All the hullabaloo about the PvE system tests and what people really want has gotten me thinking and I’d like some feedback.
Consider this:
- A given event or sub has a ceiling of points which can be obtained;
- To get good placement rewards, one has to be as close as possible to this ceiling;
- There is evidence that part of the population will do whatever is needed to approach this ceiling, including waking up at 4 AM to play or spending 6 hours grinding the same node;
- A high ceiling will not only affect the upper part of the placement rankings but also increase the effort needed to be in the top 50 or 100.
- From the devs perspective, the ceiling cannot be too low otherwise players will not spend enough;
- However, a ceiling that is too high will just make players give up on placement rewards.
A points ceiling is traditionally determined by several factors:
- Timing (potentially including rubberbanding)
- Number of full clears (defined here as beating all nodes, including essentials, exactly once)
- Scaling
These are some of the main factors that have changed from system to system when it comes to placement rewards. Scaling, while an important factor, is not really limiting the ceiling so far; the best players will always be able to beat the hardest nodes as much as necessary as of today. Let us focus on timing and number of clears.
Here’s an attempt to simplify the calculation of the points ceiling:
Let “c” be the value obtained for a full clear; “t0” a time point at the beginning of a sub, “t1” = t0 + 8h, “t2” = t0 + 16h, “t3” = t0 + 24h; “f” the “final grind” needed to clear all the nodes as much as possible (e.g. to 1 point) and maximise points.
With the old system (A), the point ceiling for a 24 sub would be approached optimally by the following strategy:
t0: 1 c
t1: 1 c
t2: 1 c
t3: f
Ceiling ~= 3c + f
According to the new system (B) that was tested before the Rocket & Groot event:
t0: 6 c
t3: f
Ceiling ~= 6c + f
And finally, according to the newest system (C):
t0: 4 c
t3: f
Between t0 and t3: beat easiest node as often as possible for 20 pts.
Ceiling ~= 4c + f + 20x, where x is the number of times a human can possibly beat the easiest node within 24h.
Now, if we assume that the time needed to do 1 full clear is 20 minutes, f = 4c, and one could spend up to 6 hours beating the easiest node before his/her hands fall off:
Roughly, system A requires 140 minutes; B requires 200 minutes; and C requires 520 minutes.
In addition, if there are two 24h subs in a row, the time allocated to consecutive clears between t3 of the first sub and t0 of the second would be:
A: 100 minutes
B: 200 minutes
C: 160 minutes
I hope these models are accurate enough to draw a few conclusions with regards to placement rewards:
- C is by far the most time consuming (broken!);
- A is the least time consuming both in total minutes and in terms of consecutive minutes between subs;
- B and C provide a little more schedule flexibility than A by requiring grinds only at t0 and t3 (no t1 and t2);
- However, A can be done with more flexibility by going t0: 1c, t1: 2c and t3: f; or t0: 2c, t2: 1c and t3: f; or even t0: 3c, t3: f, with only a small loss in the number of points compared to the ceiling.
Am I thinking about this too much? 
