Suggestions on CE of hcp system

Dear Axel,
Can you please have a look at the attachment which shows the CE of a hcp system. The fit.out is also attached. Now, when I used checkrelax, this is what I got…
0.0004 1/str_relax.out
0.0297 52/str_relax.out
0.0322 62/str_relax.out
0.0369 49/str_relax.out
0.0389 201/str_relax.out
0.0396 4/str_relax.out
0.0469 262/str_relax.out
0.0488 20/str_relax.out
0.0512 12/str_relax.out
0.0623 51/str_relax.out
0.0644 44/str_relax.out
0.0685 60/str_relax.out
0.0707 9/str_relax.out
0.0731 258/str_relax.out
0.0744 24/str_relax.out
0.0755 6/str_relax.out
0.0761 50/str_relax.out
0.0794 135/str_relax.out
0.0827 158/str_relax.out
0.0835 57/str_relax.out
0.0893 41/str_relax.out
0.0898 48/str_relax.out
0.0910 209/str_relax.out
0.0913 2/str_relax.out
0.0951 56/str_relax.out
0.0956 3/str_relax.out
0.0957 156/str_relax.out
0.0983 55/str_relax.out
0.0985 54/str_relax.out
0.1009 10/str_relax.out
0.1029 144/str_relax.out
0.1054 252/str_relax.out
0.1070 46/str_relax.out
0.1080 39/str_relax.out
0.1080 47/str_relax.out
0.1086 53/str_relax.out
0.1097 5/str_relax.out
0.1141 104/str_relax.out
0.1141 7/str_relax.out
0.1147 0/str_relax.out
0.1191 251/str_relax.out
0.1219 161/str_relax.out
0.1269 199/str_relax.out
0.1320 136/str_relax.out
0.1333 250/str_relax.out
0.1434 176/str_relax.out
0.1456 40/str_relax.out
0.1993 16/str_relax.out
0.2317 138/str_relax.out
0.5217 8/str_relax.out
0.5448 43/str_relax.out
0.6190 42/str_relax.out

Th ‘0’ structure has a strain (factor or coefficient, whatever we may say !!) of 0.1147…I have used structures which have strain ‘factors’ below 0.1 except 0…which has to be used ? Am I correct…
The maps.log shows a CV score of 0.0160831. There are no ground states predicted which seems to be correct.
Can you please comment on the ‘known gs’ line. Is the CE ok?

Thank you

Suddhasattwa

The checkrelax command output for hcp structures may be a bit pessimistic.
It counts c/a distortions from what is specified in the lat.in as a relaxation.
Technically, c/a distortions don’t break the lattice’s symmetry so any number would be allowed.
I did account for this in checkrelax because very large deviations in c/a should be double checked a bit.
So in this case I would plot the output of checkrelax as a function of composition and see if there are outliers then. If the structures just follow the trend you can expect from the end members 0 and 1, then I would not worry about excluding structures.
You "known gs" line looks reasonable to me.