### Table 3 Linear relationship between the integrated results and the number of concurrent players

in M. Ye

"... In PAGE 9: ... On the other hand, if every server followed Equation 10, the number of concurrent players for the entire game world would follow Equation 13. Table3 shows the results for the entire game world. Cov(k0, k) indicates that the linear relationship is improved (see the brown line in Figure 4).... ..."

### Table 1 Linear relationship and parameters of network traffic and number of concurrent players

in M. Ye

"... In PAGE 8: ... Two network traffic metrics were analyzed: bytes per second and packets per second. Table1 shows the results, a strong linear relationship between these metrics. Cov is the correlation coefficient calculated according to Equation 11, and uk is the parameter in Equation 7 estimated by the robust regression algorithm.... ..."

### Table 2 Server performance vs number of concurrent players using robust regression algorithm to esti- mate parameters of Equation 8

in M. Ye

"... In PAGE 9: ... Once U(t) is obtained, the number of concurrent players k(t) can be calculated by performing integration (see Equation 12). Table2 summarizes the results, which show that almost all of the servers have a strong linear relationship with the number of concurrent players except game server 4. Figure 4 plots the robust regression result of the servers.... In PAGE 9: ... Therefore, whether the load-balancing algorithm for the game server is proportion-consistent is decided by the geographical distribution of players. This is why the correlation coefficients of game servers are poorer than those of proxy servers (see Table2 ). On the other hand, if every server followed Equation 10, the number of concurrent players for the entire game world would follow Equation 13.... ..."

### Table 6 Linear relationship and parameters of network traffic and the number of concurrent players for (a) Server 1; (b) Server 2; (c) Server 3; and (d) Server 4.

in M. Ye

"... In PAGE 13: ...Table6 shows the results for all four game servers, indicating a strong linear relationship between them. Cov is the correlation coefficient calculated according to Equation 11, and uk is the parameter in Equation 7 estimated by the regression algorithm.... ..."

### Table 1. Average length of time both players were active (concurrent interaction), one player was active (sequential interaction), or neither player was active. Note, the total session time was 600 seconds.

"... In PAGE 11: ...ime), the amount of time users interacted sequentially (i.e. only one player active), as well the amount of time when neither partner was active2. The results for the three experimental conditions are shown in Table1 . Figure 3 shows three segments from the video annotation system timeline that illustrates the concurrent nature of interactions in the paper condition and the two-mice condition, compared to the forced sequential interactions in the one-mouse condition.... ..."

### Table 1. Average length of time both players were ac- tive (concurrent interaction), one player was active (sequential interaction), or neither player was active. Note, the total session time was 600 seconds.

2000

"... In PAGE 4: ...cted sequentially (i.e. only one player active), as well the amount of time when neither partner was active. The results for the three experimental conditions are shown in Table1 . Figure 3 shows three segments from the video annotation system timeline that illustrates the concurrent nature of interactions in the paper condition and the two-mice condition, compared to the forced sequential interactions in the one-mouse condition.... ..."

Cited by 14

### Table 4 Server performance and number of con- current players

in M. Ye

"... In PAGE 10: ... From these figures, a strong linear relationship between the number of concurrent players and the CPU utiliza- tion can be clearly seen. Table4 summarizes the results, which show that all of the servers have a strong linear relationship with the number of concurrent players. We also summed the total number of players for the four game servers and evaluated the relationship between the total number of players and CPU utilization.... ..."

### Table 6: Units Won by each player (Player 1 is the evolving player)

2001

"... In PAGE 10: ...Figure 5: Learning Curve, against Tight Aggressive Players Table6 shows the results when the evolved player (i.e.... ..."

Cited by 10

### Table-2: Players and their roles Player Roles

### Table 6: Thread Concurrency (%)

1998

"... In PAGE 20: ... The reason is that most of the time the main thread runs alone or concurrently with a few nanothreads. Table6 shows the thread concurrency. i thread means that the main thread plus (i -1) nanothreads are active.... ..."

Cited by 53