With the outbreak of Covid-19, a 15 years old event surfaced; the World of Warcraft (WOW) pandemic. At the time, discussions appeared about how online video games could be revolutionary tools to predict the evolution of outbreaks. The current events make these discussions more newsworthy than ever.
The corrupted blood epidemy
In 2005, this massively multiplayer online role-playing game (MMORPG) experienced an epidemic due to an unexpected consequence of software update. The developers introduced an additional area in the game with a new enemy to fight. It was meant for high-level players (at the beginning of WOW a player’s avatar starts at level 1, and throughout the game he can increase its level and therefore access more content). Indeed, to make the combat more interesting, the avatars were sometimes infected by a disease called “Corrupted Blood” which was not supposed to spread outside this area. However, developers made a coding mistake and the disease did not disappear once players completed this additional section. Soon, an outbreak occurred because of the avatars’ capacity to teleport themselves in cities, the infection of their pets, and the infection of the non-player characters (characters who are not controlled by players but are part of the game contents, like shopkeepers). Lower-level players died, and cities were literally covered with their bodies. So, before the developers resolved this issue, a quarantine was established both by developers and players. Indeed, because a character resurrection implies to lose some resources or capacities gained during the game, death is undesirable.
Some interesting aspects for epidemiology
This event held epidemiologist’s attention because the “Corrupted Blood” outbreak seemed like a credible representation of reality (Oultram, 2013). Firstly, the abilities of players such as teleportation could be assimilated to airline travel, an important parameter when studying pandemics. Secondly, the capacity of the “Corrupted Blood” to jump between species (between pets and characters) was similar to some real diseases like the avian flu due to the H5N1 virus. Thirdly, there were asymptomatic carriers such as the non-player characters, which the Covid-19 epidemic showed as an important aspect to take into account to halt an outbreak. Finally, the evolution of the “Corrupted Blood” epidemic was influenced by the behaviour of players; some were curious and went to the infected areas, others tried to heal the infected low-level players but contributed to spread the disease instead, and some others deliberately infected players.
Moreover, epidemiologists saw the potential of such data because MMORPG are not only played on a large-scale (5.5 million players reported in 2015 by Blizzard Entertainment, the company selling WOW) but are implying an important commitment from players who play regularly. Therefore, the evolution of their avatars becomes integrated with their real-life preoccupations, making their virtual behaviour comparable to their real-life behaviour (Lofgren and Fefferman, 2007). This commitment can also be expressed by the monthly fee to access MMORPG platforms. For instance, it is approximately 12 euros per month for WOW.
An apparent breakthrough in pandemic studies
There are two main sources of data for epidemiologists. First, traditional epidemiological studies which consist in observation of past epidemics because of the perpetual problem of morality when implementing controlled experiments in this area. Second, large-scale computer simulations where all the parameters are controlled, and which often rely on an economic borrowed axiom; the rational choice theory (Lehdonvirta, 2005). The behaviour of people is modelled by considering that an agent always makes the more rational choice. With MMORPG these two sources were combined; unprogrammed human behaviour on a large-scale with the possibility to control disease’s parameters (Lofgren and Fefferman, 2007). Therefore, it seemed to be a way to settle both the immorality of real-life experiments and the controversy of rational behaviour in modelling. Consequently, epidemiological modelling could be improved through a better representation of behaviour.
Some difficult issues to resolve
This event happened 15 years ago but still today no concrete application has emerged. However, some scientists claim that the “Corrupted Blood” outbreak influenced how they construct pandemics models. For instance, Dr Nina Fefferman who wrote one of the early papers on the WOW pandemic said that its work is still largely influenced by the social side the WOW outbreak revealed as an important aspect of epidemics’ evolution. When the outbreak occurred, players decided to quarantine their avatars because they discussed the risk to spread the disease and came to an agreement. Dr Nina Fefferman states that this type of apparent informal interactions can have important consequences for the whole society. She tries to construct models with a better representation of people’s behaviours in epidemics, based on these observations. But this is an improvement of computer modelling through a more realistic programmed human behaviour and not through an unprogrammed behaviour. In reality, there are difficult issues to resolve to be able to represent unprogrammed human behaviour in a controlled environment.
Firstly, some authors challenged the idea that the negative outcome following a death in WOW (loss of resources or abilities) was sufficient to represent a real-life situation (Oultram, 2013). In the early papers, authors corrected this issue by introducing a risk parameter when modelling behaviour. In video games, this parameter is by definition low because you are not risking your life, but authors argued that by increasing it you could model a situation where your life is really at stake (Lofgren and Fefferman, 2007). However, it did not convince the scientific community, as a high-level of risk is still different from the threat of death (Oultram, 2013).
