Can't we discuss that on the board? I think it'll be more appropriate.
But in short words (maybe...; actually I'm not sure what you don't understand):
1. you cannot sum on things that are given. P(ro) above as nothing to do with e.g. P(rob), where b is given
2. don't confuse probabilities in general and parameters of the model. P(ro) above is **NOT** a parameter; P(rob) would be, for a 3-gram model
But in short words (maybe...; actually I'm not sure what you don't understand):
1. you cannot sum on things that are given. P(ro) above as nothing to do with e.g. P(rob), where b is given
2. don't confuse probabilities in general and parameters of the model. P(ro) above is **NOT** a parameter; P(rob) would be, for a 3-gram model
