[ In Brief … ]Blackjack extends beyond strategy into psychology and neuroscience. Research shows how optimal decisions were mathematically defined, how learning algorithms model play, how the brain processes risk, and how problem gambling differs biologically from casual play. Understand decision-making, recognize addiction risks, and treat gambling as an activity you can always walk away from. Blackjack has been studied by science for many reasons, including gambling addiction.If you have read any books on Blackjack or followed some of the better blogs out there, you know that all the blackjack strategies have been done to death. And yet though the odds and probabilities have all been thoroughly documented and discussed there is another aspect to playing Blackjack that hasn’t been so deeply investigated by the pontiffs and prognosticators of pop gambling culture: the psychology of the players.I thought it would be interesting to check out some of the research that looks at how Blackjack players think or how the game of Blackjack can be used for modeling algorithms. You don’t have to understand all this stuff to appreciate it. I don’t even expect you to read all these articles. Scientific papers, after all, can be tiring to read if you don’t spend much time reading them. But the idea that Blackjack is a topic of intense scientific interest just appeals to me.We’ll take them in approximate chronological order.“The Optimum Strategy in Blackjack” was published in The Journal of the American Statistical Association in September 1956. This is the earliest research on the subject that I was able to find but I would not be surprised to learn there were earlier papers. This paper has been quoted and copied many times over on the Internet, and maybe in a few Blackjack books, too. My link points to a .PDF file of the original article.If you are curious about what the paper has to say, be warned: it is loaded with mathematical formulae for computing probabilities and distributions. This is not some Blackjack guru’s advice column from the 1950s. It’s a serious look at what happens when a deck of cards is used to play the game of Blackjack. A lot of this information has been condensed and reorganized for the lay reader in other sources, but it divides the player’s strategy into three questions that must be answered:When to draw cards or standWhen to double downWhen to split pairsUnder the “Drawing Strategy” section the authors note that when you have a soft hand (Ace and a non-Ace card) you have to use an alternate strategy. Their basic proposition is that if your total unique hand (no splits or soft cards) is less than the dealer’s face up card, you should draw until your hand exceeds the value of the dealer’s up card.With a soft hand, you simply use the larger of your two possible totals (count the Ace as an 11, not a 1). The paper mentions alternative strategies that had been proposed by then previous authors.Under the “Double Down Strategy” section the authors include a small table that breaks out the patterns. I can simplify this by summarizing as follows: If your total unique hand equals 9, 10, or 11 and the dealer’s up card is a 2 through 6, double down. Also, if your total unique hand equals 10 or 11 and the dealer’s up card is 7 through (1 less than your total) then double down.For soft hands where your total is 13-18 and the dealer’s up card is 5 or 6, double down. If you have 18 total and the dealer shows a 4, or if you have 17 total and the dealer shows (3 or 4), or you have a total of 12 and the dealer shows a 5, double down. The authors noted that their strategy suggests doubling down far more often than the experts of their day.Under the “Splitting Pairs Strategy” section the authors use another small table to summarize the player’s best options. I will rewrite the table here:WHEN TO SPLIT IN BLACKJACKPlayer HoldsDealer Up CardAces or 8s*any card*9s2 thru 6, 8, or 97s2 thru 82s, 3s, or 6s2 thru 74s55s or 10s*any face card* “Blackjack as a Test Bed for Learning Strategies in Neural Networks” was published in 1998 in Neural Networks Proceedings. Learning algorithms process many choices in using a data set, looking for the most optimum sets of choices that produce a desired outcome. This paper looks at how a learning algorithm can develop its own strategy for playing Blackjack. What makes Blackjack a useful exercise or “testbed” for programmers who develop learning algorithms is that they can compare their algorithms’ performance to the already well-established strategies (such as the 1955 strategy discussed above) that have been reached through human computation.There is virtually no mystery about how good the algorithm is because you can easily compare its choices about when to draw, double down, or split to the choices laid out by other strategists.