[molpro-user] Energy doesn't match wrt to opt and Freq for BH2 radical in SA-CASSCF

[Gerald Knizia]

On Wednesday 05 January 2011 20:04, t s venkatesan wrote:
> I am doing a SA-CASSCF optimization and frequency calculation for BH2
> radical. Unfortunately, the total energy doesn't seem to match.  [...]
>
> Optimization point 5
>
>  Variable                      Last           Current        Next
>
>  E(MULTI000) / Hartree     -25.79212071   -25.79212099     0.00000000

Maybe I'm misunderstanding you, but your input does seem to work fine for me. 
The multi following the optg says

 !MCSCF STATE 1.1 Energy              -25.792120998771
 !MCSCF STATE 2.1 Energy              -25.744937326564

And the first of which (which is the state energy the optg was done for) 
agrees with the value optg obtains. Agreed, the following line is somewhat 
confusing:

>        MULTI        OPTG(MULTI)        MULTI           HF-SCF
>     -25.74493733    -25.79212100    -25.73507280    -25.43015545

But to me it seems that the issue is simply that the energy reported in these 
timings/statistics lines is the last energy calculated by the program. In 
this case this would be the energy of the first excited state---which is 
calculated last, but which is not the energy actually used in the geometry 
optimization.
-- 
Gerald Knizia