So the base game could be the exact same in free play and real play, but the very different mathematics could kick in only in real play, thus allowing a line of wilds. For all we know it could act similar to a progressive jackpot.
That's an interesting thought..... DoA is capable of insane hits, a couple of high-rollers hitting one of those 4000-5000x stake monsters at one casino in an evening could cause some serious carnage.....
So yeah, how about everyone playing for real feeds into a hidden 'progressive' pot, and when there's enough cash in there, randomly release the 'five wilds winline' to a player, if it's a high-roller, then the pot will take a lot longer to fill up again, to protect casinos from business-disrupting payouts.
After all, DoA is unique in NetEnt's slot library, they do have a couple of other high variance efforts in there (JATB/Reel Steal) but nothing that seems to hit as big and as regularly as DoA, and they certainly haven't produced anything like it for years.
Tinfoil hat stuff for sure..... But then again when you've chucked as many spins at this slot as me and dunover have in free play mode, you do start to ask if there's a difference between real and free, especially if you compare that to rolastan and Blathon's history in real. (Just go back to July 2013 and look through the 1000x screenshots thread.)
Still doesn't explain why they couldn't just model that behaviour into free play though, make it a 1 in 200 chance on the bonus round or something, unless it's an oversight, and the code to 'release' the five wilds simply doesn't exist on the freeplay calls to the server.
Fucking hell, time for my own tinfoil hat, unless dunover wants to lend me his