What is the Actual House Edge on Blackjack?

Let us assume a good (for the player) 6 shoe game - 0.26% house advantage with perfect basic strategy

Now, let's factor in all the drunks, stupid people, non perfect basic strategy players, cheats, card counters, new players, etc.

What do you think the actual house advantage is for blackjack? Are they returning 1%, 2%, 3%, more?

I have read somewhere that even on 100.7% expectation video poker, the house still has a 3% house advantage on average due to all of the sub optimal play.

For blackjack I will go middle of the road and guess 1.5% advantage for the house on average. Your thoughts?

You forgot to include the idiots that play BJ for 6:5 when a 3:2 is on the next table, etc.

[edit] -- deleted, did not understand OP's question.

--Dorothy

Hmmm... it depends on the distribution of the players. Out of a group of 100 players, I would say only 5% know perfect basic strategy. They get the .26% HA.

Another 20% are good strategists. They make a few errors, like standing a soft 18 on a 10, standing on a 12 vs a 3, and make a few iffy doubling and splitting decisions (Splitting 4 on 4, not splitting 9s on a 9, not doubling a soft 15 on a 6, etc). Let's say they make 10 errant basic strategy decisions consistently that costs them 10% of the EV vs the correct decision. Multiply this by the % of hands that this occurs (there are 26 different decisions and 10 dealer positions, so getting six wrong probably means the combination where the wrong decision is made on 4% of hands. So, .10 x .04 = .004. They get the .66% HA.

Another 50% make alot of mistakes but know some strategy, like to make a decision based on the dealer's up card. But they will always stay on 12 on a dealer 2 and 3, they will make bad splitting a doubling decisions. Say, out of the 260 combinations, they make bad decisions on 10% of hands which costs them, say 20% of EV (.10 x .20 = .02). They get assigned a 2.26% HA.

The other 25% are out to lunch and have no idea what they are doing. According to the Wizard, they follow a strategy like "Mimic the Dealer" or "Never Bust". They get assigned a HA of 5%.

So, when you put it all together .26% x 5% + .66% x 20% + 2.26% x 50% + 5% x 25% = 2.525%.

That's my guess anyway, about 2.5%. The thing is that a drunkie is far more likely to sit at a 6/5 table than stumble into a game with a .26% HA.

Nevada publishes something every month called a "Win percent" for every game. Sometimes the win percent is called the casino hold. I will quote the wizard here:

The house advantage is the percentage you will lose, on average, of each dollar bet. The hold is the ratio of money the casino wins to chips purchased. This is going to be much higher than the house edge because in table games players circulate through the same chips for a while.

This percentage is larger than the house edge even considering wild unorthodox play, because the winnings are circulated back in. Sometimes when it goes up it is an indication that people are staying at the tables longer.

The win percent for the last year for blackjack in all of Nevada was 11.38%.

I don't think there is any way to actually calculate the true house advantage which is what you call house edge including all the stupid plays. No casino keeps track of the number of times chips get re-circulated. It would be possible if chips were replaced with electronic systems. However, the few systems (like Shufflemaster) that have all electronic systems don't publish that data. The Wizard has done some calculations to account for various bad strategies like mimic the dealer, or never double, or never split anything other than aces.

One of the most unlikely strategies I have ever heard was reported by a dealer in Ask the Wizard . The player never looked at his cards playing blackjack....just tucked them. The wizard calculated that under typical Vegas rules (6-deck, dealer hits soft 17) the house edge by always standing is 15.7%. But in the real life situation this player sometimes walked away ahead. It just goes to show you that empirical observations are almost meaningless.

Peter Griffin devoted a whole chapter to the question of the cost of playing errors in his book 'Extra Stuff -- Gambling Ramblings.' His study was based on observing 11,000 actual hands of play in 1987. The following table summarizes his results of the cost of errors made.

City	Cost of Errors	Margin of Error
Atlantic City	1.13%	0.12%
Las Vegas	1.67%	0.17%
Reno	1.48%	0.19%
Lake Tahoe	1.39%	0.54%
Total	1.41%	0.10%

In my opinion, play has improved a lot in the 23 years since the study. If forced to guess, I think the cost due to errors is only about 0.5% now, based on total amount bet. The smaller the bet, the higher it tends to be. Griffin standardized the bet size in his study. I would agree with Griffin that Atlantic City players are more skilled than Vegas players.

Quote: Wizard

If forced to guess, I think the cost due to errors is only about 0.5% now, based on total amount bet. The smaller the bet, the higher it tends to be. Griffin standardized the bet size in his study.

I don't know Wizard. If what you say is true, then a good blackjack table (0.26% house edge) only has a 0.76% house advantage when you factor in errors. While I agree with you that a $5 per hand player is likely to be less skilled then a $100 per hand player, I have seen more then my share of bad "high rollers" out there.

On a side note, let's say you were a god awful blackjack player. Would you get more comps then a good basic strategy player with the same per hand play? Or would they just rate you as the same $100 a hand (for example) player.

