What to Expect from Fits using <K>

Last Update: 02-Aug-05

The critical point here is the relationship between Lxy or VPDL and the effective ct0 used in the predictions:

P^unm = ( K/cτ ) ( M/P_T^m ) exp[ -( K/cτ ) ( Lxy M/P_T^m ) ] { 1 + D cos[ (dm/c) K ( Lxy M/P_T^m ) ] }
VPDL = Lxy M/P_T^m

Source K Contrib to Var from Contrib to Var Spread

Variable: VPDL histo bin

Data many ct0's K

Pred conv many ct0's K

Pred <K> single ct0 ---

Variable: Lxy histo bin

Data many ct0's K

Pred conv many ct0's K,P_T^m

Pred <K> many ct0's P_T^m

So - the bottom line is that chi2 fits to the asymmetry using <K> will return pred's without any smearing of VPDL at high values, while logL fits will have some effective smearing of the predicted Lxy histogram.

Source	K	Contrib to Var from	Contrib to Var Spread
Variable: VPDL histo bin
Data		many ct0's	K
Pred	conv	many ct0's	K
Pred	<K>	single ct0	---
Variable: Lxy histo bin
Data		many ct0's	K
Pred	conv	many ct0's	K,P_T^m
Pred	<K>	many ct0's	P_T^m