I'm posting this on behalf of kamikaze viking.
Kami's Lookup Table PSR Shift Proposal.
the reasoning behind My proposal is
a) To fit within the scope given. Minimal work for Best effort gains.
A compromise to cater to both camps of "everything is Win/loss" and "everything is personal skill"
I realised early in the discussion regarding how Match Score affect PSR movement that due to score increase based on the WIN The current system is 'double dipping'. Hence the Solution appeared to me to be to neutralise the Raw Match score by using a relative system separated by team. This would seem to be to cancel out the Match Score granted for winning, so that the total MS for each player reflects better on their skill within the team.
The result is that this system grants a factor for the WIN/loss as well as your personal skill relative to your team, and these should be easily tweakable within this system (values within a lookup table).
Method
==============
This graphic may be familiar to some as ive posted it a few times in recent weeks.
https://i.imgur.com/nSl4rQ9.jpg
PSR Shift Values
================
These are base shift values - totalled as zero sum.
Win = -1 +0 +1 +2 = +1
Loss = -2 -1 +0 +1 = -1
Win + loss = +1 -1 = 0
The higher the number the faster people will move through the psr range per match. initial Simulations they found that 0, 12 & 24 worked best (and fit with Pauls proposed range).
Applying this across the match. Many systems suggested will sort all 24 players and then rank them, mine MUST keep the two teams separate and rank them by MS within their team.
==================
When applied to each team in a 3/3/3/3 team split
Win = -1x3 +0x3 +1x3 +2x3 = -3 +0 +3 +6= +6
Loss = -2x3 -1x3 +0x3 +1x3 = -6 -3 +0 +3= -6
Win + loss = +6 -6 = 0
Example Match using 3/3/3/3 split
https://i.imgur.com/FJoNrQO.jpg
==================
As Jay Z was testing in his simulator (as seen in this post
https://mwomercs.com...66#entry6337266), he implemented this using a lookup table. Which showed once tweaked, a more granular version seemed a better distribution and still remains zero sum. I derived my own version as can be seen below;
Winning team PSR Shift
|
24 | 24 |
24 | 20 |
24 | 16 |
12 | 12 |
12 | 8 |
12 | 4 |
0 | 0 |
0 | 0 |
0 | 0 |
-12 | -4 |
-12 | -8 |
-12 | -12 |
Losing team PSR Shift
|
12 | 12 |
12 | 8 |
12 | 4 |
0 | 0 |
0 | 0 |
0 | 0 |
-12 | -4 |
-12 | -8 |
-12 | -12 |
-24 | -16 |
-24 | -20 |
-24 | -24 |
Sum = 0
Edited by Dogmeat1, 13 June 2020 - 02:55 AM.