Quote: gambler

I don't know Wizard. If what you say is true, then a good blackjack table (0.26% house edge) only has a 0.76% house advantage when you factor in errors. While I agree with you that a $5 per hand player is likely to be less skilled then a $100 per hand player, I have seen more then my share of bad "high rollers" out there.

On a side note, let's say you were a god awful blackjack player. Would you get more comps then a good basic strategy player with the same per hand play? Or would they just rate you as the same $100 a hand (for example) player.

I know of one major casino in Vegas that assumes a 0.75% house edge in blackjack (source). Griffin also said about 5% of hands were not played correctly. It seems to me that most errors are relatively minor ones, like standing on 12 against a 2, and/or don't happen often, like not soft doubling when you're supposed to. It isn't just blackjack, players are more skilled across the board than they used to be. It is much easier to get good information about gambling. Hopefully I've played a part in that.

To answer your question, most casinos generally don't take skill into consideration. Jim Kilby's 'Casino Operations Management' shows a rating slip with three levels of skill, but everyone I speak with says they just go by average bet and hours played. There would be exceptions for whales, where play gets watched more carefully. Indeed, they could expect to be treated better if they played badly, but it would likely be an informal adjustment.

However, mark my words, the day will come where every bet and every mistake will be recorded, and your exact skill level will be known. This will be done by sensors for the bets, and either card readers or cameras for the cards. This will be true for all games of skill. Milk the comps now, while you still can.

Quote: Wizard

I know of one major casino in Vegas that assumes a 0.75% house edge in blackjack (source). Griffin also said about 5% of hands were not played correctly. It seems to me that most errors are relatively minor ones, like standing on 12 against a 2, and/or don't happen often, like not soft doubling when you're supposed to.

That's an interesting statistic. Even allowing that 0.75% is for traditional blackjack, and must be increased to cover 6:5, superfun, and other variations, that means that each chip is played on the order of ten times to get the win percent reported in the NGCB publications which is currently 11.38% statewide.


I wouldn't have though that people replay their chips that many times.

Side Question Players must be passionately attached to chips. You would think that casinos could still play with dealers, cards, felt and dice, but replace chips with electronic devices that would allow them to easily collect data, distribute comps, look for cheating, and check disputes. I must assume that all experimental moves in this direction have met with dismal failure.

They might have a higher theoretical for 6 to 5 blackjack, but I'm not sure. Perhaps croupier can tell us.

Player tracking will be able to be done with chips and live dealers. Sensers under the felt will be able to read bet amounts. Shufflers can read cards as the come out, but something else would have to indicate which player received it. Such technology was displayed at the Global Gaming Expo, but I heard it was still buggy. Just give it 5-10 more years.

Hi Wizard, What do think a Players edge could be if they were a Good card counter with a perfect Strategy? I know it would depend on the houses rules. But just a rough average with basic BJ rules. Thanks

I think Hunches cost more than mistakes, at least in my case they do.

I know basic strategy like the back of my hand, but I still deviate once in a blue moon if I get that "Feeling".

One time a dealer pissed me off at Binion's Horse Shoe, (not by doing anything other than beat my pants off) so I went on a quest to kick her butt and left the table $300 up. I broke all the rules.

But of course, that was the exception to the rule, but it sure was fun.

What about the flip side... by how much do you think the house advantage goes down due to dealer errors?

Quote: jeremykay

What about the flip side... by how much do you think the house advantage goes down due to dealer errors?

I tend to think it goes up because of dealer errors. In my opinion, about 80% of errors are in the dealer's favor. Of course, those are more likely to get challenged. Let's say players catch 60% of errors either way, but say nothing when it goes in their favor. So given there is an error, the probabilities are:

20% = Player's favor, player keeps his mouth shut.
48% = Dealer's favor, player opens his mouth.
32% = Dealer's favor, player didn't catch it.

Of course these are rough estimates.

Is that your hunch that 80% are in the houses favor or have you heard or read that figure? Just curious. I was at a roulette table one time and a new dealer was giving out 6 stacks of 20 chips when someone bought in with a $100. I bet he gave out $160 in free chips before the pit boss caught him.

Yes, that is my position, based on 23 years of playing. I think that is because:

1. Dealers also know the players are likely to speak up if a mistake goes against the player, giving them a safety net in that direction.

2. If the mistake goes against the house, the players won't give the dealer a chance to correct it. If surveillance catches the mistake, the dealer will have to answer for it later.

I'm sure the dealers want to deal a perfect game, but the consequences of mistakes against the house are greater, so they likely err on the side of the house more.

Old conversation, but what do you guys think the true house edge for other games are, namely video poker? Seems even an expert (I'm thinking of those that go through thousands and thousands of hand when the jackpot is right) would have a really hard time of not making a mistake and each individual mistake could be very costly.

Quote: johnbarley

Old conversation, but what do you guys think the true house edge for other games are, namely video poker? Seems even an expert (I'm thinking of those that go through thousands and thousands of hand when the jackpot is right) would have a really hard time of not making a mistake and each individual mistake could be very costly.

I'm sure everybody makes mistakes either due to mental lapses, visual failure, or machine/hand failure. The game will have a big role, too. For example, the strategy for jacks or better is much easier than the strategy for many of the bonus poker variants.