. Backgammon, you certainly have to take into
account your opponents play. Knowing what beats what in poker is
fundamental knowledge. Knowing what starting hands play well, and what
plays well with lots of players, and knowing position, pot odds, how
to deal with tight players, loose players, and passive and aggressive
ones as well is not fundamental knowledge.

Temperament, preparation and execution matter a great deal

Patience is probably the main asset a poker player has. Temprerament,
and execution matter as well (preparation isn't detailed).


I would also posit that a good yardstick is the level of play of
computers in the game. They put forth a strong game (albeit at the
whim of the dice) of backgammon. How about poker?

On Jul 13, 11:48 pm, Patrick Volk wrote:
I think the methodology is suspect. He tries to make poker a 2-player
equivalent game, which it isn't.

I agree that multi player poker is more complicated than 2 player
poker. But my posting the link was
in response to your comment about 5% winning chances against a poker
expert. If you are refusing to consider
the 2 player version what did you mean then by saying "Where you
against Kramnik might leave
you at 1-199 if you played him 200 games, it would be more like 10-190
if you played Phil Ivey or someone."


If Ivey waits for a 75% sure thing (he mentions 10-10, which
is the 7th best hand out of 169 different ones... But your chances of
getting that or better are less than 5%!).

Not sure what you are getting at it since this is an if-then
statement without the then portion. Note though that the
chances of getting such a good hand being low argues that it is hard
for the expert to have a high winning% for the match.


Informally thinking about it, if the blinds are 10% of the stack,
and my 75% hand comes 1 out of 4 times (A-x, K-x Q-x, and any PP). Of
course, I could get a hand worse than the all-in guy, but just as
easily I could get one that has more of a 90% chance. I should hit
before I go less than half of his stack.

You get a 75% hand or better after 2 cards much less than 1 out of
4 times.



A raw beginner (someone who has just learned the basic rules and a few general strategies) would have almost no chance of defeating a world-class expert in a 25-point match
The expert would almost certainly be greater than a 95% favorite

This is a bit of apples and oranges. Poker isn't a 2-player game,
but I think you'd have to say a RAW beginner (keep in mind, the 'new
faces' you see in poker have been playing online for years, and for
money) would have about the same chance in poker.

So you seem to be holding to the 5% figure for 2-player poker.
After two cards the worst situation to be in
would be 2-7 off suit vs AA of those same suits. In an all-in
situation this gives more than 12% to the 27
player (http://www.holdempoker4u.com/poker_calculator.html). Can you
see why that means the all-in player
has at LEAST a 12% chance of winning the match regardless of how small
the blinds are compared to the starting totals?
This is a very crude lower bound. A better lower bound, though still
crude would be the chance of winning if one has 2
unknown cards vs. AA.

If a mathematical argument doesn't sway you, you could try a
simulation out for yourself.
You don't even need another person there. Just pay the blinds and
always assume that the opponent's
action will always be "all-in" and see what % of the matches you can
win.

There is extreme volatility as a result of the dice
Single game winning chances can swing 5 to 10% on the opening moves and up to 100% in late game situations

Game winning chances swing 100%? But skill is still involved? Huh?


The possibility of certain rolls in certain situations swinging
the game does not preclude that there
may have been skill involved prior to that point. He never said that
the backgammon is devoid of luck.


I would also posit that a good yardstick is the level of play of
computers in the game. They put forth a strong game (albeit at the
whim of the dice) of backgammon. How about poker?


I agree that computers play backgammon much better than poker. It
is a game more suited
to computers though so I don't see how that necessarily is proof that
it is a game requiring
more skill for humans.

Bob Koca

On Jul 13, 10:08 pm, Iceman wrote:


The "best theoretical strategy" is not known, even for heads-up
holdem, but surely would depend a lot on how your opponent plays.

For heads-up holdem, basic game theory tells us that there is indeed
an optimal
theoretical strategy. It doesn't depend on the opponent's play though.
You are
mixing up the ideas of theoretical strategy and practical strategy. As
backgammon example is
suppose I have two checkers on my 3 point vs my opponent who has 2 on
his ace point. The theoretically
optimal cube play for me is to not double. Practically though it might
be a double (if my opponent is so bad as
to pass such positions).

Bob Koca

On Jul 14, 1:05 am, bob wrote:
On Jul 13, 10:08 pm, Iceman wrote:



The "best theoretical strategy" is not known, even for heads-up
holdem, but surely would depend a lot on how your opponent plays.

