Method:

Shooting gammas with pt of 15 GeV at sector 11 in the center of a tower (phi=pi/2) of the endcap at an eta of 1.3 (where scale factor is relatively flat vs. thrown energy in previous tests).  Looking at reconstructed tower energy. (Note this pT, eta corresponds to thrown energy of ~30 GeV).  Running with either stand-alone endcap or full STAR geometry.  I do not use the LOW_EM option.  I did lots of "diff"'s on the files and tried to find the changes.  There were so many cosmetic changes with each revision this was quite difficult to do.  So my list changes are only guaranteed accurate at the Wikipedia level.

Version           EEMC Only            Full                  Note

1. cvs        28.4                    28.7                 CAIR fix
2. v23       27.7                    27.3                 latest version
   
3. v1          28.4                    28.7             cosmetic changes after CAIR fix
4. v2         27.9                    28.6                 block EXPS added  = G10 layer in shower max                   
5. v3         28.0                    28.0                 ERAD change to stainless steel, ELED change to pballoy
   
6. v4         27.9                    27.7                 tie rod rmax change from 1.0425 to 1.50/2 (block ERCM), 3 degree cut on either side of spacer (block EXPS)
7. v5         27.9                     27.7            seems to only be cosmetic changes
8. v6         27.9                     27.7            cosmetic changes + a few print statements
9. v7         28.0                     27.6            more cosmetic changes
10. v8         27.9                     27.5             EFLS,EBLS added, ESMD options added, GapSMD changed from 3.4 to 3.6, spacer layers extended to 5 degrees on either side
11. v9        30.2                      29.7             ecal_set_cuts added, pre/post shower have same cuts as megatiles for particle tracking
12. v10      XX                        XX               Rtie added as option for block EMCS (would not compile, did not test)
13. v11      32.3                     32.2            refinements to material in front and back of SMD plane
14. v12       XX                       XX            "Medium standard" replaced by "Medium ecal_standard"    refinements to building of SMD planes (would not compile)
15. v13      XX                        XX            seems mostly cosmetic - "plastic" replaced by "megatile" (crashes), "Medium ecal_standard back to "Medium standard"
16. v14       XX                       XX                       cosmetic changes, changes to way the drawing is done
17. v15      XX                        XX                      StrAluminum replaces Ecal_Al is EALP (crashes)
18. v16      27.5                     27.3                     seems cosmetic - jiggle option is added, but set to 0, StrAluminum removed in EALP - add AluPlate to this block to get it to run, block ESCI is changed (not sure of effect).   At v16 we are getting very close to what is observed in the final version, but I had so many versions not run it's unclear what is responsible for the change. I have tried to disentangle the changes, but cannot create versions which compile.  Know that it is not changes to EALP that are responsible.
19.  v17     27.5                     27.3                      changes to construction of SMD planes (ecal_get_strip replacing code is EXSG - test version)
20. v18      27.7                      27.2                      ecal_get_strip implemented
21. v19      27.7                     27.2                    cosmetic changes
22. v20      27.7                     27.3                          cosmetic changes
23. v21      27.7                     27.2                          deltaz shifts added to move megatiles, pre,and post shower layers
24. v22      27.7                           27.4                    change to PVC, ESPL and EXSG now have kOnly = "MANY " option, EHMS, EBLS now have kOnly = "Only" option.  These changes change the way the spacer overlaps with the SMD.
25. v23      28.4                     27.9                         scintillator thickness changed to 0.475 from 0.5, latest version without new layers in place

 Next I try to make changes to cvs version to see what biggest effects are.

1. Change ERAD block.  No effect
2. Keep change to ERAD block + add change to ELED block.  See a change where sampling fraction drops for eemc stand-alone vs. full
3. Keep changes to ERAD+ELED block.  Changed cuts to be the same for pre, post and megtiles = cuts used in Jason's file.  This causes reco energy to drop again.
4. Tried changing the tie rod radius - has no effect.

Notes for me:

To do full STAR geometry .kumac file looks like:

GDrop all
DetP geom y2006g
make geometry
GClos all

To do eemc only geometry .kumac file looks like:

GDrop all
DetP geom y2006g
make geometry/cavegeo
make geometry/ecalgeo

For versions> 15 and <23 these changes to ecalgeo.g are also needed:

1. Change IF(emcg_OnOff==2) THEN... to  if (emcg_OnOff==2 | emcg_OnOff==3) then
                Position ECAL in CAVE z=-center ThetaZ=180
           endif
    
2. In EALP block add material AluPlate isVol=0