That's a valid point, but affects only one in 216.
I just checked the percentile distribution:
Damage mod,: New% vs. Old RQ3%
-1d4: 9,3% vs. 0,1%
-1: 16,7% vs. 0
0: 36.6% vs. 60,0%
+1: 21,3% vs. 0
+1d4: 11,6% vs. 39,2%
+1D6: 4,2% vs. 0,7%
+1D8: 0,5% vs. 0
Sort of surprised me to see how low the chance of having a +1d6 modifier with a natural roll the old rule actually gave.