vpFREE2 Forums

Two disturbing articles on casino gambling

As someone who plays both video poker and live poker, I would agree with all of the explanation below. I like the concept of there being a continuum or spectrum of the element of chance vs. the element of skill in various games, and it would be interesting if there were a way to measure that, so you could place games on that continuum in a quantified manner. Right now, all we can do is say that one game is more or less dependent on chance than another, and be fairly certain of our statement, but we can't put a real number on it (except, I guess, to use expected return calculations, which can be calculated on the less complex games, but not on the most complex ones).

Video poker requires skill, but it is skill based on a fairly fixed strategy, that is dependent on known variables (on an "honest" or "fair" game), such as the type of game and makeup of the deck (does it have wild cards, jokers, etc) and the paytable. It is much like playing blackjack, where the strategy depends on the type of game (as determined by number of decks, what "plays" are permitted - e.g. double down after splitting a pair, doubling down on soft hands, re-splitting aces, and so on) - and where the strategy can be further refined if one counts cards, providing additional information as to the overall makeup of the remaining cards.

Sometimes the strategy is varied in VP, e.g., for advanced players, based on penalty cards. Sometimes the strategy in blackjack is varied, e.g., for card counters, the amount of the bet, and some play variations, based on the count. Likewise, there are "simplified" strategies for both games that are fairly quick and easy to return, but which are not optimal for minimizing the house edge.

In live poker, there is strategy, but it is NOT a fixed strategy; there are some mathematical principles that can be applied, but they require that one make assumptions about the hand held by one's opponent(s), as well as how they might play it, and those assumptions almost always have a degree of uncertainty as to their correctness. The ability to make correct assumptions based on incomplete information is what makes the skill in live poker very complex to master.

The variables that must be dealt with are NOT "known"; e.g., range of hands that the opponent might currently hold, whether he/she will bet or not, and/or how he/she will respond to a bet. And the historical responses of an opponent in given circumstances may not be predictive of future respones, since opponents can vary their play from hand to hand and hour to hour. In fact, poker is all about making good decisions based on incomplete information, while in VP and blackjack, the decisions are based on complete information EXCEPT for information as to what cards are actually coming on this particular occasion.

Finally, live poker has a psychological component; opponents can sometimes be manipulated by various means into taking actions that, at least in the long run, lower their expectation of winning, and likewise they may manipulate you; no such psychological component exists in VP, except perhaps as a need to have fun might drive a player to play a less optimal machine, or to vary from correct strategy "on a hunch" for some "reason".

The real difference in the games can be best understood by considering this: if you are playing blackjack at a table, and the other players at the table are all better players than you (they know more detailed strategy and/or play it more perfectly), you can still have a positive expectation from the game that is not dependent on chance alone, since you are not competing with those players (who would, however, have a higher positive expectation than you do).

Likewise, if you are playing VP at a bank of the same machines, and the other players at that bank are all better players than you (they know more detailed strategy and/or play it more perfectly), you again can still have a positive expectation from the game that is not dependent on chance alone, since, again, you are not competing with those players (who, again, would have a higher positive expectation than you do).

BUT, if you are playing live poker at a table and the other players at that table are all better players than you, even if you are ordinarily a very skilled player who ordinarily has a positive expectation, at this particular table, you have a negative expectation, and only elements of chance (getting lucky) can save you. Worse yet, you may not recognize the situation as one in which you are out-skilled for a while (and perhaps not at all if your skill level at such recognition is low).

In this way, while all three games have strategy as a key component, unlike most reel slots, the strategy for live poker is dependent on so many variables and so much incomplete information as compared to blackjack and video poker, that one can never "master" the game and play it perfectly at all times. Like chess, the factor of human opponents of varying skill levels makes the game of live poker far more complex.

I do agree that Lederer was in error in grouping video poker with roulette and lotteries. Roulette "strategy" is essentially to minimize losses by using minimum bets and avoiding the one or two bets with a higher house edge than the rest, and lottery "strategy" is essentially not playing at all until the jackpot is large enough to justify the enormous odds against hitting it (and factoring in the unknown variable of how many tickets will be sold, which can affect the probability of needing to split the jackpot even if you hit it). Nothing that you do beyond this can alter the probability of a win on a particular bet, once placed.

But to say that VP is not a game of strategy is simply not correct -- as I tell friends, if you think strategy is not important to VP, try any one of these approaches: always hold either nothing, or always hold either the first one, or first two, or first three, or first four cards dealt, or all five cards (or any other predetermined selection decided before you know what you will get) - no matter what they are and what the remaining cards are -- and you'll soon see that it's a game where strategy matters.

