As for the subject of the thread, I believe that JoB is really 99.54% if you don't make any mistakes and play perfectly -- in other words, it's not really that good for most human beings, who are at least once in a while subject to imperfection, error, distraction, etc. So that's the expected return of the game, and I see no reason why a casino would risk their license (in states where they are expected to run machines that deal from software that emulates a real 52-card deck and associated expected card frequencies) to set the machine for a different return. Where the laws are not the same, and they are not required to emulate a real deck of cards, anything can happen; it's a "regular" slot machine, set to return what the casino wants it to return.
However, for the casino, the percentage "hold" (1.000 minus the actual percentage of coin-out divided by coin-in) will almost always be greater than 0.46% over time -- just as for an imperfect human being playing the game, the expected return on playing will be less than 99.54% to the extent that one makes mistakes, which may be a little or a lot for any individual player. And for the casual player who doesn't have a clue as to correct strategy, and may play less than full coin as well, the casino does even better, so the total mix of players always makes the return for the casino better than the game's rated return with perfect play.
Many on this board seem to think they never make a mistake when playing video poker (and a few don't limit it to video poker); I am sceptical of such a claim, although I'll concede that their mistakes may be uncommon enough that they don't have a big impact on their expected return. Is it possible that a human being can play perfectly all the time forever? Yes. Is it likely? No. And perfect play on practice computer runs, away from the distractions of a casino, are not proof of perfect play in the casino setting, although there is undoubtedly a correlation.
For games other than video poker, using blackjack as an example...
Most casinos look at the "hold" as a measure of how their business is doing for table games. What a particular casino thinks is "good" business is likely entirely their own call. I'm sure they balance the net cash in against expenses, which include everything from personnel to property costs, and I'm sure they also think in terms of whether that particular patch of casino real estate would be better off, for example, populated with "other" slot machines than with video poker, or blackjack, or roulette -- and then, they need to balance that against what brings players in (for example, poker rooms are not a great use of casino floor space in terms of dollars per square foot, but they often attract people with money who are gamblers at other games as well, who would not attend their particular casino if it didn't have a poker room).
This is part of the reason that they want to know when a player buys in (which is why a blackjack dealer is usually required to announce the amount of cash being changed into chips -- "changing $500!!", getting the pit boss's approval -- plus verification of the amount and correctness of the transaction sometimes), and why the pit boss will sometimes come to the table after a player leaves to get the dealer's estimate of how many chips the player walked away with (and why the dealer will almost always offer to "color up" the chips the player has, in order to make that estimate more accurately).
The computer will allow a casino host to pull up this information and tell the player "you played for two hours on Monday with an average bet of $25 and you won/lost $xx" -- and to tell the player how much more they need to bet or play to get more comps. Of course, video poker and other slots track this automatically via the players card (I haven't played blackjack for a long time, so don't know if they directly scan the players card at those tables now, like they do at many poker rooms).
The players, on the other hand, may look for chances to take a large-denomination chip or two off the table and sneak it discretely into a pocket, to make their losses seem larger than actually is the case when the dealer reports the amount the player "walked" away with.
As for what they do with this information, I have no idea, but I would think they would have industry benchmarks and would all want to be above average (obviously not possible), and that they might be adjusting the rules (which affect the expected return of the game with and without expert strategy and card counting) to increase the hold if they wish, balancing that against a "guess" of how much action they'll lose as a result (often, very little, since many players don't appreciate the affect of, for example, multiple decks vs single deck, ability to double down after splitting a pair, ability to double down on hands other than 10 and 11, whether the dealer is required to take a card with soft 17, and so on). But even an unsophisticated player will usually have some awareness, if a regular, that there IS a cost to more obvious rule changes, such as paying less than 3:2 for blackjack.
The pit boss does, however, ALSO make an estimate of average bet, and may make an estimate of average NUMBER of bets per hour (or may just use a fixed figure for that variable), and notes hours played, for purposes of rating the player. I have been told that some also make an estimate of quality of play (for blackjack). The total figures are apparently used to estimate the casino's expected win from the player, and based on that (and casino policy) the amount of comps that can be given.
When I played blackjack, I had an expected edge of 1 to 2% over the house, with a large variance (I expected to win one or two average-size bets per hundred hands, plus or minus twenty bets in a one hundred hand sample). At a full table, I would play sixty to eighty hands an hour, while I could play one or two spots and get 200 to 300 hands an hour heads-up against a fast dealer -- obviously, this affected my expected return per hour a lot, but I don't think the casinos gave me full credit for the amount of action I gave when heads up. (Oh yeah, and I sometimes would make a mistake, from speed when playing fast and from boredom when playing too slow -- I'm human too).
This is 1990's era rating system, it may be different today, but it certainly makes sense that the casino would have a policy to return in comps a more or less fixed percentage of their expected win from each player, and that they would have algorithms they use to determine that expected win. It also makes sense that they would be concerned about the "hold" -- and in fact, it makes even more sense that they would be concerned about TOTAL $$/hour of hold, on average, as that really reflects their expected income vs known expenses to forecast profit. A profitable game is not only one with a casino edge, but one where players like to bet more money per hour, and where the casino attracts more players. The wheel of fortune is a terrible game for players, but makes up for it with a big casino edge, even though you don't see someone playing the wheel very often -- while craps has less than 3% house edge (and further reduced with odds), but gets a lot more
action.
Even though the latter (being concerned with dollars per hour of "hold") makes more sense, casinos will employ anti-counter measures such as eight deck shoes, which can eat up 10 minutes of time every hour or less, where no one is betting anything while the shoe is shuffled -- and once a shift or once a day, there's a 45 minute drill when new decks are brought in.
