DISCLAIMER: Obviously, with only 1 week of data, it can be inaccurate to extrapolate, which is why I’ll continue updating this post as new data become available.
Steam peak users per day for ToS (data on concurrent users @ 13:00 hours each day)
For the linear fits: y is the number of active users @ 13:00 and x is consecutive days. Projections based on the previous weekly linear model will be denoted with “[ ? ]”
DISCLAIMER 2: Projections especially based on linear trends will likely be highly inaccurate. It is nonetheless interesting to see whether a simple linear model might have predictive power for ToS active user trends. Trend predictions will be crossed out and updated with an +/- error and actual value when new data are available.
Week ~1 of ToS (Founders):
SAT 4/02/2016 33437
SUN 4/03/2016 31023
MON 4/04/2016 27172
TUE 4/05/2016 28953
WED 4/06/2016 26793
THU 4/07/2016 26392
FRI 4/08/2016 26304
Linear fit for Week 1: y = -1108.6x + 5E7 (R^2=0.767)
Week ~2 of ToS (Founders):
SAT 4/09/2016 [25195?] 29950 (underestimate by 4755) this number is prior to the server issues
SUN 4/10/2016 [24086?] 29804 (underestimate by 5718)
MON 4/11/2016 [22977?] 24359 (underestimate by 1382)
TUE 4/12/2016 [21868?] 24055 (underestimate by 2187)
WED 4/13/2016 [20769?] 21869 (underestimate by 1100)
THU 4/14/2016 [19650?] N/A (Servers were down @ 13:00)
FRI 4/15/2016 [18541?] 21460 (underestimate by 2919)
Linear fit for Week 2:
Week ~3 of ToS (Founders):
SAT 4/16/2016 24600
SUN 4/17/2016
MON 4/18/2016
TUE 4/19/2016
WED 4/20/2016
THU 4/21/2016
FRI 4/22/2016
Generally for other games, you see a spike of user activity corresponding to Friday and Saturday, with a decrease on Sunday and a huge decrease on Mondays. Then, lull in user activity around Tuesday/Wednesday but building back up on Thursday. Basically looks like a parabola. The point is that usually linear fits don’t work well and result in R^2 values under ~0.3 for most other games. It’s interesting that there is such a high R^2 for ToS.
What does it mean? It’s perhaps too early to conclude, so stay tuned for next week I guess!