For heads-up holdem, basic game theory tells us that there is indeed
an optimal
theoretical strategy. It doesn't depend on the opponent's play though.
You are
mixing up the ideas of theoretical strategy and practical strategy. As
backgammon example is
suppose I have two checkers on my 3 point vs my opponent who has 2 on
his ace point. The theoretically
optimal cube play for me is to not double. Practically though it might
be a double (if my opponent is so bad as
to pass such positions).

That would be the backgammon equivalent of a bluff then?

Will in New Haven

--


"Have faith in the Yankees my son and remember the great Dimaggio."
Ernest Hemingway, THE OLD MAN AND THE SEA




Bob Koca

Patti Beadles wrote:

In article om,
Hank Youngerman wrote:
It is much easier to count a deck in bridge than in blackjack.

I'm not a bridge player, but in my experience this is false.

I am a pretty average bridge player, but I think he's right, at least
for the sort of people who play abstract games well. The reason is that
Bridge counting is dependent on tactical patterns, so there's a mental
framework that the cards fit into. You're not remembering where each
individual card is, instead you're figuring out how the cards fit
tactical patterns, and in some cases you can tell immediately that
certain details are irrelevant.

You get to see half of the cards as soon as dummy comes down -- this
plus the bidding allows you to build a very good picture of the various
hands, usually with missing information reduced to a fairly small number
of questions like "Who has the Kc?" "How do the spades split?" On many
hands, it's not necessary to track *any* suit in complete detail (i.e.
more than just how many cards are played), and it's pretty rare to need
the detail in more than two suits.

If you have the kind of mind that puts these things into abstract mental
pictures, then once you understand the tactics of the game, keeping
track of the cards becomes fairly intuitive. I never put much energy
into learning how to do it, I just started doing it better and better
once I "got" the game, eventually it became a habit.

No way that would happen with blackjack.

For example, I also play casino, a simple two player card game where
card counting is de rigeur and critical. I am unable to keep track of
the whole deck in that game, because the particular tactical maps don't
encompass the whole deck the way they do in bridge or hearts.

People who play casino for real (if there are any left) absolutely count
every card, and you can pretty much destroy anybody who doesn't if you
do. If I played it for money or if there were bigtime tournaments, I'd
have to practice counting just like people do with blackjack. I never
had to practice counting in bridge. Not because I'm some kind of freak
with a photographic memory, but just because it makes more sense.


Michael

--
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

Iceman wrote:

Chess games rarely have the advantage shifting back and forth the way
backgammon games do, however, so I wouldn't think possession of the
cube would be nearly as important in chess as it is in backgammon.

I think that tends to keep the cube smaller, but it doesn't seem like it
provides any less advantage. If the skills aren't even close (after
adjusting for the handicap), then I think you're right that it isn't
very important -- the weaker player will have few chances to double.
But if the weaker player would win as much as 45% of the time, owning
the cube gives a significant advantage. Break even should be somewhere
around 40%. If the pawn alone was enough to equalize them, then Will
had a huge advantage with the cube.

The player who owns the cube can turn .6+ equity into a win. The player
who doesn't cannot.

Of course, fractional equity is an imaginary concept in chess, where the
true equity is always either -1, 0 or 1. But consider that Will is the
weaker player (he needed a pawn as well). If I'm playing somebody a
couple hundred ELO points ahead of me but I'm getting a pawn, it will be
fairly common for me to have a won game at some point but be at real
risk of making a critical mistake and ending up with a draw or loss.
But if I toss the cube, I don't have to worry about that (or if I do, at
least we're playing from doubled stakes in a position where I should win
75% of the time). My opponent doesn't have that luxury, unless he's
*already* turned the game around from my doubling him.

Most of his wins will be single wins that he must play out to the bitter
end, while I will only have to play mine out to the end when the stakes
are doubled.



Michael


--
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

On Fri, 13 Jul 2007 21:52:45 -0700, bob wrote:

On Jul 13, 11:48 pm, Patrick Volk wrote:
I think the methodology is suspect. He tries to make poker a 2-player
equivalent game, which it isn't.

I agree that multi player poker is more complicated than 2 player
poker. But my posting the link was
in response to your comment about 5% winning chances against a poker
expert. If you are refusing to consider
the 2 player version what did you mean then by saying "Where you
against Kramnik might leave
you at 1-199 if you played him 200 games, it would be more like 10-190
if you played Phil Ivey or someone."

And the article qualifies further as making a match against an amateur
a 25-point one. Why not a 1-point one?



