"From what I understand it is the breeding back to the parents that can produce the polyploid forms and many of the odd and unusual traits."
I don't think that is true of zinnias. As far as I know, there are only two commercial tetraploid strains of Z. violacea, State Fair and Burpee Tetraploids. If you cross a tetraploid with a diploid, you get a triploid. As a rule, triploids are sterile, in that they can't set viable seed. There are a few annual flowers that are deliberate triploids, because their inability to set seeds means you don't need to deadhead them. And you can't save seed from them, either. For example, I think there are a few triploid marigolds. Seedless watermelons are triploids. I don't know of any triploid zinnias. I seriously considered producing triploid zinnias, but I reconsidered because I wouldn't be able to save seeds from them.
I also considered producing additional tetraploid zinnias. But, referring to the book, Flower Breeding and Genetics, http://www.amazon.com/Flower-Breeding-Genetics-Challenges-Op... in Chapter 12 titled ZINNIA by Dennis Stimart and Thomas Boyle, it says "Relative to diploids, tetraploid zinnias have larger flowers and thicker, stronger stems but also have poorer seed germination, less branching, delayed flowering, and fewer capitula [flowers]. I considered that the disadvantages of tetraploid zinnias outweighed the advantages for me, so I gave up the idea of breeding tetraploid zinnias. That's not to say that others wouldn't enjoy breeding tetraploid or triploid zinnias. Just not me.
On a somewhat different note, I have enjoyed crossing Scabiosa Flowered Zinnias with "regular" zinnias. The disk florets of scabiosa flowered zinnias are not the usual yellow, furry stars, but are like little florets with the same color as the petals. The hybrids of scabiosa flowered zinnias with conventional flowered zinnias can produce some interesting results. Like, for example, this specimen.