What is "Hold Percentage" ?
November 9th, 2010 at 6:55:59 AM
permalink
In the thread about Borgata's hot streak, there is a link to an article about it, which includes the following:
I.E. In that example, if the player bought in for $600, but still only lost $30, would the hold be 5%?
It seems to me that if that's really the way it's calculated, then it's a pretty worthless figure. If the player's loss remains unchanged, a lower hold means nothing except more thumb licking for the cage's money counters.
Is that REALLY what 'hold percentage' is?Quote:The hold percentage is the term for the ratio of gaming chips the casino keeps compared with the number it sells to gamblers. For example, if a blackjack player buys $200 in chips and loses $30 in bets, the hold percentage for the casino is 15 percent.
I.E. In that example, if the player bought in for $600, but still only lost $30, would the hold be 5%?
It seems to me that if that's really the way it's calculated, then it's a pretty worthless figure. If the player's loss remains unchanged, a lower hold means nothing except more thumb licking for the cage's money counters.
I invented a few casino games. Info:
http://www.DaveMillerGaming.com/
Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
November 9th, 2010 at 7:03:47 AM
permalink
Yes. On table games there is no accurate way to count total wager volume or "handle". The only measure you have is how much money was in the chip tray at the beginning of the day vs. at the end, and how much cash you took in (loosely - it's more complicated than that). It's not an accurate measure of much, but it's the only yardstick there is, so regional jurisdictions have come to understand what's a healthy "hold" and what's not. For example, blackjack in Nevada holds around 10-11%. We all know the EV is nowhere near that - it's probably 1.5% when you factor in the bad players. In Washington State, blackjack holds closer to 18%.
The only way you can really know bet-by-bet is if you have some mechanism for table tracking. That's a perennial product offering at the gaming shows, but most casinos don't do it. As a result, that happens to be one of the selling points of the electronic table game systems.
The only way you can really know bet-by-bet is if you have some mechanism for table tracking. That's a perennial product offering at the gaming shows, but most casinos don't do it. As a result, that happens to be one of the selling points of the electronic table game systems.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice."
-- Girolamo Cardano, 1563
November 9th, 2010 at 8:24:56 AM
permalink
Thanks.Quote: MathExtremistIt's not an accurate measure of much, but it's the only yardstick there is...
It just seems to me that a hold DOLLAR would be more useful than a hold percentage...
I invented a few casino games. Info:
http://www.DaveMillerGaming.com/
Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
November 9th, 2010 at 8:40:07 AM
permalink
They do -- table game numbers usually report win (in dollars), drop (in dollars), and percentage (win/drop). Here are the NJ numbers:
http://www.state.nj.us/casinos/financia/mthrev/docs/2010/201009revenue.pdf
See columns d, e, and f respectively.
And you're right - Borgata got killed in BJ last month. 6.8%? Wow.
http://www.state.nj.us/casinos/financia/mthrev/docs/2010/201009revenue.pdf
See columns d, e, and f respectively.
And you're right - Borgata got killed in BJ last month. 6.8%? Wow.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice."
-- Girolamo Cardano, 1563
November 9th, 2010 at 9:59:52 AM
permalink
Dave,
Table Hold is crucial - it's the buy-in/chip-in versus what the table keeps as profit, a bottom-line number for a table game in the real world.
For game designers, you review your live table's hold from the info that casinos may supply to you and your distributor (IF THEY do that for you!), to determine the performance and health of your game. Fiesta did that for me in the first month, and I have it on a spreadsheet. Only Me, my business partner, and DEQ see it. It's the figure of profitability "from where the rubber hits the road," and is what the casino looks at in deciding to keep or drop your game.
Usually, you need about 12 "table months" or more to get a very good and stable idea of how a game is working for a casino, kept and analyzed in a spreadsheet, with the first three months really making an impression on a casino.
The Missouri gaming commission publishes this info on the Internet for Each game (Regular Blackjack, BJ+21, Three card Poker, EZ Pai Gow, Fortune Pai Gow, Emperor's Challenge Pai Gow, etc.), and Lubin-Jones and DEQ systems reviews this info on EZ Pai Gow, to see if it is overholding or overdumping.
You can review this at http://www.mgc.dps.mo.gov/2010_fin/FY10_slot_table.htm. Select the bottom-row tab of "AMERKC," to see Ameristar Kansas City Casino, for instance. You see the individual brands of games (EZ Pai Gow, Emperor's Challenge, and Fortune Pai Gow) listed out seperately, so you can compare the popularity of the product through "Drop" - which is buy-in cash drop into the drop boxes, and "Actual Hold," which is what a casino keeps of that action.
