I made another graph which might be interesting to some of you. Technically I made it to further my point about how long it will take for lowered player input to have a noticeable impact on player population, but I think it can be interesting elsewhere too.
This graph shows how old (in days) the players who played in the last 24 hours. This can however easily only be made for "today", whatever today is when the graph is made. It can be done for other points in time too, however it's much harder. And I think this sufficiently proves the point.
As you can see, the vast vast vast majority of players is VERY FUCKING OLD. There is a tiny spike in the 0, 1, 2, 3 day age area (10, 6, 4, 1 player(s) - respectively), however that is really not anywhere close to enough to maintain the long tail.
Statistically speaking, the dropoff is like this (actual numbers)
out of 29315 players who have ever connected to TGS servers:
5813 connected only once (20%; 80% remain)
12084 did not connect after 24h of their first connection (41%; 59% remain)
14568 did not connect after 7 days of their first connection (50%; 50% remain)
17115 did not connect after 30 days of their first connection (58%; 42% remain)
So you can expect to keep about 42% of people for more than a month. These percentages are not 100% correct, since I didn't reduce the 2nd number to account by the number of players who have not yet actually had the chance to reach this goal, but the difference is negligible. Basically the 29315 value would go down by about 400ish in the 30 days calculation, which would change fuck all in the percentage calculations.
Also, here's that same graph from above, just with people of various ages clumped into groups. Again, it shows how old the players who played in the last 24 hours are.