Switchers and low monthly mileage drag locomotives cannot post the economic numbers that mainline power can. As a matter of fact, many railroads considered switchers as a user of income and not a generator of income, although they are obviously a necessity. Therefore, a smaller or older engine such as a 4-4-0, a 2-8-0 or a Mike probably can't make the cut. So that leaves the larger mainline power, and the ranking should reward high monthly mileages and the ability to achieve GTM/train hour, a measure of horsepower and productivity.
The engine should be trouble free, and broadly applied by more than one railroad, i.e. a standardized design if you will. An engine type that has been compared economically with a diesel of the period and tied or won should also be on the list. So that list is pretty small....
-a Lima Berkshire used by NKP, C&O, P-M, etc was a standardized design, in spite of differences in weight and minor construction details. The NKP Berk also was tested vs. diesels and about matched them economically.
-a NYC Niagara, while it was a design unique to NYC, economically outperformed a 3 unit E7 set on the RR, and set world records for both monthly mileage and availability, and also overhaul interval, so that has to be on the list.
-The N&W three engine test also put up good numbers, although three different engine types were necessary due to the N&W terrain and the fact that the RR narrowly optimized each design for specific assignments.
-you might also make a case for the use by three different western railroads of 4-6-6-4 Challenger type locomotives, although their specs differ by quite a bit. I am not aware of a comparison test of any Challenger type with a diesel, however.
-the "non standardization" of "mountain power" in the east is a footnote to this question. B&O used 2-8-8-4's, PRR used 2-10-4's, Clinchfield used 4-6-6-4's, D&H used 4-6-6-4's of a different design, C&O used 2-6-6-6's, and N&W used Y's. I think that is why each RR has its own fans.....