A good table hold for a table game is in the 17% to 26% range or so; above that, players feel their losses really hard, and below that level, the casino laments the poor income produced, unless the game drops a huge amount: a 5% table hold of $1,000,000 drop is $50,000; a 25% table hold of $100,000 in buy-ins is $25,000. Baccarat play affords this.
Table hold is tricky for side bets, because the side bet is played in conjunction with (and overlaid on) a base game, masking the individual side bet performance to a great degree.
You can assume very roughly that one percent of house edge equals a 10% of table hold, for a game between 1% up to 4% HE; It's a remarkable corrolation.
The 10% BJ table hold in NV is because of experienced players making BJ a 1% HE game or so for them. The 18% BJ hold in other areas may be due to poorer play. A LOT of other factors are involved, but it's rough ballpark rule-of-thumb that is used, and that is usable from data of a high-volume extended period.
Table Hold is crucial - it's the buy-in/chip-in versus what the table keeps as profit, a bottom-line number for a table game in the real world.
For game designers, you review your live table's hold from the info that casinos may supply to you and your distributor (IF THEY do that for you!), to determine the performance and health of your game. Fiesta did that for me in the first month, and I have it on a spreadsheet. Only Me, my business partner, and DEQ see it. It's the figure of profitability "from where the rubber hits the road," and is what the casino looks at in deciding to keep or drop your game.
Usually, you need about 12 "table months" or more to get a very good and stable idea of how a game is working for a casino, kept and analyzed in a spreadsheet, with the first three months really making an impression on a casino.
The Missouri gaming commission publishes this info on the Internet for Each game (Regular Blackjack, BJ+21, Three card Poker, EZ Pai Gow, Fortune Pai Gow, Emperor's Challenge Pai Gow, etc.), and Lubin-Jones and DEQ systems reviews this info on EZ Pai Gow, to see if it is overholding or overdumping.
You can review this at http://www.mgc.dps.mo.gov/2010_fin/FY10_slot_table.htm. Select the bottom-row tab of "AMERKC," to see Ameristar Kansas City Casino, for instance. You see the individual brands of games (EZ Pai Gow, Emperor's Challenge, and Fortune Pai Gow) listed out seperately, so you can compare the popularity of the product through "Drop" - which is buy-in cash drop into the drop boxes, and "Actual Hold," which is what a casino keeps of that action.
A good table hold for a table game is in the 17% to 26% range or so; above that, players feel their losses really hard, and below that level, the casino laments the poor income produced, unless the game drops a huge amount: a 5% table hold of $1,000,000 drop is $50,000; a 25% table hold of $100,000 in buy-ins is $25,000. Baccarat play affords this.
Table hold is tricky for side bets, because the side bet is played in conjunction with (and overlaid on) a base game, masking the individual side bet performance to a great degree.
You can assume very roughly that one percent of house edge equals a 10% of table hold, for a game between 1% up to 4% HE; It's a remarkable corrolation.
The 10% BJ table hold in NV is because of experienced players making BJ a 1% HE game or so for them. The 18% BJ hold in other areas may be due to poorer play. A LOT of other factors are involved, but it's rough ballpark rule-of-thumb that is used, and that is usable from data of a high-volume extended period.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 10:21:57 AM
permalink
Guess I don't understand the interest in this metric at all. Consider two scenarios:
(1) I go to a table, buy in for $100, place $5 bets for an hour, and leave while down $20.
(2) I go to a table, buy in for $1,000, place $5 bets for an hour, and leave while down $20.
Why should a casino (or a table game inventor) care which one of these scenarios took place? But (as I understand it), the Table Hold Percentage would be different by a factor of ten. Is that right?
(1) I go to a table, buy in for $100, place $5 bets for an hour, and leave while down $20.
(2) I go to a table, buy in for $1,000, place $5 bets for an hour, and leave while down $20.
Why should a casino (or a table game inventor) care which one of these scenarios took place? But (as I understand it), the Table Hold Percentage would be different by a factor of ten. Is that right?
November 9th, 2010 at 10:40:49 AM
permalink
Its further complicated when you arrive at the next table lugging all those chips and start betting them... what happens to the statistics at the new table? That new table didn't sell you any chips but they are filling up their drop box with them.
November 9th, 2010 at 11:05:55 AM
permalink
Doc -
That's EXACTLY what got me scratching my head. I'd assume that they only care about the $20.
Flea -
Chips brought to the table gets recorded on the comp cards.