More sense for the casino (and death for the card counter) is the continuous shuffling machine, which many use -- it kills the counter's ability to track remaining cards in the deck, and assures that table time is all "active" for betting purposes, as long as there are players at the table betting. I don't know if casinos have measured the effect on player participation -- many I know who are NOT counters just don't like the shuffling machines, suspicious that something funny is going on (which I do NOT believe to be the case at all).
For other table games, such as craps or roulette, the same principles apply, except that, for most players, there is only one variable the casino needs to note, and that's average bet size, since skill is not really a factor in those games (absent wheel tracking or dice control, about which I have not yet come to a good conclusion as to whether they are "real" or not), and number of bets per hour are pretty much the same for all players, although, like blackjack, there is some variation according to how full the table is.
Of course, one important variable for the player is how generous the casino is in returning comps for expected win (do they return half of the expected win, or 10%, or what?) -- and how they value those comps -- and it's not likely that a casino host or executive will share that information with a player.
For example, a casino returning $25 of their expected win in comps may give that out as two buffets, the face value of which may be $25 per person (twice as much), and for which the cost to the casino may be $15, and the value to the player will vary according to what the player considers a "fair value" for the buffet, ie, what he/she might have actually paid for and felt it was reasonable, if paying out of their own pocket. A room with a rack rate of $300 a night may not have that much value to the player getting it for free, and in fact may rarely sell at the full rack rate to non-comp'ed players. Getting something for free that you would not have otherwise purchased has less value than getting something for free that you would have paid for. A $150 dinner at the casino's perhaps-overpriced gourmet restaurant may compare to a $50 dinner at a nice restaurant at home.
When getting "rated" at these games, it used to be a good strategy to take note of when the pit boss was watching the table, and to try to assure that larger bets were out at that time (count notwithstanding) to maximize comps, as often the pit's estimate of average bet is made based on a very small sample of observation. Again, this is 90's era, things may be more sophisticated now.
Although I don't know if it's actually used, nor, if so, to what extent... I know there is now the highest level of technology available to track play and catch card counters. The pit boss's role in this activity CAN be markedly reduced, perhaps even eliminate. Chips and tables can have tracking devices in them (I think they use an RFID chip), so the casino knows EXACTLY how much is bet on each and every hand at each spot, and the video monitoring of play can automatically monitor every card dealt, with fairly unsophisticated software needed to identify the plays being made by each player.
It used to require a degree of suspicion by the pit boss or dealer, and then someone reviewing tapes or live play from the eye in the sky, to make a determination of a player's skill and whether they suspected care counting. However, with this new information, the right software could easily and automatically determine to what degree a player is adhering to "correct" basic strategy, whether they are increasing / decreasing bet amounts according to the card count, and whether they are varying play from correct basic strategy randomly, or appropriately for a high or low card count. Such software could rapidly determine whether a player is a good player or not, and could assign a mathematical probability that the player's betting and playing decisions are consistent with those of a perfect card counter -- and could notify the appropriate personnel that a player is ripe for being barred from the game.
With such software, for example, a 99% degree of certainty could be assigned to the decision that "this player is a card counter" -- and this could probably be done without a huge sample of hands. But the casino always has to decide if they want to "trespass" the 1% who correlate with perfect play due to random factors, as opposed to skill factors -- as those players are where the real money is made for the casino, in the long run. Throwing out the baby with the bath water is very bad business -- barring someone who's not a counter is not well received.
Furthermore, most card counters make sufficient mistakes to wipe out their narrow edge over the game, to the point where I personally think it would just be good business to let them play except in the most egregious situations (ie, high stakes team play) -- it's "good for the game" to have non-counting players KNOW that good card counters are usually allowed to play the game (non-counters who understand that counting is skill, not cheating, don't usually think such "good players" should be barred) and that they are long-term winners, and it's also good to allow imperfect (or inadequately bankrolled) card counters to play and blame their short-term bad luck for their lack of profit.
In the long run, I would think that most casinos will be smashing successes if they don't offend their customers, have a reasonable volume of business, and provide good entertainment value to the average player, and that their success should be enough that the return on investment of kicking people out is not enough to justify that kind of behavior. Unfortunately, that is long-term thinking, and most businesses today think short-term only.
--Barry Glazer
====================
1a. Re: Is JoB REALLY 99.54%?
> Date: Wed Aug 8, 2012 10:24 am ((PDT))
>
>
>
> When I dealt in Las Vegas, I was surprised to see that in
> BlackJack etc, the execs were always concerned about the
> "hold" in terms of what percent of the buy-in was retained
> by casino. Say someone buys in at the BJ table for $100.
> They place $5 bets on 100 hands, then cash out for $95. In
> mathematical terms, the player lost 1% on his entire amount
> wagered. The casino execs never cared about that; they cared
> that the player had lost 5% of the money he bought in with.
>
> This might be because the casino's definition of "hold"
> gives them results that more closely relate to casino
> profits. Or it might be because it's much much much easier
> for the casino to track results of a table game using their
> definition of "hold." (Think of how easy it is for casino to
> track the cash going into the BJ table slot and how many
> chips go out of the dealer's tray... vs how difficult it
> would be to track the total amount wagered.)
>
> So perhaps they talk about video poker "hold" in the same
> way. And it might make sense from a business perspective. If
> someone comes into a casino with $100 and sits down at a VP
> machine, perhaps the most meaningful metric for the casino
> to track is how much of that $100 will be left whenever the
> player leaves.
>
