I was probably not too clear, sorry! For geeks and formula wranglers
-
CSR:
For 4+ races, we use 100% of line 1 SR + 62% of (line1+line2)/2 + 38% of (line1+line2+line3)/3 + 24% of (line1+line2+line3+line4)/4
For 3 races, we use 100% of line 1 SR + 62% of (line1+line2)/2 + 38% of (line1+line2+line3)/3 + 24% of (line1+line2+line3)/3
For 2 races, we use 100% of line 1 SR + 62% of (line1+line2)/2 + 38% of (line1+line2)/2 + 24% of (line1+line2)/2
For 1 race, we use 100% of line 1 SR + 62% of line1 + 38% of line1 + 24% of line1
100-62-38-24% is a Fibonacci progression (an organization of natural and mathematical phenomena). Works for me ...
It's a smoothing and averaging of recent lines which conceptually depends on how useful the Adjusted Speed Rating concept itself is - i.e. the distance equalization adjustments, the daily-track and inter-track adjustments and the surface to surface adjustments turf/dirt/poly (which is really just another track to track adjustment).
Of course no set of adjustments is perfect, hence the minimal ability to exclude one
low rating (especially in Top 2) which disproportionately affects CSR. But this concoction seems to have pretty good legs, especially when married with other ratings, and
especially with other
non-correlated ratings (e.g. an oddsline built of not only speed/pace/deceleration factors, but also positional pace, class, public ratings like Prime Power/Profit Line/Morning Line -- and public tote odds --> hence
Rx).
Hope that helps.
Ted