On a somewhat related note, when I visited the Wynn in September, I steped up to an empty BJ table to purchase a $1 chip. The dealer actually looked to the floor person to get permission. Is it because that $1 skews the table hold?
That's EXACTLY what got me scratching my head. I'd assume that they only care about the $20.
Flea -
Chips brought to the table gets recorded on the comp cards.
On a somewhat related note, when I visited the Wynn in September, I steped up to an empty BJ table to purchase a $1 chip. The dealer actually looked to the floor person to get permission. Is it because that $1 skews the table hold?
I invented a few casino games. Info:
http://www.DaveMillerGaming.com/
Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
November 9th, 2010 at 11:12:29 AM
permalink
Table Hold Simulator for a gaming mathematician
Actually,
This House edge to table hold correlation might be a very interesting "Let's SEE" project for a gaming mathematician such as Stacy or Mike or Charles M.
Is there a correlation between HE and table hold, and can it be simulated??!!
Assume on the base game:
1. 3 player types on Pai Gow Poker table;
2. The house uses the Clarion AC house way (listed in WOO PGP house ways appendix)
3. ONE Strong player: Player type #1 bets: $20 main bet; Player leaves if a) up 300%, b) Busts trying, or c) played the equivelant of three hours or 60 hands ("typical green player") using Clarion House way (Very Good), but plays a straight or a flush with one pair as one pair if the S or F has a J-10 or less top, and the one pair is Queens+ with an AJ or better top, also splits NO two pairs with AK top except Kings and 6+. (very strong player).
4. FOUR Average players: Player type #2 plays $10...leaves if a) makes 200% of buy-in or goes down to 25% of buy-in, or plays 40 hands. Player uses Texas Sation house way against Clarion AC house way.
5. ONE Novice/Drunk player: Player #3 plays $5/$0/$0; Player's stratgey is Novice Pai Gow poker: same as any strategy for pai gows and one pairs; Splits ALL two pairs without an ace, and all face-card pairs; splits all full houses in spite of with low pairs and a AK top; plays all straights and flushes as straights and flushes; doesn;t split three aces; splits all four of a kinds. Leaves after 20 hands.
Session ends after all players leave.
This is a typical real-world session mix of Pai Gow Poker players.
A Simulation can represent 1,000+ session simulations.
Now...For tweaking the house edge:
Analyze the 2.4% game, and see if it produces a 24% table hold:
Assume 0.9% from copy effect on two card side.
Assume 5% commission adds 1.5%, as player's win 30% of the time at 5%, for a 2.4% average house edge of Pai Gow Poker (1.5%+0.9%).
This should have about a 24% hold.
Analyze a 1.5% house edge:
0.9% copy effect + 2% commission = 1.5% house edge; should have about a 15% table hold.
Analyze a 3.9% house edge game:
With a 10% commission: 0.9% copy effect +3.0% (10% * 30% player wins) = 3.9%; should have like a 40% hold.
Then add:
Bonus/Insurance bets.
1. Strong player plays $5 + $5
2. Avg player plays $2 and $1.
3. Nove player doesn't play bonus bets.
and re-run simulation.
This can be tweaked no end.
It would be good for a mathematician to also be able to project/simulate table hold/profit. HAs anything like this been done in table gaming math?
Actually,
This House edge to table hold correlation might be a very interesting "Let's SEE" project for a gaming mathematician such as Stacy or Mike or Charles M.
Is there a correlation between HE and table hold, and can it be simulated??!!
Assume on the base game:
1. 3 player types on Pai Gow Poker table;
2. The house uses the Clarion AC house way (listed in WOO PGP house ways appendix)
3. ONE Strong player: Player type #1 bets: $20 main bet; Player leaves if a) up 300%, b) Busts trying, or c) played the equivelant of three hours or 60 hands ("typical green player") using Clarion House way (Very Good), but plays a straight or a flush with one pair as one pair if the S or F has a J-10 or less top, and the one pair is Queens+ with an AJ or better top, also splits NO two pairs with AK top except Kings and 6+. (very strong player).
4. FOUR Average players: Player type #2 plays $10...leaves if a) makes 200% of buy-in or goes down to 25% of buy-in, or plays 40 hands. Player uses Texas Sation house way against Clarion AC house way.
5. ONE Novice/Drunk player: Player #3 plays $5/$0/$0; Player's stratgey is Novice Pai Gow poker: same as any strategy for pai gows and one pairs; Splits ALL two pairs without an ace, and all face-card pairs; splits all full houses in spite of with low pairs and a AK top; plays all straights and flushes as straights and flushes; doesn;t split three aces; splits all four of a kinds. Leaves after 20 hands.
