Hello,

In ITensor3 tutorial/finiteT/metts.cc, what puzzled me is "psi.ref(1) /= norm(psi(1));". Does this relate to the probability {<cps|exp(-0.5*beta*H)|cps>}^{-0.5}(New J. Phys. 12(5), 055026(2010))? If this is true, how to get the result and why not the overlap of two MPS? Does this apply to different models?

```
println("Doing time evolution");
for(int tt = 1; tt <= nt; ++tt)
{
psi = applyMPO(expH,psi,args);
psi.ref(1) /= norm(psi(1));
}
```

Thanks

Zong-Sheng Zhou