I recommend you to wait for my Byteball test report. I have already spotted things which need to be improved, once they are fixed the price will shoot up. Just accumulate more GBs now.
First, let's discuss the main problems with Iota vs Byteball:
1) Anti-spam PoW (unprofitable PoW) for emails was entirely designed for one-offs. You put in some labor and the state changes from 0 to 1 once it's been satisfied and this information is now meaningless and can be thrown away afterwards. Satoshi decided that in order to chain the past to the future in a continous ledger, he would need to create a linked list with a get rich quick scheme built on top of it, otherwise there are no incentives for the burden associated with the cost of holding this data forever.
It seems like IOTA was invented (unprofitable PoW) while ignoring everything Satoshi learned needed to be utilized in order to launch the bitcoin scheme. True, Iota could utilize a form of pruning, thus correcting the incentives problem and dragging it back to more like email PoW in nature, but then you have a very fragile, unsound money system with no valid form of state recovery if it goes down.
I'm not saying bitcoin is sound money, because it's not. It doesn't function as a store of value due to price floor being recursive based on it's own demand, meaning you can always mine an endless stream of coins as transaction fees at the new floor and the floor can crater to nothingness at any time unless bitcoin was the unit of account of something (but it never will be since it doesn't function as a store of value - chicken and egg scenario).
Most attempts to increase bitcoin scalability in altcoins, and thus lower redundancy, seem to make it even less sound money where you can no longer trick people into believing it's money at all. People will see how fragile they are and treat them as they should be, just as a throwaway currency like airline miles instead of pretending they are a valid store of value like gold or silver.
As for Byteball:
2) Is there really a purpose in a, hmm how to describe it in general, a branching/multi-threaded/non-linear system like a DAG if you're just using bitshares-style consensus to force convergence and make it linear in a single path again?