applying math to blackjack

Insurance and Even Money

EV, odds, and probabilities tell the story about one of blackjack’s more controversial subjects — whether insurance and even money are good bets. More specifically, is the payoff for these justified given the risk?

Let’s start with the right answer:

No, for insurance
No, for even money when the blackjack payoff is 3:2
Yes, for even money when the blackjack payoff is 6:5

There are two ways to look at insurance. First, insurance is a side bet that the dealer has a 10-point card in the hole with a two-to-one payoff. Since there are sixteen 10-point cards in a 52-card deck, the probability of there being a 10-point card in the hole is 30.7% (16/52); the “true odds” of this are 36 to 16 or 9 to 4 or 2.25 to 2. This means that the casino is paying $2 when they should be paying something close to $2.25 for every $1 bet. This $0.25 difference between the payoff and the true odds represents an 11.1% advantage for the house ($0.25 / $2.25).

Second, if we want to weigh the payoff / bet with the probabilities of winning / losing — the way Expected Value (EV) is calculated — the casino’s advantage with insurance is 8% as follows:

EV = (Payoff  x  Probability of Winning)  -  (Bet  x  Probability of Losing)
EV = ($2 x 31%) - ($1 x 69%) = $0.61 - $0.69 = -$0.08

The blackjack game itself has only a 1/2 of 1% house edge; the house edge for the insurance bet is 8% — 16 times worse.

The insurance bet gives the house a huge 8% edge over the player.

Most players wisely refuse insurance unless they have a big bet on the table then all logic is tossed out. Somehow the idea that the player should insure a big bet has become a blackjack tradition even though the insurance bet is a loser no matter how much money is at stake.

The math for even money is a little more complicated. Basically, if the player accepts even money, he is going to receive a one-for-one payoff for his blackjack. If the player declines even money, then 69% of the time he is going to get a 3:2 payoff (or a 6:5 payoff) for his blackjack since the probability is 69% that the hole card is NOT a 10-point card, and 31% of the time the player is going to get nothing since the probability is 31% that the hole card is a 10-point card.

This is quite a mouthful and a brain-twister, but examine the following chart and you’ll see the sense of it.

With a 3:2 blackjack payoff, the EV for declining even money is about 4% higher than accepting it, therefore the player should decline even money.

 The EV for declining even money with a 3:2 blackjack payoff is 4% more.

The EV for the 6:5 blackjack payoff is detailed in the following table.

With the 6:5 blackjack payoff, the EV for accepting even money is about 20% higher than declining it ($.17 / $.83), therefore the player should accept it when it’s offered. (This option is rare since casinos are not stupid.)

The EV for accepting even money with a 6:5 blackjack payoff is 20% more.

Note that the payoff and the bet in these tables is stated as a “betting unit,” which is a way of expressing the monetary value of the bet and payoff without referring to a specific amount. Multiplying betting units by the actual bet gives the monetary results.

Also note that the odds for even money are calculated based on 15 (rather than 16) 10-point cards and 49 (rather than 52) other cards left in the deck. This is because at least one 10-point card is in the player’s hand (for him to have blackjack) and there are at least three cards on the table (assuming only the player and the dealer are playing). The calculation changes slightly with more decks and with other player’s cards on the table but not enough to swing the results in another direction.

Most players take even money claiming that “it’s foolish to risk a push and receive nothing for their blackjack.” They are convinced it’s always the right thing to do. Players who refuse even money, as they should with the 3:2 game, are therefore open to criticism as their refusal seems like a rookie mistake. The best tactic when this happens is to smile and nod in the like you just fell off the turnip truck. Trying to convince a player that he’s wrong in the middle of a blackjack game is futile.

Calculating the Profit and Loss Range

Luck is unpredictable; however, it is possible to use a formula that employs “the standard deviation for blackjack” to calculate the probability of a profit or loss. The formula for maximum loss is:

Maximum loss (with 95% confidence) =

[(Standard deviation x 2 {for 95%}) x
(Square root of the number of hands played) x
(Average bet)] -
[((Average bet) x (Number of hands played) x (House edge)) +
((Average bet) x (Number of hands played) x (Basic strategy handicap))]

The formula for maximum profits is the same except the amount attributable to the house edge and the player’s basic strategy handicap is subtracted rather than added.

Let’s use some real numbers to show how it works:

Standard Deviation = 1.17
Number of standard deviations for 95% confidence = 2
Hands Played per Hour = 55
Number of Hours of Play = 3
Average Bet = $20
House Edge = 0.5%
Basic Strategy Handicap = 1.5%

Plugging these into the formulas for maximum loss and maximum profit, we get:

Maximum Loss = $667
Maximum Profit = $535

In other words, a player betting an average of $20, at a table that plays 55 hands per hour, for three hours, where the house edge is 0.5%, and his basic strategy handicap is 1.5% has a 95% probability of losing no more than $667 and winning no more than $535.

