The model-rival method of training relies on the fact that dogs learn by observation. By providing a model of good behavior or a rival to compete for resources, dogs can learn to mimic desired behaviors. So a trainer might have another human act as the model, praising them for completing tasks on command or scolding them for unwanted behavior. The dog, as an observer, learns what to do correctly from the model.
The model can also act as a rival, competing to do the right task for a desired toy or treat as a reward, encouraging the dog to pick up on he task and accomplish it more quickly. Mirror training relies on the same principle using the dog owner as a model, then offering rewards for mimicking good behavior. It’s meant to use the dogs natural instincts to operate socially instead of working against them.
To put it simply, the dog learns by example. This training method has been found to operate with a similar level of success as positive reinforcement and operant conditioning, though some trainers may find it more natural and preferable. If your dog has a strong bond with you and can spend a lot of time observing you and following you around, this may be a technique that you find more comfortable than sticking to regular training sessions.