I am currently a math and statistics undergraduate who is applying to graduate programs in statistics this fall. Two of my professors have confirmed they will write letters of recommendations for me. They are both well respected probabilists in the math department. However I have heard it strongly suggested to get at least one statistics faculty member to write a letter for me when applying to statistics, but I don't really know anyone noteworthy in my school's statistics faculty. I've been trying to get to know my professor who is teaching my theory of statistics course this term, but I still don't think he knows we well enough to write a strong letter for me.
However there is a third math professor who I believe would be willing to write me a letter. I've been in multiple of his classes (always coincidentally got relatively higher grades in his classes), met him at multiple conferences, we are on a first name basis etc. His work is mostly in dynamical systems and ergodic theory, the latter being a something of a sub-field in probability. The fact that he's an ergodic theorist is perhaps relevant since I've been told that strong skills in probability/analysis are highly valued in statistics programs now a days, and hence probabilists' recommendations are taken seriously. He's also been the math department head a couple times and has chaired the local mathematical sciences institute I think more than once.
Therefore I'm thinking of just getting my third recommendation from him, since we's a respected academic who could write me a strong letter. My question is how important do you think it is to get at least one reference from a statistician given that I'm applying to statistics? Do you think I should pick a more generic letter from my statistics prof or a stronger letter from a third math prof for my last letter?