The effect is poor when using Lam-Bias command to do variance reduction?

The Lam-Bias card was used to decrease the inelastic interaction length for variance reduction,
but the effect is poor.
here is the .inp file:
LamBias.inp (1.5 KB)

and the result is:

Compared to Mcnp, Phits or Geant4, the effect of variance reduction is too poor.

The effect is not poor, is null, since LAM-BIAS does not apply to low energy neutrons.
We are considering to possibly include them in the reaction probability biasing by LAM-BIAS, in connection with their recent pointwise treatment.
In the meantime, you can just increase the number of histories accordingly.

@ Francesco Cerutticeruttif, thanks!