If Ivey waits for a 75% sure thing (he mentions 10-10, which
is the 7th best hand out of 169 different ones... But your chances of
getting that or better are less than 5%!).

Not sure what you are getting at it since this is an if-then
statement without the then portion. Note though that the
chances of getting such a good hand being low argues that it is hard
for the expert to have a high winning% for the match.

I'm getting at he seems to not know what he's talking about when it
comes to poker.



Informally thinking about it, if the blinds are 10% of the stack,
and my 75% hand comes 1 out of 4 times (A-x, K-x Q-x, and any PP). Of
course, I could get a hand worse than the all-in guy, but just as
easily I could get one that has more of a 90% chance. I should hit
before I go less than half of his stack.

You get a 75% hand or better after 2 cards much less than 1 out of
4 times.

Ok, the top 51 hands (there are 169) is 24% of the time.
The top 42 is 20% of the time.



A raw beginner (someone who has just learned the basic rules and a few general strategies) would have almost no chance of defeating a world-class expert in a 25-point match
The expert would almost certainly be greater than a 95% favorite

This is a bit of apples and oranges. Poker isn't a 2-player game,
but I think you'd have to say a RAW beginner (keep in mind, the 'new
faces' you see in poker have been playing online for years, and for
money) would have about the same chance in poker.

So you seem to be holding to the 5% figure for 2-player poker.
After two cards the worst situation to be in
would be 2-7 off suit vs AA of those same suits. In an all-in
situation this gives more than 12% to the 27
player (http://www.holdempoker4u.com/poker_calculator.html).

The calculator says 11%, and that's not the worst hand (the worst is
5.3%)

AA against KK has a probability of 18%.

And your poker calculator appears to be using Monte Carlo analysis, no
pun intended (it says 13% on AA v. KK, but it's really 17%).

http://wizardofodds.com/holdem/calcu...handstrength2/ (it's quicker
too)

Can you
see why that means the all-in player
has at LEAST a 12% chance of winning the match regardless of how small
the blinds are compared to the starting totals?
This is a very crude lower bound. A better lower bound, though still
crude would be the chance of winning if one has 2
unknown cards vs. AA.

If the blinds don't exist, then I'm at 82%. If the blind is high
enough to not give me a choice, then it's 50%. If we allow seeing the
players cards, it is 95%.

The article as well made mention of a professional environment, a
25-point match, while for poker, that lassitude isn't given.


If a mathematical argument doesn't sway you, you could try a
simulation out for yourself.
You don't even need another person there. Just pay the blinds and
always assume that the opponent's
action will always be "all-in" and see what % of the matches you can
win.

If I take the lowest of the 25% hands (A7s), and the average hand (Q-7
of different suits) I get 23%.


There is extreme volatility as a result of the dice
Single game winning chances can swing 5 to 10% on the opening moves and up to 100% in late game situations

Game winning chances swing 100%? But skill is still involved? Huh?


The possibility of certain rolls in certain situations swinging
the game does not preclude that there
may have been skill involved prior to that point. He never said that
the backgammon is devoid of luck.

If I make the same statement about poker, it's called luck. He pretty
much said it is luck.



I would also posit that a good yardstick is the level of play of
computers in the game. They put forth a strong game (albeit at the
whim of the dice) of backgammon. How about poker?


I agree that computers play backgammon much better than poker. It
is a game more suited
to computers though so I don't see how that necessarily is proof that
it is a game requiring
more skill for humans.

It's the intangibles... All of the information in chess and backgammon
is on the table. There is probability in the backgammon dice, which
isn't known, and cannot be known until the dice are thrown.

I think the thing that rankles me about the analogy is he tries poker
at the worst case, and backgammon at the not-so-worst case.

Out of a field of 6,000 players in this years WSOP, there still are 3
former world champions (I think the top 100, and 2 or 3 more cashed).



Bob Koca

On Jul 14, 7:33 pm, Patrick Volk wrote:
And the article qualifies further as making a match against an amateur
a 25-point one. Why not a 1-point one?

Because a 1-point match isn't long enough for a superior player to
fully demonstrate his superior skill. A 1-point backgammon match is
too short for fair comparison to, say, a chess game with professional
time control. Similarly, if Phil Ivey and I were to play _one hand_
and one poker hand only for all the marbles, I know what my chances in
that contest would be!

Frigo's article refers to one by Bill Robertie that appeared in Inside
Backgammon in 1992. That's where the 25-point match length comes from.

