record and SOS to develop their rankings. That's a very good way to rank how good a teams' resume is, but not as good when it comes to predicting future outcomes.
There's other rankings that are based solely off of efficiencies and MOV types of things and don't really factor in W/L record at all (i.e. KenPom for college basketball). Those systems are usually better at evaluating the actual quality of a team and are better predictors of future success. However, they don't do as good of a job identifying who might have the "best resumes" or be the "most deserving" teams.
There's other rating systems that try to do some sort of hybrid of the 2. NET in college basketball tries to do that sort of thing.