Session ends after all players leave.
This is a typical real-world session mix of Pai Gow Poker players.
A Simulation can represent 1,000+ session simulations.
Now...For tweaking the house edge:
Analyze the 2.4% game, and see if it produces a 24% table hold:
Assume 0.9% from copy effect on two card side.
Assume 5% commission adds 1.5%, as player's win 30% of the time at 5%, for a 2.4% average house edge of Pai Gow Poker (1.5%+0.9%).
This should have about a 24% hold.
Analyze a 1.5% house edge:
0.9% copy effect + 2% commission = 1.5% house edge; should have about a 15% table hold.
Analyze a 3.9% house edge game:
With a 10% commission: 0.9% copy effect +3.0% (10% * 30% player wins) = 3.9%; should have like a 40% hold.
Then add:
Bonus/Insurance bets.
1. Strong player plays $5 + $5
2. Avg player plays $2 and $1.
3. Nove player doesn't play bonus bets.
and re-run simulation.
This can be tweaked no end.
It would be good for a mathematician to also be able to project/simulate table hold/profit. HAs anything like this been done in table gaming math?
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 11:18:51 AM
permalink
Quote: DJTeddyBear
On a somewhat related note, when I visited the Wynn in September, I steped up to an empty BJ table to purchase a $1 chip. The dealer actually looked to the floor person to get permission. Is it because that $1 skews the table hold?
I think it's because there are things they're supposed to do to "unfreeze" the table if you actually want to play. When I've collected chips from empty tables, I've encountered all sorts of different procedures. If the cards are fanned on the felt, sometimes they need to be moved to the (is there another word for this?) collector on the dealer's right before they can accept cash. If it's a shoe game, I think they need to decide whether or not to reshuffle or burn a card, etc. One time, I was told "no, we own the chips and we keep them all." (@ Silverton Lodge) So I bought in for $40, played 2 hands, walked to the cashier, and "lost" a chip into my pocket on the way.
"So as the clock ticked and the day passed, opportunity met preparation, and luck happened." - Maurice Clarett
November 9th, 2010 at 11:44:15 AM
permalink
Quote: DocGuess I don't understand the interest in this metric at all. Consider two scenarios:
(1) I go to a table, buy in for $100, place $5 bets for an hour, and leave while down $20.
(2) I go to a table, buy in for $1,000, place $5 bets for an hour, and leave while down $20.
Why should a casino (or a table game inventor) care which one of these scenarios took place? But (as I understand it), the Table Hold Percentage would be different by a factor of ten. Is that right?
Yes...but:
They (Casinos) care about profit. And the table hold is a function of amount bet and lost against buy-ins that reflect real-world betting habits that are proportional. If you bought in for $5 billion dollars, and played a few $5 hands and lost only $20, your "cash color-out" of $4,999,999,980 would be subtracted out from the equivelant loss of chips from the rack (assuming $B chips on the table).
Therefore, your buy-in is adjusted by your "chip-out" or color-out that's reflected in the chip rack, that's counted also, so yes, the table hold then would essentially be 0%. This is similar to buying in for $100, but insisting on playing and losing twenty pennies. But this doesn't happen in the real world. When Buy-ins are out of line, real-world limits and diminishing returns are imposed that are reflected in the hold numbers we do see. Let me explain:
What is reflected in the real world is that someone who buys in for $1,000 can be statistically counted on of betting $50 to $100 hand, and so loses on average $200 instead of twenty bucks. This is because people bet in relation to what they buy-in for. If your sitting with $10,000 in chips in front of you, you aren't betting a nickle a hand, and if you were, the pit boss would talk to you or forbid the buy-in in the first place.
The dealer will call out "Change $10,000, but this asshole wants it all in Nickles" (which the rack doesn't even hold), and ther pit boss would say, "Do NOT go ahead, let me chat with this fool!" (Actually, they will practice customer service, and so use different language to indicate that something is wrong.)
Baccarat is a good example. People often buy in for $10,000, bet $500 a hand, and lose $2,000 for that 20% hold. This is what happens. Bet a nickel and you'll get a no-action result from the dealer and floor, and the casino host will call a neurologist to check you out, and that will be the end of that table's action, which doesn't happen too often in the real world. So theoretical examples of extremes are filtered out by real world limits.
You see, in those table hold figures, both universal player habits (of betting to their bankroll) and casino employee override "correct out" aberations.
