I've always assumed that there's a "real dollar" influence to the holds and variance that the casinos set vis-a-vis the players for those machines. If you manage to hold $50 off a penny slot in 30 minutes you might have just blown grandma's budget for the whole trip, while that real dollar loss for a $5 player is less significant. I know that some people will reply that "a percentage is a percentage" but slot technology has come a long way, there has to be a time element to the range of std deviation that they are incorporating.
If you're playing penny slots at $2-4 per pull and $1 slots at $5 per pull, while the hold figures will still be as shown in the document, the short term variance could very easily show not much difference, or a reverse experience.
Also, don't forget that a higher denomination is not always a lower hold; these are general figures. I routinely go to casinos and find $1 VP with "full pay" (9/6) Jacks or Better, and when I check out the high stakes gaming area, the $5 (and up) VP may have 8/5 Jacks or Better. I've never understood the casino rationale for this, except perhaps (1) reduced risk of loss to high stakes players and (2) because they can get away with it, apparently, or don't really expect to get any meaningful action on such machines.
--BG
===================
2a. General Odds
We were in the local MGM casino in Detroit a few weeks ago and saw this Casino House Advantage & Expected Loss sheet. Nothing that the savvy player here don't know, but it was a good reminder to us not to stray too far in the direction of the more fun games because the casino hold really goes up! I went online to find it so I wouldn't have to type it all in:
http://www.americangaming.org/sites/default/files/uploads/docs/odds_brochure_2012.pdfI'm not sure I believe that the Penny slots are more than double the hold of dollars, as when I stray from VP, my money seems to last the least at $5 slots, then $1 slots, but I can play a longer time (even at $2-$4 per pull penny slots) on penny's. But that may be short term variance.