On Vegas and on other Casino-lands we see plenty of flashy slots machines that are 100% digital - no mechanical parts, only a screen and a few buttons. Isn’t it really easy to rig those machines to make the house win… a lot? I mean, it could also be rigged with the “real” machines, but this would be in theory easier to spot… right?
The ROMs in slot machines are audited and regularly inspected by the local gaming authority. In the case of Vegas that’s the Nevada Gaming Commission, which licenses and inspects casinos and game manufacturers.
Of course, they’re already rigged for the house to win. That’s the point.
As friedo said.
I’m sure they could be easily rigged. But what would the casino gain? Slot machines are casinos biggest winners as it is now. I think they make the house something like 16% of the total betting over time. If the chumps, oops I mean players, didn’t make a little from time to time they would quit playing.
Casino games are rigged only in the sense that the payback ratios favor the house. That is, to keep it simple suppose a $1 slot has 1000 combinations and 1 of those is a winning combination. It really will win 1 time out of 1000 (on average). But if you play 1000 times and hit a jackpot once you only win $900. The machine is perfectly fair mechanically but the odds are are stacked in favor of the house.
The house always wins.
I once browsed through a mechanical slot machine brochure and it noted that a machine could be set to pay off anywhere between every time and never. I assume that’s how they handle those million-dollar payoffs and can claim they have the loosest slots in town, though that may refer to the women who aren’t formally on the payroll.
A lot of the “mechanical” slot machines have been digital for quite some time now. They may have spinning mechanical parts, but the result has already been decided by the machine as soon as you pull the lever. All the mechanical stuff that happens after that is just to make you feel better.
Yup - and even the ‘skill shot’ type sub-games (where you have a light flashing between ‘double’ and ‘lose’, for example) are decided before you hit the button.
These machines are one place where the Gambler’s Fallacy is sometimes not fallacious. Electronic slot machines (aka ‘Fruit Machines’ here in the UK) may be programmed to pay out a very specific percentage of the money put in by punters - and they do this by switching between modes which pay out more, and other modes which pay out less.
Therefore, it really can be true that a machine that has been played for an extended period, without significantly paying out, is more likely to yield a jackpot.
And likewise, it can be true that a machine that has recently paid out well, is not worth playing (in addition to the payout percentage adjustment, it may be that the payout coin tubes in the machine are empty - in this case, it’s not going to let you win until they fill back up again)
And beyond that, even if the odds were perfectly even, the house would still win. It’s basic human psychology. Most people don’t play until they win - they play until they run out of money. Even if they hit the jackpot, they’ll just blow it all on the next machine or table.
The specs for electronic slot machines are filed, IIRC, with the Nevada Gaming Commission who in turn employed Ron Harris to design a tester, and then did random spot-checks. The tester, I think, basically just confirmed that the EPROM program matched the one on file with the Commission.
Unbeknownst to the Gambling Commission, Harris added another feature to his tester: it replaced the EPROM program (of video poker machines?) with one configured to respond to a particular sequence of bets and then shift modes. Later he developed a hands-off way to defeat Keno by reversing its PRNG. I saw an interesting TV documentary on this (but can’t find it with a quick YouTube search).
How does this inspection work? How can someone look at a chip and figure out whether it is rigged or not?
As a programmer myself, it seems to me that the only way to guarantee fairness is by having inspectors review the source code, and also have some method to guarantee that the compiled version of those approved programs ended up on the chip. That seems so cumbersome as to be totally impractical, but how else could it work?
My guess is that there’s really no inspection of the software at all, only a chain of sworn affadavits (“Our programs don’t cheat!”) such that if a gaming machine is ever seriously suspected of cheating, only then will anyone begin to take a good look at the source code.
In other words, there’s no real inspection of individual games, but the inspectors could take one game, play it a million times, and certify that there’s a fair number of wins. But even that won’t work if it can switch modes like Mangetout described.
When I say “switch modes”, I mean that, if the legislation is such that machines must pay out (say) 78% of money paid in, the machine may have two different schemes for (pseudo)randomising the results - one that yields lower than that percentage, one that pays out higher.
Thus, the machine can use one or the other depending on whether it is below or above the legislated payout percentage. It’s easier to cope with changes in the expected payout level this way - as you just change the setting (just like a thermostat), rather than having to actually rewrite code to try to hit a specific average.
That is exactly how they do it. They use ROMs which are designed to be easily accessible and dumped, and they compare the image to the compiled code. They also audit the source code itself to make sure it’s fair and behaves according to regulations. (ETA: That’s also why slot machine hardware in Vegas looks so dated. They’re using the same basic embedded systems and commission-approved hardware that they’ve been using for years and years. You’re not going to see slot machines with Playstation-level graphics any time soon, because the industry isn’t going to fix what ain’t broke.)
The above-mentioned Ronald Harris took advantage of his access to swap out good ROMs for ones that would pay out in response to certain inputs that he knew. Later he exploited a flaw in a Keno game’s PRNG. Then he spent some time in prison.
Yes, I have thought that. Roulette has the fairest odds where there are 37* numbers including zero but a single chip on a winning number will return you 36 chips. Yet if you watch people play roulette they lose more money faster than most other games because instead of sticking to bets that exploit the naturally low house margin, like a single number bet, or a corner or row or line, they instead place multiple bets that overlap each other.
Quite often the student of roulette psychology can witness a person placing so many chips all over the table that even if they got the largest payout possible from the chips they placed they would still have put more money down that they got back.
I am certain that if roulette specifically had no house margin it would hardly matter to the house.
*there are tables with double zero though.
Even single-zero roulette is an expensive game. Your expected-value on almost all bets is -2.7%.
Single-zero roulette tables are like Bigfoot. They talk about them on TV and, hell, we even have pictures of them. They’re just sliiiightly different than what we see in everyday life, enough to be interesting. But let’s be honest- no one’s ever really seen one and most likely they don’t really exist.
I seen one at the Bellagio! I took a picture of it but it came out all blurry and looks like a guy in the woods wearing a roulette suit.
Maybe not in the Good Ole US of A, but in most of Europe, the casinos do perfectly well with one zero.
I have had lengthy arguments with people about this. The point is that even if it paid out statistically even odds, the house would still win because the house has deep enough pockets to outlast you, and eventually drives everyone out of the game. The punters go busted, but the house never does.
Say, a bus load of players come in with $100 each. In the first couple of hours, five people lose their whole hundred. Now the house has a pad of $500, which can be used as payout, with no risk of the original losers getting it back. Somebody wins a $500 jackpot, and the house is still even. By the end of the day, the house will have busted many players, but nobody has busted the house.
If you have deep enough pockets, you can play double or nothing forever – you will eventually win.
The house doesn’t have deeper pockets than “everyone”. Not if you add together the pockets of everybody walking into the casino to play over the course of a day/month/year.
Let’s take a small example. One slot machine with even odds. Joe sits down with $100 and plays. If he busts, the house didn’t “win” because Jeff is going to sit down with another $100 and play. If Jeff busts, or gets bored, then Jimmy is going to sit down with another $100 and play.
Statistically, if the odds are even, over the course of a day, even if there are 50 different people sitting down to play, it will pay out as much as it took in. That’s what statistically even means. If one person loses money another person will gain, and not every gambler is a completely degenerate gambling addict, people DO walk away when they’re up, to do things like eat, poop, sleep, go home, or just say “I’m gonna buy a TV with this $500 I won”.
There is a well known statement about gambling. No betting system can turn negative expected value bets into a positive EV total. A corollary for you would be that no betting system (like gamblers busting) can turn even EV bets into negative EV totals. Casinos win because every single bet they offer is negative EV