Table limits kind of enforce this. You do Not go to a $100 table and buy in for $100. If you lost your only $100 hand the table hold from you is 100%, and someone else will win for a 0% hold, and someone will leave with a profit of $100 for a hold of - 100%, but the house edge on the game will enforce the average profit effect using real-world players who bet proportionately, and who can be statistically analyzed. You see a table limit of $100, you walk past it if you don't have $1,000 to easily risk.
If people did act like this, then table hold would not be primarily used, just the cash profit per table per month, - which is also listed, with the table hold a result of that datum.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 11:49:18 AM
permalink
Quote: FleaStiffIts further complicated when you arrive at the next table lugging all those chips and start betting them... what happens to the statistics at the new table? That new table didn't sell you any chips but they are filling up their drop box with them.
Coming to a table with chips is counted as "checks in," similar to a buy-in. Remember, the casino adjusts the buy-in "cash box" numbers with the table's "rack closing amount" numbers, and they see the extra new chips from walk-up players as "table-in dollars" also.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 11:51:37 AM
permalink
On the topic ...
Slot "hold" appears to not mean the same thing as table game "hold." Every time I read something by a pro about slot hold, they always seem to just be referring to the house edge. What is slot "hold?."
..D..itty...
Slot "hold" appears to not mean the same thing as table game "hold." Every time I read something by a pro about slot hold, they always seem to just be referring to the house edge. What is slot "hold?."
..D..itty...
"Who would have thought a good little girl like you could destroy my beautiful wickedness!"
November 9th, 2010 at 11:54:28 AM
permalink
Dorothy,
Slot people use "coin-in" and "coin-out" for slot accounting, with net "coin-in" being the money kept by a machine.
I think "slot hold" is a machine's expected house edge; not sure. I have to open up the "Casino operations" (UNLV textbook) I have sitting in my bookcase.
Slot people use "coin-in" and "coin-out" for slot accounting, with net "coin-in" being the money kept by a machine.
I think "slot hold" is a machine's expected house edge; not sure. I have to open up the "Casino operations" (UNLV textbook) I have sitting in my bookcase.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 11:57:57 AM
permalink
Gee Dan, tell us how you REALLY feel! LOL.Quote: Paigowdan...The dealer will call out "Change $10,000, but this asshole wants it all in Nickles" (which the rack doesn't even hold), and ther pit boss would say, "Do NOT go ahead, let me chat with this fool!"
...and the casino host will call a neurologist to check you ...
(FYI: For those that missed the memo, Dan is not only a game inventor, he's also a dealer.)
But seriously, I occasionally see guys buy in for big bucks, get a rainbow assortment of chips (he's not an asshole asking for all nickels), play a few minutes at or near the table minimum, then walk - sometimes without coloring up.
If it matters, I usually see it at the craps table.
I invented a few casino games. Info:
http://www.DaveMillerGaming.com/
Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
November 9th, 2010 at 11:58:08 AM
permalink
Quote: DorothyGaleOn the topic ...
Slot "hold" appears to not mean the same thing as table game "hold." Every time I read something by a pro about slot hold, they always seem to just be referring to the house edge. What is slot "hold?."
..D..itty...
Yeah, I believe "slot hold" and "table hold" are two entirely different things. And you're right, a "slot hold" is essentially the house edge; i.e. a slot machine that advertises a "93% hold" means the HE is 7%.
As people have explained before, the reason slots do it differently is because they can count each individual bet, which table games aren't usually equipped to do.
By the way, Dan - I play EZ Pai Gow at Ameristar Kansas City, and the game rocks! Stupid Harrah's North Kansas City can take their stupid Fortune Pai Gow and shove it.
November 9th, 2010 at 12:08:30 PM
permalink
Quote: DJTeddyBearGee Dan, tell us how you REALLY feel! LOL.
(FYI: For those that missed the memo, Dan is not only a game inventor, he's also a dealer.)
But seriously, I occasionally see guys buy in for big bucks, get a rainbow assortment of chips (he's not an asshole asking for all nickels), play a few minutes at or near the table minimum, then walk - sometimes without coloring up.
If it matters, I usually see it at the craps table.
Doesn't surprise me, some Craps players are waiting to see that streak start, so they'll buy in big, not like the shooters and then run.
Craps players are an odd and varied bunch.
