"No data" data points in the experimental data changes the FoM

**Description of the error**
In `chi/expt-plus-no-data`, I have created a file (MDMCv0.2_pilot\doc\tutorials\data\Well_s_q_omega_Ar_data_no_data.xml) which adds a bunch of "no data" points between q=0.42 and q=0.84 to the original (MDMCv0.2_pilot\doc\tutorials\data\Well_s_q_omega_Ar_data.xml).

Running the FoM calculation with epsilon=1.08, sigma=3.4 shows differences between the two. Here are three runs using argon-openmm.py for each.

`Well_s_q_omega_Ar_data.xml`

```
Step FoM CMA itera... epsilon (#2) sigma (#3)
0 0.6321 1 1.08 3.4
0 0.9843 1 1.08 3.4
0 0.6408 1 1.08 3.4
```

`Well_s_q_omega_Ar_data_no_data.xml`
```
Step FoM CMA itera... epsilon (#2) sigma (#3)
0 67.87 1 1.08 3.4
0 67.76 1 1.08 3.4
0 67.39 1 1.08 3.4
```

**Describe the expected result**
"no data" data points should not have any effect on the FoM calculation.