Key\\ 
PU2 - Blue\\ 
PU20 - Green\\ 
PU40 - Red\\ 
Energy [Gev]\\ 

{{:groups:cms:walstrom:dr2.gif|}}\\ 
This plot shows the distance between matched electrons and superclusers (dR). This shows a high percentage of good matches because the majority of dR values are quite small. Pile up appears to be related to closer matches and with significant pile up there appears to be no cases of high values of dR. This is attributed to the creation of superclusers by the particles created by pile up. As the number of pile up clusters goes up it increases the likelihood that there will be a pile up supercluster that is close to the electron. When this happens it will appear on the fractional resolution plots as an entry close to 0 GeV due to the probable low energy of the supercluster.

{{:groups:cms:walstrom:dr_close2.gif|}}\\ 
This shows a range from the same data as above (0.0 - 0.1). This view reveals the shape of the dR peak at low values. As pile up increases the number of matches with a dR value in the range increases, however the general shape of the peak is the same across different pile up values.

{{:groups:cms:walstrom:dr_match2.gif|}}\\ 
This is a plot of the same total data with the peak close to zero removed. There are no high value dR values for pile up above two because there is almost always a pile up supercluster which is close to the electron.

{{:groups:cms:walstrom:eta.gif|}}\\ 
This plot shows the |h| values of the total supercluster collection from reconstruction in ECAL. For pile up 2 There is a fairly constant value of superclusters across eta, the high eta range still contains more due to the pile up 2 effect. For pile up 20 and 40 the shape of the distribution of the superclusters across |h| changes drastically. The trend fits a constant within the barrel abs(|h|) < 1.48. There is a nontrivial peak at the outer edge of the endcap and another peak at the inner edge of the endcap at high values of |h|. Due to no obvious physical reason why there would be more pile up superclusters in those regions it is hypothysised to be an over correction from reco in CMSSW. Partly constructed superclusters and shifted to the first detectable |h| value in the endcap. Across the endcap the number of superclusers trends upward with increasing abs(|h|).

{{:groups:cms:walstrom:deta_match.gif|}}\\ 
This plot shows the d|h| value that is taken from the matching algorithm. This is done to check if the specific d|h| or dphi has a different shape than the calculated dR value. The peak and trend shown here is the same pattern seen in the dR plot. the d|h| does not appear to display any different behavior from the behavior of the dR value. 

{{:groups:cms:walstrom:dphi_match.gif|}}\\ 
This plot shows the dphi value from the matching algorithm. It again appears to have the same trends as the dR data. Pile up appears to effect dphi just as it effects the total dR value.