For what it's worth, in the data in my database, I've found that having two days off is one point better than having one day off, but that 3+ days off is slightly worse than one day off. So based on my data I subtract two points for a team with one day of rest vs. a team with 0 days rest, and three points for a team with two days of rest. I take it the OP's data was for some reason different?
The problem with all these studies is that they all come to slightly different conclusions. That's makes it close to impossible to build any of this into a model with a high degree of confidence.