assuming scrptt hashrates stay constant thats six months roi not terrible
Which would be an insane assumption, Scypt hashrates haven't been even CLOSE to flat for a very long time now.
They haven't been as bad as that big jump in December 2017, but the average this year has been about 2% per difficulty period.
With cheaper miners I would think it goes higher or at least, it would go higher than if the miners were expensive.
The other bigger risk is that this is on 28nm. If a 16nm appears, it might be 3 times better, as how s9 is 3 times better than s7
Innos A4+ is 15-20% more efficient but is on 14nm node. So, if bitmain comes up with a 14nm l3 in future, it will be alot better i think. I really hope one does not appear soon though as it would be bad for L3+ owners.... but a 14nm L3 would possibly have no possible competitor that comes close to matching it.
Just my 2 cents