Analysis of subscription characteristics and user mobility in different types of MMORPGs

Transcription

1 December 009, 16(6): The Journal of China Universities of Posts and Telecommunications Analysis of subscription characteristics and user mobility in different types of MMORPGs PENG Hui ( ), SUN Yan-li, WU Hong School of Economics and Management, Beijing University of Posts and Telecommunications, Beijing , China Abstract Massively multiplayer online role-playing game (MMORPG) is one of the fastest-growing segments of the online game industry. MMORPG subscription refers to game accounts logged in during a certain period of time. MMORPG user mobility explains the dynamics of subscriber size change. This article explores the subscription characteristics and user mobility in different types of MMORPGs. It is found that subscription characteristics in different types of MMORPGs are marked by dissimilarity. On one hand, the curve of competition-based game subscription is linear, and the stability period is short, therefore it is necessary for the game operators to introduce new versions timely. On the other hand, the gradient rate of subscription in the community-based game is slow at first and then fast, and the stability period is long, hence it will take the game operator a long cycle to launch a new version. The difference of subscription curve is caused by the fact that user mobility in competition-based game is higher than in the community-based game, as attractiveness of community-based game can maintain a longer period due to network externality. The purpose of the study is to help the game operator to understand the development stage and features of the game and to make effective decisions to attract more players. Keywords virtual reality, online game, subscription characteristics, user mobility, subscription curve, life cycle, MMORPG 1 Introduction Defined by international data corporation (IDC), online game is the game that adopts TCP/IP protocols based on the Internet, and multiple players can participate at the same time (source: MMORPG is one of the fastestgrowing segments of online game industry, in which a user controls a character represented by an avatar, and then explores in the game world, interacts with others, undertaking various tasks and so forth. The non-player character (NPC), controlled by the game operator, usually provides special virtual services [1]. MMORPG subscription refers to game accounts logged in during a certain period of time. The subscription size and its changes with time affect the interaction among players as well as the game operator. MMORPG user mobility explains the dynamics of subscriber change, which could help the operator in designing and Received date: Corresponding author: PENG Hui, gracepenghui@16.com DOI: /S (08) maintaining the game. Earlier works about subscription characteristics and user mobility of online game are limited. Woodcock studies the subscription trends in massive multiplayer online game (MMOG) and concludes that MMOG subscription movements tend to follow hyperbolic or parabolic curves with little variation among them ( downloads/). Carooe analyzes the changing subscription based on Ultima Online []. Cornett examines the usability challenges faced by new players of MMORPGs [3]. However, none of these studies has analyzed the subscription characteristics and user mobility in different types of MMORPGs. Following the in introduction above, this article describes classification of MMORPG. Sect. 3 explains the subscription characteristics in different types of MMORPGs and the complete subscription curve. In Sect. 4 the user mobility of MMORPGs is demonstrated. Finally, conclusions are drawn in Sect. 5.

