cunning indeed, significantly faster.

Benchmark: running for_loop, map_void, new_guy, nonvoid_1, nonvoid_2 f
+or at least 3 CPU seconds...
  for_loop:  4 wallclock secs ( 3.19 usr +  0.00 sys =  3.19 CPU) @ 15
+55.17/s (n=4961)
  map_void:  3 wallclock secs ( 3.29 usr +  0.00 sys =  3.29 CPU) @ 11
+30.40/s (n=3719)
   new_guy:  3 wallclock secs ( 3.07 usr +  0.00 sys =  3.07 CPU) @ 19
+67.10/s (n=6039)
 nonvoid_1:  3 wallclock secs ( 3.15 usr +  0.00 sys =  3.15 CPU) @ 33
+9.37/s (n=1069)
 nonvoid_2:  3 wallclock secs ( 3.15 usr +  0.01 sys =  3.16 CPU) @ 13
+9.87/s (n=442)
[download]

Cheers,
R.