×
The RPI functions as an active indicator of current performance, separate from lifetime cumulative points. For every match, the system calculates a shared Match RFI Quality baseline ($Q$):
Q = (RFI_Winner + RFI_Loser) / 2
This quality profile defines a maximum cap for performance target scaling using a $+1.0$ offset parameter:
Ceiling_Tier = min(8.5, (Q / 715.0) + 1.0)
Using the actual match point win ratio ($W_{act} = Points_{Win} / Total_{Points}$), the engine calculates separate Performance Targets for both competitors:
Target_Winner = min(Ceiling_Tier, ((RFI_Loser / 715.0) + 1.0) + (W_act - 0.5) * 1.5)
Target_Loser = min(Ceiling_Tier, ((RFI_Winner / 715.0) + 1.0) - (0.5 - W_lost) * 1.5)
The player's index converges toward this target using an exponential filter based on game experience ($K_{rpi} = 0.20$ if matches $< 15$, otherwise $K_{rpi} = 0.06$):
RPI_Next = RPI_Current + (Target - RPI_Current) * K_rpi
If an active player's match history shows gaps of over 30 days, they experience a monthly time slip penalty of $-0.05$. Finally, hard volume caps limit progression flags for lower sample sizes ($5.8$ cap under 10 matches; $6.8$ cap under 20 matches), with an absolute lower boundary guard of $2.0$.