2 1 The Journal of China Universities of Posts and Telecommunications 009 Classification of MMORPG So far, the online game industry has not established a standardized classification of MMORPG. Based on the nature of online game and the way people obtain psychological satisfaction, the authors hold it that the functions of MMORPG can be classified into two types: competitionbased and community-based. In competition-based games, the players should compete with external conditions (other players or NPC) constantly, and there is a strict hierarchy system to measure the players in-world achievement. It is a way to develop a sense of accomplishment among the players. Community-based games refer to an equal and harmonious environment wherein the players can collaborate and communicate with each other, and there is no hierarchy. It is a way to develop a sense of social belonging among the players. No game is an exclusively competition-based or community-based, but a combination of both games. Several exhibit more characteristics of competition-based game, while others show more emphasis on community-based. It is therefore necessary to define the type of a MMORPG by its dominant function. Sect..1 demonstrates the criterion to determine a competition-based MMORPG, while Sect.. presents the criterion to determine a community-based MMORPG. In Sect..3, an empirical classification of MMORPGs is done..1 Criterion to determine a competition-based MMORPG: subscription distribution The world of warcraft (WoW) introduced by Blizzard is the most representative competition-based MMORPG, in which the hierarchy system assigns a level to each avatar, players find pleasure through advancing the levels and accomplishing the quest set by NPC [4 5]. The subscription distribution statistics of WoW shows a peak subscription size at level 70 and level 80 (Fig. 1). As was confirmed, the highest level of avatars in WoW was at level 70 before Nov, 008 and the majority of avatars had achieved this highest level. After the game operator launched a new version named Lich King Fury on Nov 13, 008, the supreme level in WoW was upgraded from level 70 to level 80. Hence, plentiful players compete to achieve level 80, as the situation indicated in Fig. 1. It shows that players have the urge to gain achievement by mastering game avatars from the lowest level to the highest, thus verifying the function of competition-based game. Fig. 1 Subscription distribution of WoW as of Mar 7, 009 (source: Inspired by this example, the authors summed up a standard to determine a competition-based game: Xi* Xi = max ; i= 1,,..., H (1) Y Y X i represents the subscription size of i level; Y represents the total subscription size, and i * is the level with the greatest subscription size, H is the highest level.. Criterion to determine a community-based MMORPG: network externality A situation in which the private costs of or benefits to the producers or purchasers of goods or services differs from the total social costs or benefits entailed in its production and consumption, is called externality [6]. Network externality has been defined as the benefit or surplus an agent derives from a good consumption changes with the number of other agents consuming the same kind of good [7]. With the increase in number of subscribers, both the opportunity of interaction and content of game playing experience for the players increase. The increased playability of the game, in turn, increases the utility players gain from the game. To verify the network externality in the community-based

3 Issue 6 PENG Hui, et al. / Analysis of subscription characteristics and user mobility in different types of MMORPGs 13 game, it can be assumed that the value of network V is proportional to the square of the subscription size Y in the game, i.e.: V = ay () The most representative community-based MMORPG is Second Life launched by Linden Lab [8]. The value of the network in Second Life can be replaced by the volume of user to user transaction between avatars. The regression result (Table 1) of Second Life is satisfactory. The time-series regression of in-game transaction volume on the square of subscription size is conducted for the period Jan, 005 through Jul, 007. Eq. (3) shows that the transaction volume is positively correlated with the square of subscription size. The regression result presented in Table 1 confirms that Second Life should be classified as a community-based game. V = Y (3) Table 1 Regression result of Second Life (see data in Appendix A) Variable Volume of user-user transaction Coefficient t-statistics Square of subscription size R-squared Adjusted R-squared S.E. of regression Sum squared residual Log likelihood Mean dependent variable S.D. dependent variable Akaike info criterion Schwarz criterion Durbin-Watson statistics Empirical classification of MMORPGs Based on the above criteria, the authors have selected 5 MMORPGs (WoW, Second Life, War Hammer, The Lord of the Rings, and HiPiHi) to test their subscription distributions and network externalities to determine their type of game. War Hammer was originally a table top war game released by Mythic Entertainment. The Lord of the Rings (Lotro) is the world s first and only MMOG based on the novel of the