I wasn't sure on what Stack Exchange forum to ask this. I'm trying to computationally model a situation where I have an object that has two possible states and can switch between them. I know that experimentally the average stretch of time that the object spends in State 1 is 60 time points, averaged over many of these continuous stretches of time. I also know similarly that experimentally the average stretch of time that the object spends in State 2 is 6 time points. If in my model I'm making a choice at every time point whether to switch states, is there a way to analytically determine or estimate from these average lengths of time what the probabilities would be of switching? I would think that the probability of switching from State 2 back to State 1 is much higher than the probability of switching from State 1 to State 2, and that there would be high probabilities of not switching, but I'm not sure how to analytically determine what those probabilities would be.
Thanks so much.