Robertie was attempting to quantify the complexity of various games:
Go, Chess, Scrabble, Poker, Backgammon, Draughts, Blackjack, Craps,
Lotteries, Roulette. He took chess as an example: take the best player
in the world; find someone who beats the best player in the world 25%
of the time; find someone else who beats that second player 25% of the
time; and so on until you reach the bottom of the barrel -- an
absolute beginner. The number of skill differentials between best in
the world and absolute beginner is what Robertie called a "Complexity
Number." The more skill differentials, the greater the Complexity
Number, the more complex the game. Robertie's list:

Go 40
Chess 14
Scrabble 10
Poker 10
Backgammon 8
Draughts 8
Blackjack 2
Craps 0.001
Lotteries 0.0000001
Roulette 0

Why a 25-point match? Because that's what Robertie thought would make
for a meaningful comparison to chess and other games. He explained:
"We can now apply this process to any game, although we may have to
give some thought as to what constitutes a meaningful contest. In
chess, a single tournament game of four to five hours seems
reasonable. In backgammon it would probably be a 25-point match, in
scrabble perhaps a best of five series, and so on." A 25-point
backgammon match should also take about 4 to 5 hours. See David
Montgomery in the rec.games.backgammon thread "Which is greater: luck
or skill" beginning Aug 29 1995.

In chess, I believe, a players with a 200 rating point advantage has
an expected score of 0.75. Similarly in backgammon, the player with a
200 point advantage rates to win 75% of the time -- in a long 25-point
match, that is, not a 1-point quickie.

The article as well made mention of a professional
environment, a 25-point match, while for poker,
that lassitude isn't given.

Feel free to suggest some other format for a poker contest, lasting
4-5 hours, that you believe would be approximately comparable in the
amount of skill required by one professional game of chess, or one 25-
point backgammon match, or a game of Go.

On Jul 14, 10:33 pm, Patrick Volk wrote:


Ok, the top 51 hands (there are 169) is 24% of the time.
The top 42 is 20% of the time.



I don't see how you go from top 51 hands out of 169 gives 24%. Did
you do 51/(51+169) = .2318 and round?
That seems how you got the 20% figure as 42/(42+169) = .199. Also,
why not 42/169. Also why not worry about
if the cards are suited or not?

Anyways it is an irrelevant calculation. By 75% hand in the article
it clearly meant a hand that gives a 75%
chance of winning an all-in. Not a hand that is one of the 75% bests
hands for you.



The calculator says 11%, and that's not the worst hand (the worst is
5.3%)

AA against KK has a probability of 18%.

And your poker calculator appears to be using Monte Carlo analysis, no
pun intended (it says 13% on AA v. KK, but it's really 17%).

http://wizardofodds.com/holdem/calculator/handstrength2/(it's quicker
too)

Agree that you site is better. Mine never gives anything near 13%
though for AA vs KK though.
For the worst hand

On Sun, 15 Jul 2007 19:36:37 -0700, bob wrote:

On Jul 14, 10:33 pm, Patrick Volk wrote:


Ok, the top 51 hands (there are 169) is 24% of the time.
The top 42 is 20% of the time.



I don't see how you go from top 51 hands out of 169 gives 24%. Did
you do 51/(51+169) = .2318 and round?
That seems how you got the 20% figure as 42/(42+169) = .199. Also,
why not 42/169. Also why not worry about
if the cards are suited or not?

Anyways it is an irrelevant calculation. By 75% hand in the article
it clearly meant a hand that gives a 75%
chance of winning an all-in. Not a hand that is one of the 75% bests
hands for you.

I said 1 in 4 hands, and you disagreed. Also, I just quoted the
numbers, and know that all hands aren't created equal ( For any given
pocket pair, there are 6 ways to get it (.45%). For x-y suited, there
are 4, (.30%). For x-y unsuited, there are 12 (.90%))




The calculator says 11%, and that's not the worst hand (the worst is
5.3%)

AA against KK has a probability of 18%.

And your poker calculator appears to be using Monte Carlo analysis, no
pun intended (it says 13% on AA v. KK, but it's really 17%).

http://wizardofodds.com/holdem/calculator/handstrength2/(it's quicker
too)

Agree that you site is better. Mine never gives anything near 13%
though for AA vs KK though.
For the worst hand

The worst hand is AA vs. A-9 (PP covers the over card, and if the 9 is
a suit of the ace it adds about 1%). Not PP vs. PP (PP v. PP needs
only 1 card, while covered over card requires 2).