same title [9]. The subscription distribution of War Hammer and Lotro are shown in Figs. and 3, respectively. Fig. 3 Subscription distribution of Lotro as of Apr 8, 009 (source: lotro.mmodb.com) HiPiHi is an online 3D virtual world game founded by Xu Hui and Rao Xuewei in China. As there is no hierarchy system in HiPiHi, one can only test the network externality of HiPiHi. The value of the network can be replaced by the quantity of in-game virtual items ( I ), the time-series regression of virtual items volume on the square of subscription size is conducted for the period Nov 30, 008 through Feb 15, 009. Eq. (4) shows that the weekly volume of virtual items is positively correlated with the square of subscription size. The regression result is presented in Table, which confirms HiPiHi as a community-based game. I = Y (4) Table Regression result of HiPiHi (see data in Appendix B) Variable Volume of in-game items Coefficient t-statistics Constant Square of subscription size R-squared Adjusted R-squared S.E. of regression Sum squared residual Log likelihood Durbin-Watson statistics Mean dependent variable S.D. dependent variable Akaike info criterion Schwarz criterion F-statistic Prob (F-statistic) Therefore, the five MMORPGs can be classified as in Table 3. Table 3 Empirical classification of MMORPG Name of MMORPG Subscription characteristics MMORPG type WoW i * =H=80 Competition-based War Hammer i * =H=40 Competition-based The Lord of the Rings i * =H=50 Competition-based Second Life Network externality Community-based HiPiHi Network externality Community-based 3 Subscription characteristics in different types of MMORPGs Fig. Subscription distribution of War Hammer as of Apr 5, 009 (source: warhammer.db.mmosite.com) The authors use subscription curve to exhibit how the

4 14 The Journal of China Universities of Posts and Telecommunications 009 subscription size changes with time since the game was launched in different types of MMORPGs. Sects. 3.1 and 3. demonstrate the subscription characteristics of competitionbased and community-based MMORPG respectively. Sect. 3.3 compares the differences of subscription characteristics between the two types of MMORPGs in a life cycle view. 3.1 Subscription characteristic of competition-based MMORPG The statistical method of subscription in WoW is: active game accounts within one month, including both horde and alliance, two avatar races in WoW. From the subscription curve of WoW (Fig. 4), one can see: Fig. 4 Subscription curve in WoW US&EU as of Nov, 008 (source: From Apr, 005 to Dec, 006, the number of subscription in the game has been in a state of sustained growth. From Jan, 007 to Nov, 008, the subscription curve is wave- shaped, and cycled. Every substantial growth in the game subscription is well connected to a new sensational patch released in the game (source: 1) In May, 005, the film 1.4 was launched. ) In Jan, 006, patch 1.9 The Gates of Ahn Qiraj was introduced. 3) In Jan, 007, patch.0 Before the Storm was introduced. 4) In Jan, 008, patch.3.3 was released. 5) In Nov, 008, patch 3.08 Lich King Fury was released. In competition-based MMORPG, the dominant relationship is that between players and game operator. The operator provides new quest in new film while players get amusement mainly from the complex and constantly enriched game plot, and the interaction among players is merely complemen -tary [10]. If the operator does not launch a new game version for a long time, the single relationship between players and game operator is very likely to collapse, then players will gradually lose their interest in the game. Therefore, the subscription of the competition-based MMORPG is markedly affected by the new version of the game. 3. Subscription characteristic of community-based MMORPG Statistical method of subscription in Second Life is: the number of premium accounts. From the subscription curve of Second Life (Fig. 5), one can see: From Jan, 005 to Aug, 006, the subscriber size is on slow growth. From Sep, 006 to Jun, 007, there is a rapid increase in subscription. From Jun, 007 to Apr, 008, the subscription fluctuated in a small range, but relatively stable. The dominant relationship in community-based MMORPG is that among the players, whereas the relationship between players and game operator is only complementary [11]. The players in the community-based MMORPG can be seen as a part of the game plot. Even if the operator does not make any strategy at all, the game can still maintain its attractiveness for a long period of time. However, if a certain number (threshold) of players quit the game because of herding effect, a lot of other players or even all of them may quit soon. Generally speaking, this complex interpersonal relation is stable, therefore, the subscription curve of the community-based MMORPG is gentle, and the subscriber size is stable. Fig. 5 Subscription curve in Second Life as of Apr, 008 (source: Differences of subscription characteristics between competition-based and community-based MMORPGs Subscription curve during the lifetime of a game generally consists of four periods: forming period, growing period, maturing period, and recession period [1]. Different types of MMORPGs have different subscription curves (Fig. 6).

