While we can’t take the rmm forecast amplitude literally, I still think rmm forecasts have some utility where we look at multiple models/ens and look at what phases they think the mjo will be in and how quickly or slowly they cycle through. The spread gives a good idea of the uncertainty of the amplitude, which can be really noisy. Even in noisy amplitude predictions, we can get a good idea of where the mjo is going to be, phase wise, in a week or two ahead of time.