Secondly, the major issue is probably the divergence of interests between scientists and the gaming industry. Designing a game where epidemiologists could set all the parameters, from the basic reproduction number (how many people one person can infect) to the mortality rate of the disease, implies important constraints which are not necessarily aligned with the players’ expectations. It is unlikely that a gaming company targeting millions of players would risk developing a game entirely designed by epidemiologists. Certainly, scientists’ primary concern is not entertainment, and that could result in the development of less interesting games for players.
However, there are some examples of such partnerships but they always involve fewer players and are relevant only for educational games. For instance, Harvard University developed the River City Project, a computer simulation disguised in a multi-player video game where students can experience the spread of a disease and try to stop it. Thanks to the funding of a US public agency, scholars created this game for learning purposes in collaboration with a gaming company. It seems that in 2009 the project was abandoned because of lack of additional funds. The creation of a MMORPG would require even more funds to target a wider audience, so it seems quite complicated to develop such a game, particularly in the light of the River City Project example.
Another issue, related to the previous one, which was encountered in the field of economics and that one can easily imagine being a problem for epidemiologists too, is the complexity of games created by scholars. Economists have studied video games economics for a while now; from applying microeconomic theory through the rules of demand (Smith, 2017) to comparing guilds (groups of players with a common objective) as alternatives to markets (Lehdonvirta, 2005). In particular, the MMORPG Eve: Online has the advantage to have been developed in collaboration with an economist, making its economy particularly realistic and interesting to study. But this example of collaboration between ex-scholars and the gaming industry did not seem to have been a great success. Indeed, Eve: Online players’ commentaries about the game are reflecting the difficulty to get familiar with it. It seems that the complexity introduced to represent reality with accuracy is not an obvious recipe for a business success. Eve: Online has still a lot of players (around 400,000 players today) but when comparing to the 5.5 millions of WOW, one could wonder once again if it is the best strategy for the gaming industry to bet on scholars, due to the excessive complexity their participation can imply.
Conclusion
As the pandemic of WOW showed, MMORPG could improve considerably epidemiological modelling by combining unprogrammed human behaviour with a controlled environment, through similarities such games share with reality, and the degree of commitment players usually adopt. However, the last decade did not bring a satisfying evolution to this potentiality, only some observations based on the “Corrupted Blood” outbreak were used to improve behaviour representation in computer models. The main hindrance seems to be the lack of incentives from the gaming industry to develop a game exclusively designed by epidemiologists to the detriments of its attractiveness. Only educational games with fewer players succeed in such cooperation. But as the Covid-19 showed, it can be interesting not to give up the idea, as it could really improve the prediction of outbreak evolution in the future.
References
Fenlon, Wes. “The Researchers Who Once Studied WoW’s Corrupted Blood Plague Are Now Fighting the Coronavirus.” Pcgamer. Last modified March 13, 2020. https://www.pcgamer.com/the-researchers-who-once-studied-wows-corrupted-blood-plague-are-now-fighting-the-coronavirus/.
Lehdonvirta, Vili. “Virtual Economics: Applying Economics to the Study of Game Worlds.” , Proceedings of the 2005 Conference on Future Play (Future Play 2005), Lansing, MI, October 13-15, 2005.
https://ssrn.com/abstract=1630302
Lofgren, Eric T., and Nina H. Fefferman. “The untapped potential of virtual game worlds to
shed light on real world epidemics.” The Lancet Infectious Diseases 7, no. 9 (2007), 625-629.
doi:10.1016/s1473-3099(07)70212-8.
Oultram, Stuart. “Virtual plagues and real-world pandemics: reflecting on the potential for
online computer role-playing games to inform real world epidemic research.” Medical
Humanities 39, no. 2 (2013), 115-118. doi:10.1136/medhum-2012-010299.
The River City Project. “The River City Project: Introduction.” Harvard University. Last
modified June 17, 2009. https://muve.gse.harvard.edu/rivercityproject/index.html.
Smith, Christopher. “EVE: Online as a Potential Microeconomic Model.” Master’s thesis, UW
Oshkosh , 2017.
https://minds.wisconsin.edu/bitstream/handle/1793/79143/11460%20OshScholar18%20Smith.
df?sequence=1&isAllowed=y.
By Louise Damade
The TSEconomist 