5 Issue 6 PENG Hui, et al. / Analysis of subscription characteristics and user mobility in different types of MMORPGs 15 Fig. 6 The life cycle of different types of MMORPGs 1) Forming period. Formation in community-based games is a long period. At the beginning, the avatars in MMORPG are scanty and the relationship among avatars is undeveloped, hence the game attractiveness is limited and the subscription grows slowly. However, the formation in competition-based MMORPG takes a short period, and the pleasant setting and engaging plot attract many players quickly. ) Growing period. The growing period curve of community-based MMORPG is sharp. Because once there are active players in the game, the game attractiveness increases rapidly, and new players join the game at an accelerating speed. As to competition-based MMORPG, the growing period curve is similar to the forming period curve. 3) Maturing period. For community-based MMORPG, if the operator does not adopt any strategy, the maturing period is still a longer one, which means that after the game has reached maturity, this state will continue for a long time. On the other hand, under the condition of operator not devising any strategy, the maturing period of competition-based MMORPG is very short, because the players lose their interest very soon in the fixed state of the game. Therefore, the maturity of competition-based MMORPG depends on the operator s game-maintaining strategy. Launching a new version and increasing the in-game figures level from time to time will maintain the players interest and desire. The subscription will increase temporarily, and then it will decline, showing the unique wave-shaped curve. 4) Recession period. The recession period curve of community-based MMORPG is flat at first, and then steep. Several players quitting the game will affect the game subscription a little at the beginning, but with more players quitting the game, the attractiveness of the game will be reduced, then a larger number of players will also quit. The recession period of competition-based MMORPG is usually relatively short. 4 Analysis of user mobility in MMORPG User mobility, which explains why subscription size changes, mainly involves three aspects: length of game playing time, the time point of players quitting the game and time point of operators shutting down a game [13]. The user needs to make a decision whether to play or quit, which will affect subscription size; the user also needs to decide the length of game playing time. The operator s logging into the game and the power to maintain the game will be affected by all the users decision. The relationship between the three aspects can bee seen in Fig. 7. Sects. 4.1, 4. and 4.3 explain the factors affecting the three aspects respectively. Fig. 7 User mobility in MMORPG 4.1 Length of game playing time The length of game playing time depends on two factors: the cost paid for playing one more unit of time (marginal cost) and the utility obtained from playing one more unit of time (marginal utility). When the marginal utility biger than the marginal cost, the player may choose to continue the game, until the marginal utility smaller than the marginal cost. The cost of game playing includes: electricity bill (a 1 ), network charge (a ), machinery depreciation (a 3 ), in-game consumer (a 4 ), a monthly subscription fee (e). a 1, a, a 3, and a 4 are in linear correlation to the length of playing game time

6 16 The Journal of China Universities of Posts and Telecommunications 009 (t), while the monthly subscription fee is a fixed cost, decided by the game operator. The total cost of a month is: C = ( a1+ a + a3 + a4) t+ e (5) Let A= a1+ a + a3 + a4, A is a constant. The marginal cost is: CM = a1+ a + a3 + a4 (6) The utility from online game playing includes a kind of pleasure, a sense of accomplishment and belonging [14]. The player obtains the highest utility at login time. With the passage of time, the player may feel tired of the in-world circumstances, and of the fatigue from long computer usage. Thus, the utility reduces with game playing time. Assume that the marginal utility is: UM = tb+ K (7) B represents the rate of marginal utility diminishing, which is related to the game types. For competition-based MMORPG, the game plot is set by the game operator and players can easily feel fatigue, thus the marginal utility diminishes fast (U M,1 in Fig. 8); while the game plot of community-based MMORPG is not set by default, hence the marginal utility diminishes slowly (U M, in Fig. 8). Fig. 8 Length of game playing time