In addition, we can also say with 95% confidence that if this player brings a bankroll of $667 to the table, he will NOT go broke in this 3-hour game.

Luck and the House Edge

Many players believe that the house edge and not luck is primarily responsible for their losses, that somehow the casino is manipulating the game to win. This is not true on two counts.

First, that part of the house edge that’s due to the game’s rules is typically 0.5%. Even when the rules are heavily in the casino’s favor, such as with a 6:5 blackjack payoff, which adds about 1.39% to the house edge, the total advantage is still less than 2% or $20 per $1,000 bet. If the $20 bettor’s P/L Range is $655 to $523, a $20 loss due to the house edge is insignificant.

Second, the house advantage due to imperfect basic strategy play can range from 0% to 10% or more, depending on the player’s basic strategy skill. The typical penalty for imperfect play is about 2% among experienced players. Even if this typical amount is combined with a 2% house advantage due to the rules, the average loss is still only 4% or $40 per $1,000 bet compared to the $667 loss that’s due to luck.

The numbers that support the statement that “the casino always wins because they have the odds advantage” just don’t add up.

The Law of Large Numbers

Most of what’s been said so far about luck applies to the short term (for example, a single blackjack session, one evening of blackjack play). What can the player expect in the long term? Here, it’s almost all good news. Like most things in life, if you wait long enough, things even out. So it is with blackjack — over a long period, wins and losses due to luck will tend to cancel each other out, and what remains are only the losses resulting from the house edge and imperfect blackjack play.

 The more you play, the more you will approach breakeven.

It’s really just common sense. No one has bad or good luck forever. Statisticians call this effect the Law of Large Numbers which is a fancy way to say, “things average out over time.” The law states “that as a sample size grows, the mean gets closer to the average of the whole population.” For blackjack players, the “sample” is the blackjack session, the “population” is the complete history of the player’s wins and losses, and the “expected average of the population” is the average of all wins and losses less the sum of all losses due to the house edge.

The following diagram illustrates this. The blue bars represent the imagined profits or losses of a succession of blackjack sessions. The green line is the average of all profits and losses. The red line is cumulative loss due to the house edge.” Over time, the red line and the green line will converge.

This is nice to know but most blackjack players don’t track their lifetime results. It’s comforting in the abstract that profits and losses converge on breakeven, but that does little to ease the gloom that follows a recent loss. Additionally, no one knows when breakeven will happen. Like all statistics, the Law of Large Numbers is a general principle. However, understanding it helps the player put short-term effects like streaks into perspective.

Winning and Losing Streaks

Regrettably, luck is not consistently kind to blackjack players in the short term. The regular pattern of wins and losses depicted in the last diagram is rarely seen in the real world. More likely are prolonged periods of alternating wins and losses punctuated by “streaks” as illustrated below.

Anyone who has ever played the game has run into streaks. They are exhilarating when you’re winning and devastating when you’re losing. In fact, most players are astonished by the phenomenal staying power of blackjack’s losing streaks. “Astonished” is the right word as these losing streaks seem to go on forever, so long that players often rant about their “mathematical impossibility” and the “undeniable proof” that some higher power is involved with their game.

Let’s deal with this ranting first. It is not just the player’s imagination — players do have more losing streaks than dealers; however, a higher power is not causing them. As much as we would like to blame our losses on a tyrannical blackjack god, this is not the case no matter how long the losing streak lasts. The reason for a player’s long losing streaks is much less dramatic. It’s because of the way the game is structured — a blackjack player wins only four hands out of ten (42%) on average, which is offset by the player winning larger bets. When these are averaged out, the house is left with a small house edge of about 0.5% (best case).

 Players lose more hands but win larger bets.

Why is this? Why do dealers win more hands and players win bigger bets?

One big reason is that the dealer has a much better chance of winning the hand by playing after the players since the players will often bust even before the dealer is required to take a card. Other reasons are that players win big (a 3:2 payoff) when they have blackjack, when they can double down (a 2:1 payoff), and when they can split pairs (potentially an enormous 8:1 payoff by splitting and doubling). The player even has an advantage when he has a terrible hand by surrendering (a 1:2 payoff). The important conclusion that comes out of this is that a blackjack player must have patience and wait until he has the advantage and then maximize his bet.

Right.

This is easy to say but hard to do. The player who is in the middle of an epic losing streak is simply not psychologically prepared to bet the farm when opportunity knocks. Many players, even experienced ones, often hesitate at these critical moments and refuse to double, split, or surrender because they think “luck is against them.”

