Behavioral Game Theory

Part 1: Foundations

BGT provides a method to construct game activities that contain clues, rewards and paths recognized by the player at a level below conscious thought. It also helps make games easier to balance and insures that different styles of play are accommodated. This work is also applicable to several other fields, like economics, architecture, political science and education. BGT uses behavioral methods derived from animal models. 
In part 1 of this work, we will examine the foundations the theory is built upon. 


 Humans are a kind of animal. 

 We share genes with many other creatures and during fetal development our brains, hearts and other organs resemble those of other animals. 

Animals, including humans, share a biological imperative to feed and to breed. 

Deep structures in our brains respond to stimuli related to these primal activities. The smell of fresh popcorn in the theater lobby or using attractive young women to advertise razors or motor oil demonstrates that we are still susceptible to these stimuli. Certain stimuli associated with food and reproduction operate at level below conscious thought and reasoning. 

Animals are suitable models for aspects of human behavior. 

 It is standard practice among behavioral researchers to use animal models to investigate human behavior. One would not, for instance, raise a human child with a wire mesh mother, though insights gained from raising a monkey in this way can be applied to human children. 

 The Ludic experience, playing to learn, predates Homo Sapiens. 

 Many juvenile animals play to learn skills that will serve them later in life. Mock hunting and expressions of dominance are common aspects of these animal games. It is also worth noting that while pure entertainment is often a drain on a society (cost vs return), ludic entertainment, like video games are almost always a net gain, due to educational benefits. 

 Unlike many animals, humans are socially plastic. 

 Where dogs or ducks will naturally form certain social groupings with specific roles, humans may exist in any of several social groupings, simultaneously. For instance, during a break in a business meeting with his manager, a man checks his messages and responds to calls from his wife, his son and a charitable organization he chairs. In only a few moments, the man’s social role may shift from subordinate, to dominant, from follower to leader to peer and back again. 

 Humans can exist in multiple group structures. 

 The caller described above not only changes his role, from leader to follower, he also exists in multiple animal-style group constructs. His softball team might operate like a pack or pride, his company probably looks more like a herd or hive, while other groups could work more like a colony of coral or a band of apes. 

 One major bifurcation in animal behavior is associated with diet. 

 Animals are easily classified by diet as herbivores, carnivores or omnivores. The abilities required to obtain and use different food resources have contributed to profound differences in both behavior and biology between various kinds of animals. Example, whale shark vs Mako shark; Grey whale vs Orca. 

 Agriculture is trending human behavior away from pack-like, predatory behavior toward a more herd-like, herbivorous behavior. 

 Previous assumptions about our hunter-gatherer past have been called into question by recent analysis of ancient coprolites. These data reveal that prior to the rise of agriculture, about 86% of the human diet consisted of animal protein. Prior to agriculture, the primary human social structure was a small, tribal hunting unit with strong leadership, similar to a pack or pride. Using grains (grasses) as food allowed humans to achieve herd-like population densities, which required more passive behavior, due to proximity. 

 The levels of adrenal compounds produced, when in proximity to others, is the main genetic selection involved in civilization, of humans, and in domestication, of animals 

 The famous Russian experiments in domesticating foxes demonstrates the far-reaching impact to appearance, vocalization and social behavior caused by selection for reduced levels of adrenal compounds. Since those discoveries, we can see the same physical markers in every domesticated breed of animal, even those domesticated in antiquity, and in humans. For instance, a study in 1995 showed that felons jailed for violent crimes had a ‘personal space’ zone more than twice the size of people with no criminal record. 

 Due to an incomplete change in our behavior and biology, humans display both carnivorous and herbivorous behaviors. 

 Walking in a crowd, people are able to herd, flock and school. We hunt for bargains. We eat sweets alone and consume meat in a group. Our digestive tract is too long for eating only meat and too chart for eating only vegetation. We are omnivores that can behave like a lion or sheep. Sometimes we recognize the Alpha role, sometimes we work as peers. 


 So, Behavioral Game Theory offers guidelines for constructing human activities (games) that can resist, support or reward various human behaviors, by the use of constructs based on the animal behaviors still operating within us all. 

 Part 2: Definitions 

Now that we’ve established some key points on human behavior, in part 1, let’s look at games themselves. For instance, what are games? One of the first tries at understanding games, came from James Waldengrave, who wanted to do better at cards. His thoughts presaged minimax and mixed strategies by more than 200 years.  

 The philosopher Ludwig Wittgenstein made an early attempt to define games. He identified the concepts of play, rules and competition but failed to establish a formal definition. 

 Sociologist Roger Caillois felt games could be defined as activites that are fun or light hearted, that are understood to be seperate from normal reality, that have an element of fiction, that are governed by rules, that have a uncertain outcome and are non-productive. 

 Historian Johan Huizinga, in his book, “Homo Ludens”, states that culture arises from play and that the next step in human evolution may well be Homo Ludens, the game playing man. His definition of games said the activity must allow freedom, occur in a non-real world, create order and be free from real world financial impact. 

 Since World War II, the formal game theory created by mathmetician John Von Neumann has taken an important role in international finance and statecraft. This work specializes in calculating the optimal decisions, where 2 agendas are in conflict. Though their general game definition is simply conflict, they have defined categories of games (zero sum, for instance) and examined the relationship between information the players posess versus knowledge they need to make the best, next decision. Further, game theory describes how to minimize maximum harm and to predict points where motivation and potential rewards, for players, hang in equilibrium. 

 Biology too has started to include game theory among its tools. Games originated in animal behavior, with mock hunting and king of the hill, so it only surprises economists that their pet methodology works even better to understand behavior in real bears and bulls, than it does for the market versions. The influential Evolutionarily Stable Strategy theory functionas exactly as a Nash Equilibrium. The most intersting addition to game theory, from biology, to me, is altrusim. Where homo economicus, the imaginary creature used in economic simulations always acts in perfect self interest, biologists know that imperfect self interest can be an aid to group survival. Over time, imperfect self interest has become evolutionarily encoded in us, and so, is a widely recognized social currency. 

 Recently, game maker, Chris Crawford, has proposed a definition of games that builds on Caillois’ ideas. He further defined competitions, puzzles, toys and other forms that are similar to games, but lack final outcomes, active antagonists or other game requirements. A race against the clock, for instance, or on going simulation without a final win state, by his standards are not true games. 

  For our purpopses, we will define games thusly:  

  • Is a non-real activity in a non-real place
  • Has one or more goals,
  • Has one or more challenges,
  • Has rules and
  • Has an outcome.

So, the core tenets of my Behavioral Game Theory are:  

  •  Motivation (incentive) comes from the primal need to feed and breed
  • Motivation results in 3 classes of game activities: Amassing, Skills display and politics.
  • Players approach game activities using 3 behavioral modes: carnivore, herbavore or omnivore.

In part 3, we will examine these ideas in more detail. 

  Part 3: Theory

 In Part 2, we established these rules for Behavioral Game Theory:   

  • Motivation (incentive) comes from the primal need to feed and breed
  • Motivation results in 3 classes of game activities: Amassing, Skills display and politics.
  • Players approach game activities using 3 behavioral modes: carnivore, herbavore or omnivore.

We all can understand that an organism must take in nourishment to survive. This is the most primal rule, from single cell creatures right up to humans’ only instinctual behavior, to suckle. Unless we can sustain the self, we cannot go beyond ourselves to interact with others. 

 Then, beyond mere survival of the self, comes reproduction, the continuation of the species. To reproduce means having access to a food supply that is roughly 3 times greater than that required to maintain the self. 

 Reproduction also requires interacting with another in a non-lethal way. It means forming an agenda that involves another person and pressing it, without driving the other person away or killing them. 

 These realities are the core of BGT. 

 We are wired to amass food, or the territory food comes from or abstracted tokens of food like cattle gold or stock. Excess supply provides wealth, security and enhances reproductive viability. 

The acts of hunting and gathering, to get food, requires some skill and knowledge. Hunters with keen eyesight, stamina, sure aim and similar skills get better results. The self benefits from well honed skills and any potential mate would also take notice of the skills required to provide food. In primitive societies, racing and accuracy contests are literally indicative of the ability of competitors to get food. In more advanced societies, the food getting skills may be abstracted to piano playing, selling cars or programming computers. In every case, being more skilled is better and demonstrating skills to potential mates and rivals is important.   

Finally, living and working with others, in a cooperative, mutually benificial way, without undue violence, over long periods of time is required of social creatures. While orangutangs are solitary, except for mating, chimpanzies, gorillas and humans exist in social groupings. That requires us to treat our peers differently than prey. 

Having both evolved from simpler creatures and learned by observing animal totems and familiars, we can borrow the strategies of other animals. 

Before the rise of agriculture, humans were the predator ape. We hunted and lived in small, familial packs, prides and tribes. The main business of all carnivores is death. They seek and destroy. They take. They live on the edge of famine, punctuated by feasts. These conditions require high skills, strength, speed, agility, cleverness and a tight social grouping based on a skills ranking. 

Grains are grasses. Once we started growing grains, people became herd animals. Our populations grew, because, like other large herbivores, our main business was life; a lifecycle of lean and plenty, of mating and new births. This lifestyle values larger groups, foresight and patience. The winners in this game amass things like food, territory, mates, offsprings, flocks and fields. 

Still, at our core, we are omnivores. Like bears, we can enjoy berries when the season comes, eat fish or even stalk and kill an elk to prepare for winter. Omnivores are flexible enough to behave as a placid herbivore one day and display deadly preditory abilities the next. 