determined by marginal utility and marginal cost K expresses the intercept of marginal utility curve to the vertical axis. When the attractiveness of the game is strong, the marginal utility curve will shift upward, and K will increase. K keeps a positive correlation with the game attractiveness. According to the critical condition players exiting the game as in Eq. (8), one can calculate the length of game playing time as in Eq. (9). U = C (8) M M K A t = (9) B Eq. (9) shows that the length of game playing time depends on the type of game (which determine the level of B) and the attractiveness of game (which determine the level of K). Under the same level of attractiveness, the time of playing community-based game is longer than that of playing competition-based game ( t > t1). As the attractiveness of the game increases, the length of playtime also increases ( t3 > t). Assuming that the attractiveness curve is a parabola, it can be shown that the length of playtime will also follow a parabolic trend. 4. Time point of players quitting a game When paying monthly subscription fee, the players will re-assess the attractiveness of the game to determine whether they should continue to play the game or not. If the attractiveness of the game is greater than the monthly subscription fee, they continue to pay and play. If not, they stop paying the fee and quit. The authors set the attractiveness of a game at time T as G(T), which changes once a month. At the beginning of joining the game, the enthusiasm of participants is relatively high. With further exploration, the player s curiosity will increase, and the attractiveness of the game will also be stronger. At the peak time, if the game operator does not launch a new version of the game, the player s curiosity will start weakening, thus G(T) decreases. With the increase in game-play age, the attractiveness of the game increases at first and then declines, follows the parabolic shape [15]. Therefore, the authors assume that: GT ( ) = DT + ET+ F (10) The initial attractiveness of the game at the time of launching is F; while it changes with time. The movement of the attractiveness afterward is determined by two different forces: the player s psychological satisfaction from immersing in the game increases with time T at the rate of E, the player s curiosity from innovative plot decreases with the square of time T at the rate of D if no new version of the game has been launched yet. According to the critical condition of players quitting a game as in Eq. (11), one can calculate the time point of players quitting a game as in Eq. (1). GT ( ) = e (11) E + E 4 D( F e) T* = (1) D In Sect. 4.1, it has been concluded that the length of game playing time keeps a positive correlation with the game attractiveness in certain type of MMORPG. Moreover, as the attractiveness of the game increases, the number of users to whom the attractiveness of the game is greater than the

7 Issue 6 PENG Hui, et al. / Analysis of subscription characteristics and user mobility in different types of MMORPGs 17 monthly subscription fee will also grow, resulting in an increase of subscription size, and vice versa. The subscription size keeps a positive correlation with the game attractiveness. Hence, the bold line in Fig. 9 can also be seen as the curve of the length of game playing time, as well as the subscription curve under the condition of operator adopting strategies. Fig. 9 Time point for the player to withdraw from a game * The time point T calculated above means the time point at which players will quit a game if the game operator does not adopt any strategy at all. In fact, after the release of the game, the game operators will always make changes and updates to the game (such as Ⅰ in Fig. 6) before the players * reach the critical point T, re-igniting the curiosity of players and re-fuelling the attractiveness of the game. Hence the curve of game attractiveness is wave-shaped, shown as the bold line in Fig. 9. The subscription curve during the maturing * period is no longer smooth, and there is a bump, where T is the latest time of game update for the game operator to maintain a stable period. 