In this blogpost I answer the questions and comments that came up during the preliminary request on May 2nd. 
First let me show the figure I requested preliminary for, as a baseline for the cross-checks shown later.

Preliminary request for an example fit, shown in the presentation on May 2nd:

If we fix alpha to 2, then the fit quality decreases strongly, and, as it is known, R and lambda decrease.
The chi2 value increases in this example by a factor of 4.2, and the confidence level drops below 1e-5.
On the (data-fit)/error subplot, an oscillating structure appears, showing the deviation of the Gaussian shape from the data.
It is interesting to note that the combination R/(lambda*(1+alpha)) [PRC97(2018)064911] stays approximately the same (2.44 vs 2.38, with an uncertainty of 0.07, so closer than 1 sigma).

2) Data fit including more small Q points

It was investigated what happens if the first correlation function data points are also included.

In this case two checks were made. In one check the small Q points were added, except the first one.
This is also included in the total systematic uncertainties of the fit parameters shown in the preliminary request.
The results are very close to the original ones (shown at top if this blogpost), alpha changes by 1.6 sigma:

 
In a second check, this first point was also included, but now the systematical uncertainties of the datapoints were added in quadrature.
This is not optimal, as these are correlated, hence a conceptual prerequisite of the chisquare minimalization is not fulfilled.
This results again in similar parameters, alpha is different from the default (shown at top if this blogpost) by 1.1 sigma.

3) UrQMD fit with free lambda

It was also investigated what happens if lambda is released in the fit to UrQMD simulated correlation functions.
The resulting lambda is compatible with 1, for both cases (with or without Coulomb):