"Then you can admire the real gambler, who has neither eaten, slept, thought nor lived, he has so smarted under the scourge of his martingale, so suffered on the rack of his desire for a coup at trente-et-quarante" - Honore de Balzac, 1829
November 9th, 2010 at 12:18:41 PM
permalink
As I noted in the thread about my visit to the Silver Saddle, I now have in my collection three chips from casinos where I did not actually play. In each of those cases, it was because the casino was not prepared to offer me a game while I was there, so I just purchased a chip. I have never gone to a table open for business and just purchased a chip. When I start play the first time at a casino, the first $1 chip I encounter goes in my pocket. I didn't always do that, and I had a terrible habit of forgetting to keep my souvenir when I left.Quote: rdw4potusOne time, I was told "no, we own the chips and we keep them all." (@ Silverton Lodge) So I bought in for $40, played 2 hands, walked to the cashier, and "lost" a chip into my pocket on the way.
One possible consideration from the casino might be the cost of $1 chips. I don't really know what these chips cost in bulk. If you had offered to purchase and keep a $5 chip or a $100 chip (which probably cost the casino about the same as a $1 chip), they may have been more interested in selling that to you. Special issue or commemorative chips for collectors are often in $5 denominations, and I think the casinos look at this as a side business. I'm too cheap to keep anything but the $1 ones.
Often, the dealers and supervisor are well aware that I will be walking out with one of "their" chips and don't care, so long as I have given a little action. At least twice, once at the Searchlight Nugget and once at a casino in the Caribbean, a supervisor went out of their way to bring me a newer, fresher chip than the ones they had in play. The Searchlight Nugget $1 chips are metal tokens, and when the supervisor learned I was a collector, he went to the back room and got me a mint chip fresh from the box. But I had at least given a little action and left the place down $10 at a $5 table. (One of my favorite comments is that some of my $1 souvenir chips were very expensive, and some they actually paid me to take out of the place.)
November 9th, 2010 at 1:57:58 PM
permalink
Quote: The LubinThey (Casinos) care about profit. And the table hold is a function of amount bet and lost against buy-ins that reflect real-world betting habits that are proportional.
I agree with your assessment that this is how it IS done, but hold is a truly overrated statistic. It's like taking attendance at a meeting by counting the arms, legs and heads, then dividing by five -- the underlying 'logic' makes sense, but it's unnecessarily convoluted.
In the "good old days", when a pit boss would estimate profitability by the buy-in, there was value in hold statistics. He bought in $1k, the hold is 20%, he'll leave $200 behind at this table, so let's comp him x% of $200 up front. Nowadays, any "comp engine" simply multiplies hands per hour by edge by bet size, and there's no reason a pit boss shouldn't be able to do this too.
How many spots are in play?
How many hands per hour?
How skilled is each player (thus, what is the house edge over this player)?
Take the product, there's your average profit per hour. Next case.
November 9th, 2010 at 2:25:34 PM
permalink
Quote: ElectricDreamsYeah, I believe "slot hold" and "table hold" are two entirely different things. And you're right, a "slot hold" is essentially the house edge; i.e. a slot machine that advertises a "93% hold" means the HE is 7%.
As people have explained before, the reason slots do it differently is because they can count each individual bet, which table games aren't usually equipped to do.
Right - slot "hold" or "win %" is win / coin-in. Table hold or win % is win / drop. Coin-in is an exact measure of total wager volume on a slot machine. Drop is nowhere close for table games. In a table game you have no idea how many times players re-bet the same chips, but in a slot game you do.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice."
-- Girolamo Cardano, 1563
November 9th, 2010 at 2:30:05 PM
permalink
Quote: CRMousseauI agree with your assessment that this is how it IS done, but hold is a truly overrated statistic. It's like taking attendance at a meeting by counting the arms, legs and heads, then dividing by five -- the underlying 'logic' makes sense, but it's unnecessarily convoluted.
Very true, but it is a stat we still use that reflects buy-ins dropped in versus buy-ins kept, and is used for selling and picthing purposes to casinos where no installs have ocurred yet. Stating the hold is a way of stating "this is what it will/should do" - based on what your customers spending volume is in your table games area.
I did notice in a monthly gaming commission spreadsheet that a Baccarat table dropped $900,000 with a hold of only 5% - keeping $45,000, when a Pai Gow table dropped $200,000 and kept 20% for $40,000, - which is less money for the casino for a better hold. And the workload for the dealers was the same, as $1,000 aren't any heavier than $25 chips.
What casinos really look at is the "net cash" per real estate area of the table once it is installed - above and beyond its "table hold performance."
What game designers look at is live table hold performance (assuming it'll stay a desirable product), with the later assumption that'll translate into a profitable performance consistently. We have to do this, because on the drawing board or from the early live data, we don't know who will later buy in for what....so we have assume that'll later hold "in the zone" of the buy ins that it will get. This is also why a mathematician like you should come up with a "hold simulator" for a product your doing the math for, so the parameters can be plugged in (typical players at the casino, house edge, playing style, bet amounts, session duration, etc.) Not easy, but would be awesome to have.