4.3 Time point of operators shutting down a game The aim of a game operator running a game is to gain profit. With an increasing number of players quitting game or reducing game playing time, the income of the game operator becomes less. When the income is not sufficient to pay for the operating cost, the operator will be forced to shut down the game. The critical condition for a game to quit the market is: the operator s revenue equals the variable cost. The income R of a game operator includes the premium fee (e) and the in-game consumption (a 4 ); the premium fee is proportional to the number of users, which is w times the attractiveness of the game; the consumption maintains the proportion of d of the online time (t), which is r times the attractiveness of the game. It is clear that whether a game quits the market or not depends on the attractiveness of the game. The income of a game operator is: R = ( e+ dt) n= ( e+ drg) wg= ew+ drwg (13) The main variable cost of a game operator includes: labor cost, marketing cost and server cost, recorded as P. In accordance with the critical condition when a game quits the market: R = P, one can get critical attractiveness G at the time point game operator withdraws the game: 4eP dr ± ( dr) G = w (14) e The critical subscription size at which a game quits the market is: 4eP dr ± ( dr) n= wg = w w (15) e 5 Conclusions In conclusion, subscription characteristics in different types of MMORPGs are marked by dissimilarities, which are driven by user mobility. For competition-based MMORPG, the marginal utility as well as the attractiveness of the game diminishes fast, that is, the user mobility is high and the subscriber size is unstable with short stability period in subscription curve. Therefore, the operator should pay more attention to maintaining the game. As for the communitybased MMORPG, both the marginal utility and the attractiveness of the game diminishes slowly, which makes a low user mobility and stable subscription size with long stability period. MMORPGs. Acknowledgements This work was supported by the Teaching and Research Program for Overseas Returnees in Beijing University of Posts and Telecommunications. Appendix A Monthly data of Second Life Table A.1 Monthly data of Second Life Date Volume of User to User Number of premium transactions accounts Jan Feb Mar Apr May Jun Jul Aug Sep Oct

8 18 The Journal of China Universities of Posts and Telecommunications 009 Continued Date Volume of User to User Number of premium transactions accounts Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Data source: Appendix B Weely data of HiPiHi Table B.1 Weely data of HiPiHi Date Virtual items Subscription size Nov 30, Dec 7, Dec 14, Dec 1, Dec, Jan 1, Jan 18, Jan 5, Feb 1, Feb 8, Feb 15, Data source: References 1. Yoon U G. A quest for the legal identity of MMORPGs From a computer game, back to a play association. Journal of Game Industry & Culture, 005, 10 (Fall). Simpson Z B. The in-game economics of Ultima Online. Proceedings of Computer Game Developer s Conference, Mar 9 14, 000, San Jose, CA, USA Cornett S. The usability of massively multiplayer online roleplaying games: designing for new users. Proceedings of ACM CHI 004 Conference on Human Factors in Computing Systems, Apr 4 9, 004, Vienna, Austria. New York, NY, USA: ACM, 004: Harding-Rolls P. Subscription MMOGs: life beyond world of warcraft. London, UK: Screen Digest, Nardi B, Harris J. Strangers and friends: collaborative play in world of warcraft. Proceedings of the 0th Anniversary Conference on Computer Supported Cooperative Work (CSCW 06), Nov 4 8, 006, Banff, Canada. New York, NY, USA: ACM, 006: Cornes R, Sandler T. The theory of externalities, public goods, and club goods. London, UK: Cambridge University Press, 1996: Katz M L, Shapiro C. Network externalities, competition, and compatibility. The American Economic Review, 1985, 75(3): Michael R, Wagner J A, Mark W, et al. Second Life: the official guide. New York, NY, USA: Wiley-Interscience Press, 006: 4 9. Hong J C, Koch C, Kuhn D. Market potential for online games in North American. Telecom Market, 004(6) (in Chinese) 10. Castronova E. Synthetic worlds: The business and culture of online games. Chicage, IL, USA: University of Chicago Press, 005: Wooders M, Cartwright E, Selten R. Social conformity in games with many players. FEEM Working Paper No Antràs P. Incomplete contracts and the product cycle. The American Economic Review, 005, 95(4): Halepovic E, Williamson C. Characterizing and modeling user mobility in a cellular data network. Proceedings of the nd ACM International Workshop on Performance Evaluation of Wireless Ad Hoc, Sensor, and Ubiquitous Networks (PE-WASUN 05), Oct 10 13, 005, Montreal, Canada. New York, NY, USA: ACM, 005: Hussain Z, Griffiths M D. Gender swapping and socializing in cyberspace: an exploratory study. Cyber Psychology & Behavior, 008, 11(1): Yee N. Motivations for play in online games. Cyber Psychology & Behavior, 007, 9(6): (Editor: WANG Xu-ying)