I was looking for visual representations of randomness for a cover some time ago, and what I found was mostly works of art. I ended up using a generative designed drawing that used randomness. The point is that randomness is an abstract concept that means the lack of predictability in many different fields and as such it will avoid any universal representation. You may think about randomness in math, philosophy, physics, biology, astronomy, genetics, religion (!), design, generative architecture, art, etc. Ironically, to be a good metaphor, an icon of randomness should change all the time, paradoxically, and be different depending on the field you want to apply it.
I discovered (or at least this is what I like to believe, in my scientific ignorance) that the two crossing arrows may refer to a concept in physics: the concept of "arrow of time", based on the idea that time is "asymmetric" (or one way direction). This concept involves another one, the "entropy", strictly related to randomness. I am a designer, not a mathematician, and I cannot explain this well, but those two arrows crossing should refer to a simplified representation of this lack of predictability which involves time. I guess a Nobel price would tell this a bit better (or more probably laugh about this) or give you some more visual examples: as an example, think about images from "chaotic systems"... Like the famous double rod pendulum:
Anyway. If you are looking for conventional representations of randomness, you will not find any other visual icon: the one you showed in your example is arbitrary and it may look "awkward" from a certain point of view because it is just a convention and does not stick to any particular field. The good thing about that icon is exactly what you may see as its limit: it became a convention and that's it.
Consider also that an icon aims to be a stylization of a concept or its metaphor, specially when representing something totally abstract or conceptually complex.
You are free to create another arbitrary representation though (that will maybe become a new standard one day ;). A good approach, in my opinion, would be to restrict your research to the field you are interested in, by trying to create a closer metaphor of the thing you have in mind or your idea of randomness.