I expect that if the jackpot continues to roll, my model will be less predictive in terms of the actual jackpot. In fact, hidden in these tables is the fact that I am not actually predicting jackpots so much as sales.
The higher the jackpot goes, of course, the fewer instances there are of previous sales and the weaker the model is at approximating it. There are many factors at different times, jackpot fatigue, competing jackpots, economic conditions, even weather.
Graphically the exponential model I use breaks down as sales grow higher. In the past I've used other cruder models that are based on the last 5 drawings rather than the entire set of drawings.
My general impression is that while the pure exponential model of MM has worked in the past, more recently the behavior of the MM system has become more like PB, which is grows linearly in the early draws and exponentially in the later draws.
I think another rollover will almost certainly give a jackpot higher than $250 million, and maybe higher than $270M. It's hard to know, of course. It is certainly the case that the probability of another rollover is favorable, but of course, it is less likely any previous drawing. Two rollovers will certainly bring the jackpot to the mid/low 300's.
I expect that we will someday see a jackpot of over $400M, but even now with the jackpot over $200M, this still is unlikely for this run. It's always going to be a long shot, but it seems likely that it will happen simply because there are so many repeats of the situation.
It is interesting that now the winner of such a jackpot would be able to hide himself or herself relatively easily. Such jackpots don't excite as much as they used to do. This of course, leads to less drawings where the field is nearly saturated.