Module Mismatch Losses Detailed Study gives wide range of different results

[linkeshd]

In PVsyst 7.0.8:

Using the graphical tool for studying the Module Mismatch, I know there are some random seeds, but the two tools ("Show histogram" and "Show graph") give widely different results.

Using a system comprised of JKM 315PP-72 modules, 19 units in series, 20 chains in parallel, and Normal (Gaussian) distribution, I get the following two behaviors:

Using the Show histogram Tool:

Pmpp loss average =2.4 %. Clicking "More statistics" increases the random sample pool, but tends to converge around a value similar to initial value (in my case, it jumped to 2.8% then converged down to 2.4% as I increased the sample size).

This is reasonable and expected behavior of a randomized normal calculation.

Using the Show graph tool:

Each calculation gives wildly different results, ranging from 1.5% to 4.5%, and jumping widely with each recalculation.

This is counter-intuitive and dissimilar to what I experienced in PVsyst 6.8.8 wherein the value shown in the graph was relatively constant.

Please check?

Thank you!