Usually for one of two classes of reasons:
1) It's a worthless argument
2) It's an argument that takes into account the nuances of the nature of probability and the fact that Kolmogorov's measure-theoretic axiomatic formulation of probability is useful in many, many applications but is fundamentally lacking here, and that Bayesian interpretations, subjective probability, or even frequentism (adequately elucidated) is required here over and against any calculations.
Instead, we require fishing licenses.
What I see.....
The poster believes the claims of creationist 'experts' that evolution is highly improbable to the point of being impossible.
Trust in them means one need only state that it's impossible due to improbability.
Thus, instead of presenting an actual analysis, we see the mere pronouncement of sciencey/mathie sounding terminology...."improbable".
If we do get around to specifics, their premises are restricted to those which serve their agenda.
Examples:
Say a biochemical process' evolutionary origins are examined....the poster will say there is
only a single possible evolutionary path from one to another.
This ignores as yet unknown other possible paths. Other errors are to assume an unrealistically
short geological time frame, or to fail to take into account a vast number of organisms involved.
Remember the old arguments that the eye is irreducibly complex? It turns out that pathways
from blindness to vision have more intermediate steps & are more multi-faceted than they ever knew.