The reason why csg_stat and g_rdf work differently can be found here:
<
https://groups.google.com/d/msg/votca/UubXtM0GKvc/TDzTG2LNfbkJ>
In short, when you use csg_stat with a mapping (even if it is a 1:1
map) nrexcl get wiped out. This is due to the fact that a mapping
creates a new (exclusion free) topology.
After the mapping VOTCA adds back exclusions for all beads, which are
part of a bonded interaction defined in the mapping file.
In contract when using csg_stat without a mapping file, exclusions are
read from the tpr file.
So to get the same exclusions with and without a mapping you need to
add an extra block with excluded pair to the mapping file.
Compare the output of:
csg_dump --top topol.tpr --excl --cg "PAMcg.xml;watercg.xml"
vs.
csg_dump --top topol.tpr --excl
In your case you need to add:
--- PAMcg.xml.orig 2016-02-16 15:32:09.000000000 -0700
+++ PAMcg.xml 2016-02-16 15:49:19.000000000 -0700
@@ -531,6 +531,39 @@
B29 A29 B30 A30
</beads>
</dihedral>
+ <bond>
+ <name>exclusion</name>
+ <beads>
+ A1 B3
+ A2 B4
+ A3 B5
+ A4 B6
+ A5 B7
+ A6 B8
+ A7 B9
+ A8 B10
+ A9 B11
+ A10 B12
+ A11 B13
+ A12 B14
+ A13 B15
+ A14 B16
+ A15 B17
+ A16 B18
+ A17 B19
+ A18 B20
+ A19 B21
+ A20 B22
+ A21 B23
+ A22 B24
+ A23 B25
+ A24 B26
+ A25 B27
+ A26 B28
+ A27 B29
+ A28 B30
+ </beads>
+ </bond>
</cg_bonded>
</topology>
<maps>
Christoph