For discussion in HF PWG. 

I) 3D DCA Issue. 

We extract 2D DCA_XY-DCA_Z from data at the point of closeset approach to the primary vertex (KFVertex) in the XY plane. 
The data DCA_XY-DCA_Z distributions are the only input to the fastSimulator. They are binned in Vz, η, Centrality and pT.
Using 2D DCA distributions from data preserves any correlation (mis-matched tracks DCA_XY and DCA_Z are correlated).
We then compare the output 3D DCA from data and fastSim.

Centrality 0-5%:
 - low-pT (ZOOM).
 - high-pT (ZOOM).

Centrality 5-10%:
 - low-pT (ZOOM).
 - high-pT (ZOOM).

Centrality 20-30%:
 - low-pT (ZOOM).
 - high-pT (ZOOM).

Centrality 60-70%:
 - low-pT (ZOOM).
 - high-pT (ZOOM).


II) Topological variables comparison to data:
All centralities.


III) On event counting and vertex resolution of peripheral events:

Vertex resolution from sub-event method:
 - ΔX
 - ΔY
 - ΔZ

Issues with peripheral events:
1) As seen in the plots above. The vertex resolution in peripheral events is very bad. ~35% of peripheral events vertices are out of 150μm vertex resolution. These are not likely to contribute to our D mesons foreground (maybe not even the bakground). To correctly count the number of peripheral events we need to understand the vertex resolution and place a cut on the subEvents' vertecies distance. 

2) To solve the DCA problem we had to use 2D DCA (XY-Z) distributions from data as input to our Fast Simulator. Therefore, it is not easy to unfold the vertex resolution (unfolding from 2D distributions is not straightforward). 

3) Because the vertex resolution in peripheral events is at the same order of the decay length of D0 unfoldig is not reliable (subtracting two numbers that are close to each other have very large uncertainties).


-- Mustafa^2