<p>I’m applying to grad school next fall, and am thinking about asking a certain professor to be a recommender. I’ve only taken one class with him so far and have had no research projects with him, but I’ll be taking another seminar style class with him next semester, and he knows me pretty well. Out of all my options, he seems like the best choice for the third recommendation. </p>
<p>The only problem is that this professor only seems to interact with the best math majors at my school, which is already a top-10 school in math. For example, I think at least a quarter of the students who took the class I had with him were honorable mentions on the Putnam, and at least 90% qualified for the USAMO. I know math contests aren’t that important, but that should demonstrate what types of students he normally deals with. Compared to them, I’m not very good, and I did below average in his class, though I still got an A-.</p>
<p>How much would this hurt? If it’s relevant, I’ll be applying to theory cs schools, but this class is closely related to cs.</p>
<p>I would guess very important, but graduate admissions committees can usually interpret that in context. </p>
<p>I have heard both from my own references and from professors reading graduate applications at other schools that explicit comparisons are the most useful information they get from a letter of recommendation. So being one of the stronger students your professor has taught will help. </p>
<p>On the other hand, adcoms do realize that your university might have many strong students. Out of 16 math PhD students in my year at Stanford, 3 are from Harvard. I assume that even more Harvard students got accepted. So not being the best math major at Harvard would not keep one out of the top programs, because people know that there are a lot of strong math majors at Harvard.</p>
<p>If you were noticeably weaker than a fair number of very highly accomplished students, but still very strong compared to the vast majority of college students, your professor will hopefully communicate that appropriately in his letter. </p>
<p>My experience has been in math but I assume that same principles hold for computer science.</p>