Likewise, for blackjack, use one of two extreme strategies: always hit every hand repeatedly until the dealer won't give you any more cards (expected return on a $1 bet is then zero with that strategy) or always stand on every hand no matter what it is and no matter what the delaer holds - again, you'll soon see that it's a game where strategy matters. Even playing blackjack hands by the same rules that the dealer must follow will give the house roughly an 8% edge, when just a little skillful play can reduce that to half or less, and highly skillful play, even without card counting, can reduce it to 1% or less.

So, it's a matter of personal judgment of where a game is on the spectrum -- a game of chance with an element of skill, or a game of skill with an element of chance.

--BG

···

===============

1m. Re: Two disturbing articles on casino gambling
Posted by: "fivespot" fivespot55@gmail.com fivespot838
Date: Mon Aug 17, 2009 2:39 pm ((PDT))

On Mon, Aug 17, 2009 at 6:57 AM, uny1604 wrote:
> What do others in the group think of this statement?
>
> "A poker player -- unlike someone playing roulette, a lottery or "video poker"
(which Lederer says is a misnomer; it is a game of chance governed by a machine)
-- is trying to apply skill, acquired by experience, to increase the probability
of winning each hand."

"video poker" is indeed a misnomer.

poker is a family of card games of the vying type, in which multiple
players bet in turn over who has the best hand, each player being
required to match the current bet or fold, and the accumulated bets
being awarded to the holder of the best hand who has not folded at the
end.

"video poker" is a type of slot machine, a single-player game with a
single fixed bet and varying payouts. it uses the hand rankings that
are also used in many forms of poker, and many variations use the draw
mechanism that is used in a few forms of poker, but otherwise has
nothing in common with poker. calling it "poker" is wildly inaccurate.

skill vs chance is a continuum. there is chance in chess, there is
skill in roulette, just not very much of it in either case. lederer
would argue that poker is closer to the chess end of this continuum,
whereas video poker is closer to the roulette end. i'd agree with him.

however, whether or not you agree with his characterization of vp, it
seems uncontroversial that any discussion of the legality of poker
should first ensure that the audience knows what game it is you're
talking about. there are an awful lot of people out there who don't
understand that poker and VP are entirely different games...

cheers,

five

I like the concept of there being a continuum or spectrum of the element of chance vs. the element of skill in various games, and it would be interesting if there were a way to measure that,

There are ways. One is the Sharpe Ratio (adv./std.dev.). Advantage = skill, standard deviation = luck.

In live poker, there is strategy, but it is NOT a fixed strategy;

That's being worked on, Chen and Ankenman's "Mathematics of Poker" is an excellent start.

···

--- In vpFREE@yahoogroups.com, "Barry Glazer" <b.glazer@...> wrote:

"Barry Glazer" <b.glazer@> wrote:

I like the concept of there being a continuum or spectrum of the
element of chance vs. the element of skill in various games, and
it would be interesting if there were a way to measure that,

<nightoftheiguana2000@...> wrote:

There are ways. One is the Sharpe Ratio (adv./std.dev.).
Advantage = skill, standard deviation = luck.

The actual relationship of skill vs chance in a game is more complex
than what could be expressed in a single number.

For most live poker games in the casino, there's a big gap in
expectation between a somewhat skilled player (someone who's read a
book or two about poker and understands a few strategy concepts) and
a completely unskilled player (someone who's played some kitchen-
table poker, but has never really thought about the game, and perhaps
hardly realizes that skill is involved.) This gap is generally greater
than the gap between a somewhat-skilled and a highly-skilled player.

The particular type of poker, particularly the betting structure
(whether it's limit or no-limit, and the size of the bets vs the
players' table stakes and total bankroll) will change this. In
general, though, there will be certain games in which a skill
differential between the players is highly significant (i.e.,
predictive of who will win in a very long session) if it's between
a decent player and a very bad one (meaning it's mostly a game of
skill), but hardly significant in a game between a decent player and
an expert (making it mostly a matter of chance).

Stuart
http://stuart-randomthoughts.blogspot.com/

Very interesting comments Barry. I couldn't have said it better.

···

--- In vpFREE@yahoogroups.com, "Barry Glazer" <b.glazer@...> wrote:

As someone who plays both video poker and live poker, I would agree with all of the explanation below. I like the concept of there being a continuum or spectrum of the element of chance vs. the element of skill in various games, and it would be interesting if there were a way to measure that, so you could place games on that continuum in a quantified manner. Right now, all we can do is say that one game is more or less dependent on chance than another, and be fairly certain of our statement, but we can't put a real number on it (except, I guess, to use expected return calculations, which can be calculated on the less complex games, but not on the most complex ones).......

