There's no need to introduce parallel universes for this question. Although there's also no problem with thinking that way if it helps you -- sometimes when we talk about the results of an experiment probing a random variable like a dice roll, we imagine a hypothetical "ensemble" of experiments, which is similar to a set of parallel universes each of which has a different realization of the experiment.
Anyway...
You are right that rare outcomes will happen. If you roll a dice enough times, you will find a string of 20 dice rolls each of which is 6. Therefore if you roll a dice and find a string of 20 6's, you cannot conclude with absolute certainty that the dice are rigged. This is not special to your dice experiment; it is a fundamental limitation of observational science that we always have some uncertainty in our observations.
However, we can quantify the uncertainty that we have. If you do roll 20 6's in a row, you can ask how likely you were to observe this outcome. The probability of rolling a 6 with 2 dice is $5/36\approx 0.14$, and therefore the probability of rolling 20 6's in a row is $(5/36)^{20}=7.1 \times 10^{-18}$. Needless to say, this is quite unlikely.
Now you have to be careful how you interpret this number. It means that, if the dice are not rigged, you have a extremely small chance of observing 20 6's. It does not mean that the dice are definitely rigged; it does not mean the probability that the dice are rigged is $1-7.1 \times 10^{-18}$; it does not mean that the probability of there being parallel universes is large or small. However, it does mean it would be a bad idea to place any bets that are based on the assumption that the dice are fair; I wouldn't bet that the next dice roll would be a 7.