“An electrophysiological analysis of coaching in Blackjack” was published in the Oxford journal Cerebral Cortex in January 2007. The authors wanted to determine if coaching behavior elicits the same physiological responses from Blackjack players as when they realized they made mistakes and adjusted their choices on their own. But the study was looking at how the coaches reacted when their advice was rejected. In other words, do people who give advice learn from the mistake of giving rejected (though not necessarily bad) advice? The research suggests that advice-givers do learn from these experiences much the same way as people learn from analyzing their own mistakes.“Decision-Making in Blackjack: An Electrophysiological Analysis” was published in Cerebral Cortex in May 2007. This article falls into the field of physiological psychology, which looks at the biological (or physiological) processes that the body uses to make decisions or to learn behavior. Pyschology is defined as “the study of the behavior of animals” and so physiological psychologists study how the body’s biochemical systems act during learning, decision-making, and the execution of learned behaviors (among other esoteric scientific concepts).I think we can sum up this behavior as a study in how the body learns not to repeat poor decisions in the game of Blackjack, which is a structured environment (sort of like a mental maze) where decisions can be evaluated on the basis of risk-versus-reward, and that is because Blackjack’s rules make it possible to learn the probabilities of the game, even if only intuitively (as opposed to memorizing tables in advance).“The neural basis of risky decision-making in a blackjack task” was published in June 2007 in the journal Neuroreport. This paper explains how decisions are divided into two categories: “decisions involving risk” and “decisions involving ambiguity”. Risk is based on knowing the probability of an expected or desired outcome; ambiguity is based on not knowing the probability of an outcome. The study uses Blackjack because its probabilities are known and therefore player decisions are considered to be risky rather than ambiguous.The study authors used neuroimaging to see which areas of the brain lit up when Blackjack players had to make certain kinds of risky decisions. For the lay person, the key takeaway from this study is probably that the authors found a way to identify when players were conflicted about the risks versus potential rewards of their choices for requesting another card in the game.“Blackjack in the Kitchen: Understanding Online Versus Casino Gambling” was published in the Journal of Consumer Research in 2009. The paper is long and uses analysis of anecdotal evidence (they conducted interviews). Although the paper does not present any formal conclusions, the authors do speculate that “online gambling has more of a whiff of scandal and transgression (than casino gambling), and so it may be more desirable to gamblers who can easily choose between casino and online (gambling)”. The study also cites research about problem gambling among adolescents and young adults, predicting that problem gambling increases as access to gambling becomes easier.The focus of the research is to identify changes in consumer behavior as they gamble less in public (at casinos) and more in private (at home on the Internet). The paper does describe online gambling as “unregulated”, but though there are illegal online casinos most are, in fact, licensed by government authorities and audited, so the idea that online gambling is unregulated is misleading (even for 2009). The first online gambling authorities were set up in the 1990s.Nonetheless, some of the recommendations in the paper have been adopted by the online gambling industry, such as is represented by the Responsible Gambling movement, in which players are encouraged to take time away from gambling by the various online casinos and governmental organizations.“Neurobiological correlates of problem gambling in a quasi-realistic blackjack scenario as revealed by fMRI” was published in the journal Psychiatry Research: Neuroimaging in 2010. This study determined that blood flows to certain regions of the brain differed in problem gamblers from the blood flows to those same brain regions in occasional gamblers as both groups played a game of Blackjack; the blood flows changed when they had to make decisions in high risk situations, but problem gamblers’ physiological signs were the opposite of those of occasional gamblers.The problem gamblers’ brain activity was described as “a cue-induced addiction memorynetwork which was triggered by gambling-related cues”.Although most of this research won’t directly help you become a better Blackjack player, it does help to know that there really are “problem gamblers” who are distinct from occasional gamblers. Problem gamblers may succumb to the fallacy of thinking “it can’t happen to me” (an assumption that is common across all risky human behavior and which, normally, does not define addictive or problem behavior) more easily than occasional gamblers. Learning to recognize your gambling dependence (if you have one) could set you on a path toward living a happier, fuller life.Gambling should always be something that you walk away from, not something you try to convince yourself that you can walk away from.