Funkified
Talking of funky updates, I have updated a few casinos with this months releases, and I noticed again the trend for each game, old and new, updating again. What was different this time is that in many cases this was NOT just downloading and verifying a checksum, I noticed that the old games really were fully updating, noted by the continual packet exchange through my modem. When just checking, there is a brief flicker while the checksum downloads, followed by a pause while it is checked - this time, the checks triggered the full download of a number of new games.
To my earlier point about the streak envelope. An example might be the 5 cards for the starting hand at any Video Poker. If only one ramdom number is used, it could be equated to, say, 52% just a high card, 48% a pair or better. The result would determine the class of hand drawn, and an algorithm would just make up that structure from appropriate cards for display. Simplifying the draw would probably not be an option, as the mathematics would only behave if the discards were replaced by independently drawn cards. This would give interesting behaviour, while you would forever get bad starting hands over good, the draw will occasionally come good by pure chance, giving the isolated big hit which is quickly swallowed up over the next few hands.
Another possibility is that the first card is selected ramdomly, but the other four are selected through a weighting envelope such that the cards will either favour working with, or against, the previously drawn cards - streak enhancement. The envelope will be designed to replicate a fair proportion of weightings such that the overall return on the game is exactly as it should be. This would benefit the business, as it would provide for winning streaks that players will remember, and they will have many of these streaks to remember! To compensate, there will be long losing streaks to balance the return.
This would also work on slots, and would explain certain "patterns" that seem to repeat, such as on 5 reel drive, there is a very marked run of some 30 or so spins without a single win on any of the 9 lines, but a scatter win or two hits around the middle of the run. Much of the rest of the time, the slot hits something small on nearly every other spin, and occasionally something big.
On the Thunderstruck slots, it seems that you either hit lots of scatters, or lots of wins, mostly small. This leads to many bonus rounds being void of any wins except for a couple of hits on two scatters. Very occasionally, a bonus round hits at the changeover, and leads to many hits and a big total - just seems everything comes together. There is also the possibility of a big total from just one very big hit in one of the free spins, which will pretty much be the entire contribution. Bonus rounds starting from 4 scatters seem poorer yielding than those started by three.
There is also the question of why the previous hand is saved, and why a bad run thus continues even the next day when you log in and play on from that last hand the previous day. This last hand COULD be used to continue the streak weighting, although there is no proof it has such a use. The last hand is stored on the server, and is not just localy stored as the last state - indeed the server subjects play to a delay for "refreshing state" just so that it can recover and display the last spent hand, yet, MG claim they took a short cut with the doubling game because they said players did not welcome a delay while a second transaction resolved the players choice of card - why is one unnecessary delay OK, but another needs a short cut work around?