There is a pretty simple way to compare the element of chance vs skill in various games. Almost all games played professionally have a ranking system. A luck quotient, LQ, could be defined by how often someone ranked in the bottom 10% beats someone in the top 10%. In games like tennis or chess this is a very rare event. A recreational tennis or chess player has a 0 probability of beating a professional. In many of the usual poker tournament formats the recreational level player often beats a pro to advance in a tournament.

Chris

···

--- In vpFREE@yahoogroups.com, "Stuart" <sresnick2@...> wrote:

"Barry Glazer" <b.glazer@> wrote:
> I like the concept of there being a continuum or spectrum of the
> element of chance vs. the element of skill in various games, and
> it would be interesting if there were a way to measure that,

<nightoftheiguana2000@> wrote:
> There are ways. One is the Sharpe Ratio (adv./std.dev.).
> Advantage = skill, standard deviation = luck.

The actual relationship of skill vs chance in a game is more complex
than what could be expressed in a single number.

For most live poker games in the casino, there's a big gap in
expectation between a somewhat skilled player (someone who's read a
book or two about poker and understands a few strategy concepts) and
a completely unskilled player (someone who's played some kitchen-
table poker, but has never really thought about the game, and perhaps
hardly realizes that skill is involved.) This gap is generally greater
than the gap between a somewhat-skilled and a highly-skilled player.

The particular type of poker, particularly the betting structure
(whether it's limit or no-limit, and the size of the bets vs the
players' table stakes and total bankroll) will change this. In
general, though, there will be certain games in which a skill
differential between the players is highly significant (i.e.,
predictive of who will win in a very long session) if it's between
a decent player and a very bad one (meaning it's mostly a game of
skill), but hardly significant in a game between a decent player and
an expert (making it mostly a matter of chance).

Stuart
http://stuart-randomthoughts.blogspot.com/

"kcace1024" <cy4873@...> wrote:

A luck quotient, LQ, could be defined by how often someone ranked
in the bottom 10% beats someone in the top 10%.

Suppose the contest is between the bottom group (people who play
VP entirely by intuition) and a middling group (people who have
spent < an hour learning/understanding a simplified JOB strategy).
There are 100 people in each group, and they each play 10,000 hands
of JOB. There's no time limit, so participants can play slowly to
make sure they're following their strategy/intuition correctly for each hand.

The LQ is very low. The probability is high that the more skilled
group will beat the bottom group.

Change the experiment to pit the middling group against a group of
expert players, who understand in detail the precisely correct
mathematical choice in every possible JOB situation. Now the LQ is
much higher. While the difference in skill between the two groups is
still huge, it will have a much smaller effect on the results, so
Luck will often predominate.

Stuart (RandomStu)
http://stuart-randomthoughts.blogspot.com/

"kcace1024" <cy4873@> wrote:
> A luck quotient, LQ, could be defined by how often someone ranked
> in the bottom 10% beats someone in the top 10%.
Suppose the contest is between the bottom group (people who play
VP entirely by intuition) and a middling group (people who have
spent < an hour learning/understanding a simplified JOB strategy).
There are 100 people in each group, and they each play 10,000 hands
of JOB. There's no time limit, so participants can play slowly to
make sure they're following their strategy/intuition correctly for each hand.

The LQ is very low. The probability is high that the more skilled
group will beat the bottom group.

If the difference between the two groups is 4.5%, and the variance is 20, the N0 is about 10,000 hands, meaning there's still about 15% (30 people) whose results are in the other group due to luck alone. You need about 4N0 to clearly differentiate the two groups (about 2.5% error rate). Also, the fact that your sample size (10,000 hands) is less than the top jackpot cycle time (about 45,000 hands for the royal) is problematical, but the effect of royals on the dataset can be compensated for. The number of royals a player gets generally is not a strong function of the strategy employed, except in extreme cases (like the royal only strategy used in tournaments). Strategy mostly effects the cost of getting an average royal.

Change the experiment to pit the middling group against a group of
expert players, who understand in detail the precisely correct
mathematical choice in every possible JOB situation. Now the LQ is
much higher. While the difference in skill between the two groups is
still huge, it will have a much smaller effect on the results, so
Luck will often predominate.

If the difference is only a percent, the N0 goes out to 200,000 hands. At 10,000 hands, luck (variance) is dominant.

Just to summarize, the significance of N0 is that below N0, luck (variance) is dominant (technically, 1SD > average), while above N0, skill (or lack of it) is dominant (technically, average > 1SD). For games with higher N0, a larger number of hands needs to be played for skill to become dominant over luck.

http://members.cox.net/vpfree/Bank_NO.htm

···

--- In vpFREE@yahoogroups.com, "RandomStu" <sresnick2@...> wrote: