De Dreu and Gross (D&G) show that in games of attack and defense, it is in the defenders’ best interest to match the attackers’ strategy, whereas it is in the attackers’ best interest to mismatch the defenders’ strategy. They propose that coordination emerges more easily and spontaneously among defenders, who have stronger in-group identification and are expected to invest more resources in conflict. While their model usefully draws attention to the different payoff structures of attack and defense, it underestimates the advantages that attackers have in setting the stage of conflict and overestimates the ability of defenders to match the attackers’ strategy. As a result, in contrast to predictions from their model, attackers usually fare better than defenders, having significantly less mortality and a greater rate of success. These patterns are found in both chimpanzees and in human groups. Unlike defenders who are pushed into conflict with only two strategies, conflict or flight, attackers choose whether or not to initiate conflict, as well as the time and location of their attack. This creates on-the-ground difficulty in matching strategies, because defenders may simply be incapable of this. The result is an extraordinarily high success rate for attackers compared with defenders.
Accounts from a vast number of human societies, as well as chimpanzees, show that attackers in intergroup conflict often seek to minimize risk to themselves and wait to attack until they have a strategic advantage (Gat Reference Gat1999; LeBlanc Reference LeBlanc, Allen and Jones2016). When chimpanzees patrol the boundaries of their community, or during hunter-gatherer territorial skirmishes, individuals may flexibly switch between defense (or flight) and attack, depending on the balance of power between sides (Manson & Wrangham Reference Manson and Wrangham1991; Wrangham & Glowacki Reference Wrangham and Glowacki2012). Chimpanzees patrol the edges of their territory with hypervigilance, as D&G predict in the defense condition; however, when they encounter an enemy and have a significant numerical advantage, they can shift immediately to attack (Langergraber et al. Reference Langergraber, Watts, Vigilant and Mitani2017). The Ngogo chimpanzee community at Kibale National Park commit more intergroup killings per year than any other studied chimpanzee community, at more than twice the amount of the next closest group (Wilson et al. Reference Wilson, Boesch, Fruth, Furuichi, Gilby, Hashimoto, Hobaiter, Hohmann, Itoh, Koops, Lloyd, Matsuzawa, Mitani, Miungu, Morgan, Muller, Mundry, Nakamura, Pruetz, Pusey, Riedel, Sanz, Schel, Simmons, Waller, Watts, White, Wittig, Zuberbühler and Wrangham2014). Yet they also have the lowest rate of mortality from intergroup killings of any recorded chimpanzee community (Falk & Hildebolt Reference Falk and Hildebolt2017). Ngogo chimpanzees patrol their territory in large groups, averaging over 17 individuals per party, and their attacking parties outnumber the victim's parties by 10 to 1 (Wilson et al. Reference Wilson, Boesch, Fruth, Furuichi, Gilby, Hashimoto, Hobaiter, Hohmann, Itoh, Koops, Lloyd, Matsuzawa, Mitani, Miungu, Morgan, Muller, Mundry, Nakamura, Pruetz, Pusey, Riedel, Sanz, Schel, Simmons, Waller, Watts, White, Wittig, Zuberbühler and Wrangham2014; Wood et al. Reference Wood, Watts, Mitani and Langergraber2017).
The most common form of attack during intergroup conflict for many small-scale or uncentralized societies is the ambush, where attackers take advantage of the element of surprise, leaving defenders at a disadvantage in attempting an organized response (Gat Reference Gat1999; Otterbein Reference Otterbein2009; Wilson & Glowacki Reference Wilson, Glowacki, Muller, Pilbeam and Wrangham2017). Walker and Bailey (Reference Walker and Bailey2013) provide data on mortality rates for within and between-group violence across 44 traditional lowland South American societies. They note that out of 238 death events (such as duels, homicide, and raids), in only 5 (2%) did an attacker die (Walker & Bailey Reference Walker and Bailey2013). While D&G emphasize overconfidence as one psychological mechanism contributing to a greater willingness to attack, the ethnographic and primatological evidence indicates that attackers tend to act as though they have a keen awareness of their likelihood of success, among humans and chimpanzees alike.
Intergroup violence among the Andaman Island hunter-gatherers conformed to the general pattern found across many small-scale societies, where “The whole art of fighting was to come upon your enemies by surprise, kill one or two of them and then retreat” (Radcliffe-Brown Reference Radcliffe-Brown1922). Attacks would only occur when the attacking party was certain it could take the enemy by surprise (Radcliffe-Brown Reference Radcliffe-Brown1922). Attackers would retreat if they met resistance, while if the defenders were truly caught by surprise, they would often flee to save themselves (Radcliffe-Brown Reference Radcliffe-Brown1922). While D&G propose that coordination should emerge more readily among defenders, it is worth emphasizing how successful ambushes can lead individual defenders to defect quickly, fleeing for reasons of self-preservation. While there is theoretical and empirical work demonstrating the importance of in-group defense in promoting altruistic behavior (Böhm et al. Reference Böhm, Rusch and Güreck2016; Rusch Reference Rusch2013; Reference Rusch2014a), a greater consideration of the situations where flight is a higher payoff strategy for individuals than defense may also help elucidate the reasons why attacks have such high rates of success.
D&G emphasize that games of attack and defense create stronger selection pressures on defenders than attackers, because the costs of failed defense are often larger than the costs of a failed attack, but they underemphasize the benefits of successful attacks. There is evidence that successful warriors have increased reproductive success in some small-scale societies (Chagnon Reference Chagnon1988; Glowacki & Wrangham Reference Glowacki and Wrangham2015) and cultural incentives such as material benefits or increased social standing are commonly accorded to warriors across many traditional cultures (Glowacki & Wrangham Reference Glowacki and Wrangham2013). Males can also benefit from alliances formed in a raiding context, cooperating with kin members or other close allies, and leveraging participation in raids to net better opportunities in marriage exchanges (Macfarlan et al. Reference Macfarlan, Walker, Flinn and Chagnon2014; Reference Macfarlan, Erickson, Yost, Regalado, Jaramillo and Beckerman2018). Chimpanzee males who patrol the boundaries of their communities and engage in coalitionary violence also tend to have increased reproductive success (Gilby et al. Reference Gilby, Brent, Wroblewski, Rudicell, Hahn, Goodall and Pusey2013; Langergraber et al. Reference Langergraber, Watts, Vigilant and Mitani2017; Williams et al. Reference Williams, Oehlert, Carlis and Pusey2004).
D&G make a unique contribution in examining conflict through asymmetric matching-mismatching strategies in games of attack and defense, and their framework provides possible avenues for further fruitful work. Investigating the conditions where defenders are incentivized to defect, and expanding on the advantages of the attack condition, may help reconcile the empirical record of high success of attack with D&G's prediction of easier coordination and greater success while defending.
De Dreu and Gross (D&G) show that in games of attack and defense, it is in the defenders’ best interest to match the attackers’ strategy, whereas it is in the attackers’ best interest to mismatch the defenders’ strategy. They propose that coordination emerges more easily and spontaneously among defenders, who have stronger in-group identification and are expected to invest more resources in conflict. While their model usefully draws attention to the different payoff structures of attack and defense, it underestimates the advantages that attackers have in setting the stage of conflict and overestimates the ability of defenders to match the attackers’ strategy. As a result, in contrast to predictions from their model, attackers usually fare better than defenders, having significantly less mortality and a greater rate of success. These patterns are found in both chimpanzees and in human groups. Unlike defenders who are pushed into conflict with only two strategies, conflict or flight, attackers choose whether or not to initiate conflict, as well as the time and location of their attack. This creates on-the-ground difficulty in matching strategies, because defenders may simply be incapable of this. The result is an extraordinarily high success rate for attackers compared with defenders.
Accounts from a vast number of human societies, as well as chimpanzees, show that attackers in intergroup conflict often seek to minimize risk to themselves and wait to attack until they have a strategic advantage (Gat Reference Gat1999; LeBlanc Reference LeBlanc, Allen and Jones2016). When chimpanzees patrol the boundaries of their community, or during hunter-gatherer territorial skirmishes, individuals may flexibly switch between defense (or flight) and attack, depending on the balance of power between sides (Manson & Wrangham Reference Manson and Wrangham1991; Wrangham & Glowacki Reference Wrangham and Glowacki2012). Chimpanzees patrol the edges of their territory with hypervigilance, as D&G predict in the defense condition; however, when they encounter an enemy and have a significant numerical advantage, they can shift immediately to attack (Langergraber et al. Reference Langergraber, Watts, Vigilant and Mitani2017). The Ngogo chimpanzee community at Kibale National Park commit more intergroup killings per year than any other studied chimpanzee community, at more than twice the amount of the next closest group (Wilson et al. Reference Wilson, Boesch, Fruth, Furuichi, Gilby, Hashimoto, Hobaiter, Hohmann, Itoh, Koops, Lloyd, Matsuzawa, Mitani, Miungu, Morgan, Muller, Mundry, Nakamura, Pruetz, Pusey, Riedel, Sanz, Schel, Simmons, Waller, Watts, White, Wittig, Zuberbühler and Wrangham2014). Yet they also have the lowest rate of mortality from intergroup killings of any recorded chimpanzee community (Falk & Hildebolt Reference Falk and Hildebolt2017). Ngogo chimpanzees patrol their territory in large groups, averaging over 17 individuals per party, and their attacking parties outnumber the victim's parties by 10 to 1 (Wilson et al. Reference Wilson, Boesch, Fruth, Furuichi, Gilby, Hashimoto, Hobaiter, Hohmann, Itoh, Koops, Lloyd, Matsuzawa, Mitani, Miungu, Morgan, Muller, Mundry, Nakamura, Pruetz, Pusey, Riedel, Sanz, Schel, Simmons, Waller, Watts, White, Wittig, Zuberbühler and Wrangham2014; Wood et al. Reference Wood, Watts, Mitani and Langergraber2017).
The most common form of attack during intergroup conflict for many small-scale or uncentralized societies is the ambush, where attackers take advantage of the element of surprise, leaving defenders at a disadvantage in attempting an organized response (Gat Reference Gat1999; Otterbein Reference Otterbein2009; Wilson & Glowacki Reference Wilson, Glowacki, Muller, Pilbeam and Wrangham2017). Walker and Bailey (Reference Walker and Bailey2013) provide data on mortality rates for within and between-group violence across 44 traditional lowland South American societies. They note that out of 238 death events (such as duels, homicide, and raids), in only 5 (2%) did an attacker die (Walker & Bailey Reference Walker and Bailey2013). While D&G emphasize overconfidence as one psychological mechanism contributing to a greater willingness to attack, the ethnographic and primatological evidence indicates that attackers tend to act as though they have a keen awareness of their likelihood of success, among humans and chimpanzees alike.
Intergroup violence among the Andaman Island hunter-gatherers conformed to the general pattern found across many small-scale societies, where “The whole art of fighting was to come upon your enemies by surprise, kill one or two of them and then retreat” (Radcliffe-Brown Reference Radcliffe-Brown1922). Attacks would only occur when the attacking party was certain it could take the enemy by surprise (Radcliffe-Brown Reference Radcliffe-Brown1922). Attackers would retreat if they met resistance, while if the defenders were truly caught by surprise, they would often flee to save themselves (Radcliffe-Brown Reference Radcliffe-Brown1922). While D&G propose that coordination should emerge more readily among defenders, it is worth emphasizing how successful ambushes can lead individual defenders to defect quickly, fleeing for reasons of self-preservation. While there is theoretical and empirical work demonstrating the importance of in-group defense in promoting altruistic behavior (Böhm et al. Reference Böhm, Rusch and Güreck2016; Rusch Reference Rusch2013; Reference Rusch2014a), a greater consideration of the situations where flight is a higher payoff strategy for individuals than defense may also help elucidate the reasons why attacks have such high rates of success.
D&G emphasize that games of attack and defense create stronger selection pressures on defenders than attackers, because the costs of failed defense are often larger than the costs of a failed attack, but they underemphasize the benefits of successful attacks. There is evidence that successful warriors have increased reproductive success in some small-scale societies (Chagnon Reference Chagnon1988; Glowacki & Wrangham Reference Glowacki and Wrangham2015) and cultural incentives such as material benefits or increased social standing are commonly accorded to warriors across many traditional cultures (Glowacki & Wrangham Reference Glowacki and Wrangham2013). Males can also benefit from alliances formed in a raiding context, cooperating with kin members or other close allies, and leveraging participation in raids to net better opportunities in marriage exchanges (Macfarlan et al. Reference Macfarlan, Walker, Flinn and Chagnon2014; Reference Macfarlan, Erickson, Yost, Regalado, Jaramillo and Beckerman2018). Chimpanzee males who patrol the boundaries of their communities and engage in coalitionary violence also tend to have increased reproductive success (Gilby et al. Reference Gilby, Brent, Wroblewski, Rudicell, Hahn, Goodall and Pusey2013; Langergraber et al. Reference Langergraber, Watts, Vigilant and Mitani2017; Williams et al. Reference Williams, Oehlert, Carlis and Pusey2004).
D&G make a unique contribution in examining conflict through asymmetric matching-mismatching strategies in games of attack and defense, and their framework provides possible avenues for further fruitful work. Investigating the conditions where defenders are incentivized to defect, and expanding on the advantages of the attack condition, may help reconcile the empirical record of high success of attack with D&G's prediction of easier coordination and greater success while defending.