Shower shape variables show unusual distribution when simulating strips in each layer

[SmartinaK]

To see how the shower shape variables change if the strip layer is pushed even further back, instead of only setting one layer to be a strip layer and running that configuration for each layer, I run one simulation and set strips in every layer. Since I'm not adding any noise or cross talk, and I only want to look at the shower variables which use cells in the respective layer only, I thought that the segmentations in each layer shouldn't affect each other.
However, there is a different distribution for the front side energy ratio for the photons, looking at the SW.
What I mean by "different distribution": Looking at the same layers, but in the first plots (10GeV) I set only strips in the third layer, while for the second plots I put strips everywhere. In the fside plot for photons (upper right), the peak is not concentrated around 0 anymore. Which is the distribution I see across other energies too. What is the reason for this?

[zwu0922]

Hi everyone, I tried to produce distributions of these 6 shower shape variables with the two settings of readout segmentation. For the 3rd layer (or "layer 2" in my version), distributions are compatible between two segmentations, up to the effect of cluster reconstruction efficiency. Please find the following attached plots.
SummaryPlots.pdf
After a discussion with Martina, we think the issue is most likely from a bug in the plotting code. The calculation of shower shape variables in the full simulation code seems OK.

[giovannimarchiori]

Thank you much @zwu0922 for following up on this and investigating with @SmartinaK . I will close the issue then