How Do Time Series Patterns Become Predictable?

We're given the diversity of time series data, that's kind of a high order o achievement. Do you have any insight, or could you ga share some of what those features are? I mean, they tend to actually capture the types of categories th a lot of people would think of when they're thinking of what types of time series analysis methods have been developed in the past. And so i mentioned before, there's things like the distribution of values, or does a time series have a bunch of outlies, or is it approximately normally distributed, or does it have a particular skew?