Landau-gaussian convoluted fit looks weird

ROOT
2

It looks like the “GSigma” can get too big. Try to define this variable as “limited” and set the upper limit to something like 10.