From the Dao, to Hegle to deconstructionism, much has been written about dualism. Our walnut-like, bicameral brains predispose us to having 2 minds that view the world in 2 ways. My contribution to this long line of thinking, is to propose that just as their are right and left handed people, there are also people predisposed to carnivore and to herbivore behaviors. 

For instance, I often poll people by asking: “in your opinion, are we all in this together, or is the world really every man for himself?” I only allow people to choose one option and I usually provide examples of windfalls, excess, insufficiency, famine and so on that require a commitment to one strategy or the other. The results show that there is a predictable distribution of people in each camp. Predatory behavior is common in crime, in sales and sports. Herd or prey behaviors are common in academia, in religion and in government. 

We also automatically understand and appreciate the predator/prey conflict even when astracted. For instance, imagine a story about a race between two drivers. One is a wealthy young man, of only average skills, but driving the best car money can buy. The other competitor is a fellow from a humble background. His car is simpler and mostly home-made, but he is a supremely skilled driver. 

Already we have feelings about these characters. Many people favor the poor, but skillfull underdog. Except, when we analyze this scenario deeper, we see that the wealthy driver practices amassing, which marks him as the prey. The poorer driver is behaving like a predator, by depending on superior skills. By favoring the skillful driver, our ancient selves are pulling for the predator. 

In the next part, we’ll apply this framework to shopping, to games and other areas of life. 

Part 4: Applications 

Imagine for a moment, that a large, chic, health food grocery chain has decided to work with a maker of very efficient, hybrid automobiles. They see that their audiences have overlapping demographics. Their customers are essentially the same people, with with environmental motivation and money to spend. Both companies are growing at a good pace and they believe working together will “create synergy and save energy”. 

So they decide to put a manned kiosk, to sell cars, on the grocery store floor. The stores are large and the cars are small, so on occasion, they park a shiny hybrid in the produce section of the store. Other hybrids await test drives, on the store parking lot. The kiosk booth is as small as it can be and still offer a bit of privacy for financial discussions and the sales people are selected for compatibility with the shoppers. 

Despite their best effort, car sales are awful and produce sales are also trending lower. Shoppers are willing to look at the cars, in the parking lot, but will not engage the salesperson inside the store. A student of BGT would not be surprised at this outcome. 

While food shopping, people are amassing. Like an early human moving through a forest, they pluck fruits and nuts from bush-high racks. Even the meats are cut down to fruit size, since grocery shoppers are acting very much like herbavores. They one thing they do not want to encounter in this modern jungle of plenty is a predator. A car salesperson is exactly that – a skilled, patient hunter looking for a score; a kill. 

Since the kiosk is stationary, the shoppers quickly learned to avoid the predator’s territory, unless he had already captured prey, in which case, they could scamper past. The salesperson thought, “I never see a customer, until I already have one; when it rains it pours”. 

When shopping, we are either amassing, like a browsing herbavore, or we are on the hunt for a deal, like a predator. We bring home the bacon, not the carrots. Car sales are particulalry interesting since, going to a car dealership is very much like going to the watering hole; dangerous but sometimes required. We know we are being predated. The salespeople are openly predatory. Though, one of their better techniques is to recast the roles as a male rivalry over a desirable female, with the car as the cow. This tactic of playing on reproductive drive, as hunters know, makes males reckless and bold. 

In games, we often see classes of players that take a conservative path through a level, amassing items, staying defensive and thinking of space as territory. Other groups of players trade on high skill. They use items as they find them and view space as paths that the prey must traverse. 

In a shooter, for instance, an herbavore will move to a defensible spot (territory) and hold it briefly by taking a defensive stance. Meanwhile, a predatory player will prowl, looking for movement. For that player, space is a collection of paths, not stops, so a choke point, intersection or corner is an opportunity. Still others, randomly veer from one strategy to the other. These omnivore style players pick and choose, improvisationally, on a moment-by-moment basis. Bulls, Lions and Bears. 

Notice too, that the group relationships are also Bull or Lion. Bulls defend and herd their helpless group through dangerous territory. Lions demand their underlings rise to the occasion, take risks and use skill in assisting with the hunt. Bears, of course, do all of that. 

Further, as game designers, or grocery store designers, or car dealerships, we need to build paths suited to these behavioral groups. Game level designers get three paths for the price of two, when they create a bull and tiger path, since the bear will use both. The store manager would do well to allow the meat market to do special sales with tricky conditions, coupons and gotchas, while produce remains a placid forest. Car sellers could reduce the fear of predation, just like a modern slaughterhouse, by leading the buyers through a geographical migration, from point to point, for each part of the sale, until the deal is finally closed, in a private area where the waiting crowd cannot see or hear the kill. 




Comments are closed.