Don’t fall into this trap; luck has no mind and no memory.

Maximize every advantage by doubling, splitting, and surrendering.

Again, this is easy to say but hard to do when you are losing your shirt. Which is why players change tables, change seats, skip hands, leave then return to the casino, and practice any number of other voodoo tactics. We all do it, it’s human nature; we all have an instinctual “faith” in the supernatural power of luck.

Which leads us to the Gambler’s Fallacy.

The Gambler’s Fallacy

The gambler’s fallacy is the belief that what has happened in the past affects the future. For example, most people think that if you flip a coin and get five tails in a row, the chances of getting another tail are lower. This is just not true; the probability of flipping the coin the sixth time and getting a tail are the same 50% they were on the first flip. Why would they change? Who would change them?

Whatever. It doesn’t matter how often or how many ways it’s said, we are all influenced by our belief that there is balance in the universe. We are all also confused by perspective. When we are at the sixth flip, we are looking back and saying, “the next flip has got to be heads.” It would be a different story if we were at the first flip and looking forward.

For example, the probability of tossing five tails in a row is only 3% (calculated by multiplying the probability of each flip together [0.5% x 0.5% x 0.5% x 0.5% x 0.5%]) even though the probability of tossing the coin on the sixth try and getting tails is 50%. The same is true with blackjack. A player who loses five hands in a row has no better chance of winning the sixth hand than the first. However, looking forward, the odds against losing five hands in a row with a 0.5% house edge are about 30 to one.

More to the point, a player who is losing nine out of ten hands for hours “just knows” that his luck is about to change or that it is going to remain bad. It’s only common sense, right? Wrong. Neither of these is true.

Players should certainly take a break or leave the casino when they are tired, discouraged, or broke, but not to change their luck.

Play as if Luck has no mind and no memory even if you don’t believe it.

Players try to affect their luck in many ways. One of the most appealing is to use a betting system.

Betting Systems

Betting systems are methods of betting designed to minimize losses and maximize winnings.

The popular Martingale system, for instance, requires the player to double his bet after each losing hand and to cut his bet in half after each winning hand. For example, the player starts with a $100 bankroll and loses the first bet of $1, reducing the bankroll to $99, and (per Martingale) requiring him to double his bet to $2. This second bet of $2 is also lost, reducing his bankroll to $97, and requiring him to increase his bet to $4. This third bet is lost, reducing his bankroll to $93, and requiring him to double his bet to $8. This fourth bet is won, increasing his bankroll to $101, and requiring him to halve his bet to $4, and so on. 

This approach to betting allows a player to systematically bet more to recover losses and systematically bet less to preserve winnings. It works well until the player reaches the limit of his bankroll or he encounters the table’s maximum bet limit then it’s all over. This is the problem with all betting systems. They eventually reach limits that cause them to fail. More importantly, they have no effect on luck or on profits.

Betting systems don’t work to change a player’s luck or increase profits.

This is not to say that players should not vary their bets. Part of the fun of blackjack is trying to beat the odds, which will certainly happen more quickly if the player is lucky enough to have large bets on winning hands. Of course, the opposite is true as well—the player is going to lose more quickly if he is unlucky enough to have large bets on losing hands. Again, it is a matter of luck.

playing the 6:5 blackjack premium game

Many casinos offer a 6:5 blackjack payoff at $10-minimum tables and a 3:2 payoff at $25-minimum tables. Since the 6:5 payoff increases the house edge from 0.5% to about 2%, some players think the 6:5 game is for suckers. We don’t see it in such black and white terms.

Sure, it’s harder to win the 6:5 game, and of course players should favor the 3:2 game whenever its possible and practical, but there are other considerations like $10 versus $25 betting minimums.

In our view, a player should play the 3:2 game even when this game has higher minimums if and only if he can a) afford the higher potential loss, and b) he can play a perfect basic strategy game. If he cannot afford the potential losses, then he should go ahead and play the 6:5 game -- if he plays a perfect basic strategy game -- with the understanding that he will lose slightly more often. If he cannot afford the losses and he cannot play a perfect basic strategy game, then don’t play; it’s no fun losing, which is what’s going to happen.

This highly subjective recommendation -- which the player may choose to ignore if he has no other options, like being trapped on a cruise ship -- is informed by the typical numbers shown in the following table. Given  our goal of having fun by winning as often as we lose, it’s the only recommendation we can make.

We’ve picked $10 and $25 as the representative minimums for the 6:5 and 3:2 blackjack payoff games because this is what we see in most casinos. Some casinos offer the 3:2 game with $15-minimum tables. In our view, this $15 game would be the better choice over the $10-minimum 6:5 game even assuming the player can play a perfect basic strategy game. Also, we’ve assumed that the player is betting the minimum.