Distributors pitch games based on what its table hold is from earlier installs, so that customers can assume what it will later make for the casino, if and when installed.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 4:51:37 PM
permalink
Regarding simulating the relationship between house edge and hold, it could be done, but it would only be useful in the real world if the simulated players exactly matched real players. Nobody is going to know the exact mix of players they will get. It would also depend on other factors like day of the week, reason the player is there, and maybe the weather.
I predict in the future every bet made will be tracked by camera and table sensors. We'll know both then.
I predict in the future every bet made will be tracked by camera and table sensors. We'll know both then.
"For with much wisdom comes much sorrow." -- Ecclesiastes 1:18 (NIV)
November 9th, 2010 at 5:09:36 PM
permalink
Quote: Wizard
I predict in the future every bet made will be tracked by camera and table sensors. We'll know both then.
Perry Stasi, the man behind the FireBet in Craps, has a company that is doing just that and more.
His site:
https://www.visualimits.com/index.cfm
7 winner chicken dinner!
November 9th, 2010 at 7:05:24 PM
permalink
Quote: WizardRegarding simulating the relationship between house edge and hold, it could be done, but it would only be useful in the real world if the simulated players exactly matched real players. Nobody is going to know the exact mix of players they will get. It would also depend on other factors like day of the week, reason the player is there, and maybe the weather.
I predict in the future every bet made will be tracked by camera and table sensors. We'll know both then.
True - but a forecast or a projection from a table game hold simulation is simply that - a forecast that may gives leads as to performance when compared or applied to the mix of gamblers you have. No! - It doesn't have to be 100% accurate - it has to reveal general trends that weren't anticipated or seen without the forecast being run.
And it can indeed be based on the typical parameters of a specific casino - which they already know well. In other words - casinos DO know the mix of players they're gonna get, and to a very accurate degree! They know what they play, when they play, what they buy in for, and what they bet! It's their marketing department's business, and their table games Manager's business - and they consult with each other every day on it. We know how many tables to open and when, and at what limits. What we'd like to know now is that if we tweaked the mathematical or procedural parameters of a game, or brought in a new version of a game, - how would that generally affect the forecasted table income? You can even include a projected player "like or dislike" factor by increasing number of players, hands played, or bet size changes based on what an experienced shift manager anticipates, with various scenarios. Of course include day of the week, the season, and the weather - which is also forecasted for us.
A Forecasting program is simply that: it produces a forecast based on assumptions and within limits, and not as a guaranteed 100% prediction of the future's actuality. If reveals trends, and thereby revealing potential income, and problems to be ready for or to anticipate that we couldn't anticipated without it.
It runs "what if" scenarios that provide predictions based on parameters, and these theoretical scenarios can provide "idea leads" or anticipate or reveal unanticipated problems, - all the while not being an "actuality guarantee." We know that going in.
I've written college enrollment forecast programs based on certain student characteristics: attrition rate, number of students in what majors and at what levels, an anticipated percentage of transfer coming into the college at which levels and disciplines, etc., and it gave pretty damn good estimates as to what size classroom should have which course offered there, what tuition revenue would be on that group or in certain disciplines, etc. I also wrote a system that examined different professors' grades against the same group of cohorts to spot "easy A" teachers and the "tough bastard" teachers, within standard deviations, to inform the Dean who is slack and who is not. We could raise and lower forecasted departmental GPA's by tweaking the professorial mix. Used to be my career, before teaching high school math, and before joining the gaming industry. I Can See This As A Good Gaming Application. Casinos groups would use it like crazy to forecast slot and table game product and offering procurements, in terms of "will this product work with our mix, and what might it make."
Gamblers, like college students, exhibit predictable group behaviors that can be used in forecasts.
Beware of all enterprises that require new clothes - Henry David Thoreau. Like Dealers' uniforms - Dan.
November 9th, 2010 at 8:39:31 PM
permalink
Its often easier to buy a chip as a momento from a busy table. I've seen dice dealers hand a white chip directly to a buyer and take the dollar bill from her hand. Nothing about this putting it on the table or clapping his hands or anything. Going up to a BJ table that was manned but devoid of players might actually take you more time but usually if its just buying a chip to keep as a momento... say so right out and get the transaction done with. They won't be worrying about their hold statistics. They will sell you a chip if you want one and get on with their work.Quote: rdw4potusWhen I've collected chips from empty tables,
December 4th, 2010 at 9:51:43 AM
permalink
Quote: WizardRegarding simulating the relationship between house edge and hold, it could be done, but it would only be useful in the real world if the simulated players exactly matched real players.
