minor change in ESS estimation in mcse_sd based on more simulations

[avehtari]

Summary

mcse_sd was estimating ESS from abs(x-E[x]). In my old simulations, this had worked better than (x-E[x])^2. It's likely that the change in how the rest of MCSE for sd is estimated (PR #233) did affect that now the more logical (x-E[x])^2 works better, as demonstrated by simulations run by @sethaxen.

Here are the plots for the new simulation results

Copyright and Licensing

By submitting this pull request, the copyright holder is agreeing to
license the submitted work under the following licenses:

• Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
• Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

[paul-buerkner]

Thanks. Looks good.

[paul-buerkner]

Merging now.