Mr. Wizard - I'm looking for an educated response to this question on another thread as well: If Roulette's house advantage is 5.26% and Pai Gow Poker's is roughly 1.75% how can Pai Gow Poker possibly hold over 20% while Roulette only holds 17% (I was surpirsed at how low the Roulette hold % was in NV)? Those figures were from a link on the other thread showing NV holds for past 12 months. The HA on Roulette is 3xs that of PGP. Why is the relationship of hold percentages not even close to that of the HAs? I understand Roulette won't hold 3xs that of PGP but how can it possibly hold LESS? The other thread is "does this make sense" under math. Sorry I'm not savvy creating links. I'm old school - pencil and paper. By the way I've used and loved the wizard of odds site for over 10 years.Don't get me started on House Advantage vs. Hold Percentage!
December 4th, 2010 at 10:13:58 AM
permalink
Although hold % is not a very scientific number, it is as someone mentioned earlier, the only thing we've got. There are a lot of variables that can manipulate this number but over the course of several years it should give us at least an accurate way to compare the performance of different games.
Here's a real life example of why, in the short term, hold % is misleading. I had a player buy in for $2000 on a BJ table with limits of $25-$2000. I later learned that was all the cash he had brought that day therefore the most the house could win was $2000. The player caught a monster run and in less than 90 minutes cashed out $52,000. I offered him dinner at our steakhouse and he accepted. About 2 hours later he returned and eventually bought back in for all $52,000 and lost it. Had he lost the $2000 during that first 90 minutes of play the hold would have been 100%. Instead his buy in was $54,000, $52,000 of it was "false drop" as we call it, and we won the same $2000 but our hold was only a lousy 3.7%.
But again, over the course of several years this scenario will seldom occur and the hold % will give us a good idea how much of the "drop" we can expect to win.
Here's a real life example of why, in the short term, hold % is misleading. I had a player buy in for $2000 on a BJ table with limits of $25-$2000. I later learned that was all the cash he had brought that day therefore the most the house could win was $2000. The player caught a monster run and in less than 90 minutes cashed out $52,000. I offered him dinner at our steakhouse and he accepted. About 2 hours later he returned and eventually bought back in for all $52,000 and lost it. Had he lost the $2000 during that first 90 minutes of play the hold would have been 100%. Instead his buy in was $54,000, $52,000 of it was "false drop" as we call it, and we won the same $2000 but our hold was only a lousy 3.7%.
But again, over the course of several years this scenario will seldom occur and the hold % will give us a good idea how much of the "drop" we can expect to win.
Don't get me started on House Advantage vs. Hold Percentage!
December 4th, 2010 at 10:18:08 AM
permalink
Expected value is the result from in single play.
Hold is from multiple plays, players and sessions. How many here in a session, how they bet and what they bet with all varies.
Hold is from multiple plays, players and sessions. How many here in a session, how they bet and what they bet with all varies.
"Then you can admire the real gambler, who has neither eaten, slept, thought nor lived, he has so smarted under the scourge of his martingale, so suffered on the rack of his desire for a coup at trente-et-quarante" - Honore de Balzac, 1829
December 11th, 2010 at 9:54:01 AM
permalink
Coming to a table with chips will not have them counted as a buy in for hold purposes. It happens occassionally where a game can hold over 100% for the day due to a player coming to the game with a large amount of chips and losing and then the game having very little other action the rest of the day.
Don't get me started on House Advantage vs. Hold Percentage!
January 19th, 2015 at 6:11:10 PM
permalink
Sorry for the bump, but I felt it better than starting a new thread.
Question for Dan. You say that hold percentage usually turns out to be about 10x the house advantage. Any idea why that is? Does a player's bankroll turn over about 10 times on average before they're done? (I.E. Someone who buys in for $100 will place about $1,000 worth of bets?)
Also, what about games with house advantages higher than the 4% you mentioned, like Lucky Ladies? Do their holds trend towards 100%?
Question for Dan. You say that hold percentage usually turns out to be about 10x the house advantage. Any idea why that is? Does a player's bankroll turn over about 10 times on average before they're done? (I.E. Someone who buys in for $100 will place about $1,000 worth of bets?)
Also, what about games with house advantages higher than the 4% you mentioned, like Lucky Ladies? Do their holds trend towards 100%?
Casinos are not your friends, they want your money. But so does Disneyland.
And there is no chance in hell that you will go to Disneyland and come back with more money than you went with.
- AxelWolf and Mickeycrimm
