Downloadable slot systems

I have heard of downloadable slots where the casino can remotely change the payouts and the type of game to be played. Pioneered by IGT. Is this the remote "switch" we have been suspecting all along?

Yes and no.

All of the slot manufacturers have downloadable games systems and, if I'm not mistaken, a similar system has been in use in Europe for many years. I don't know that it's correct to say that IGT pioneered downloadable games, but their system certainly is getting all the press right now.

In a downloadable games system, slot directors load new game programs on a central server and then deploy the new game to one or more slot machines with a few clicks of a mouse. No more having to go to each machine and having to change one or more chips in the machine. The new game program could be an entirely new game or a different payback program for an existing game.

So, yes, this technology does allow slot directors to make changes on a machine without the telltale sign of a technician at the machine. But no, this technology does not mean they are going to make or will be able to make changes at the drop of a hat.

First, the casino has to license all of the game programs it has available on its central server. If the casino wanted to have loose programs for weekdays and tight programs for weekends and holidays, it has to license all of those programs. It still may not make economic sense to license the additional programs because the time to break-even is too long.

In addition, regulatory safeguards are in place to prevent changes at a machine while it is being played. In Nevada, for example, a machine has to be idle for four minutes before it can accept a configuration change. While the change is being made, which may take a minute or many minutes, a message indicating the machine's configuration is being changed is displayed on the machine. Then the machine must remain idle for four minutes after the change has been made. The program in the machine cannot be changed while someone is playing it.

Despite this fact, I know it will be impossible for me and other writers to convince players that the cold streak that just followed their hot streak was due to randomness and not due to a configuration change that was just downloaded to their machines.

Best of luck in and out of the casinos,
John

John,

Your discussion of "cycles" in slots makes pretty good sense. But isn't it leaving out one important factor?

Sure, each spin is like the heads/tails coin flip, the same odds every time. But a coin doesn't have a chip in it programmed to pay back 91% over a set period of time. Doesn't it stand to reason that if the slot machine pays out a lot in a short period of time, it's going to have to hold back a while to maintain that 91%? In short, go into a "dead cycle"?

Yeah, I know; we don't know the length of time that 91% is calculated over. For all we know it might be 100 years. But given the casino owners' profit demands, isn't it more likely to be a much shorter time — short enough to create such "cycles"? But you're certainly right that the human mind demands and thus creates patterns — and my discomfort at throwing money away at play will go right on producing rationalizations allowing me to do it.

Thanks for all your commonsense cautionaries; they've saved me money.

Jon

Dear Jon,

The coin may not have a chip in it, but it does follow the rule that the number of heads (or tails) divided by the total number of flips will be very close to 1/2 after a certain number of flips. Random Sampling with Replacement tells us so. And we can calculate how close we expect the ratio to be with a given level of certainty after a number of flips. The ratio may not be exactly 1/2 — after all, the results are random.

Random Sampling with Replacement governs slot machine math too. There is no governor function to force a generous slot machine to go into a "dead cycle." Such a function would be illegal in the United States. Your machine pays back 91% in the long term because it uses the RNG to select a combination at random from the pool of all combination possible on the machine — and if you found the total of how much all the combination are worth and divided by the total number of combinations, the ratio would be 91%.

We do know the length of time that 91% is calculated over. Time on a slot machine is measured not in ticks of a clock but in spins. We can calculate how close we expect a machine's actual payback percentage to be to its long-term payback percentage for a given number of spins. The more spins on a machine, the closer we expect the actual payback percentage to be to the long-term, just as the more we flip a coin, the closer we expect the probability of having flipped heads (or tails) to be to 0.5.

The actual payback on a machine is not expected to exactly equal its long-term payback, just be close to it. Depending on the volatility of the machine, the actual payback percentages on most machines are within a few percentage points of their long-term paybacks after a few million spins.

Let's say a 90% machine has been out on the slot floor for a year. This is a 3-coin dollar machine and has had 1,000,000 spins played on it, so it has had $3,000,000 played through it. Let's also say it has paid back $2,700,000. It's actual payback percentage is spot on the long-term, 90%.

Now let's say you hit a $10,000 jackpot on this machine. Its new actual payback percentage is 2,710,000/3,000,003 = 90.3%, well within the range of paybacks expected for a machine with this much play. And as the machine gets more play, payouts — even large ones — have less and less of an effect on the machine's actual payback percentage.

There's no need for the machine to have to go into pay and take cycles to achieve the long-term payback percentage. Random Sampling with Replacement takes care of it automatically.

Best of luck in and out of the casinos,
John

Send your slot and video poker questions to John Robison, Slot Expert, at slotexpert@comcast.net. Because of the volume of mail I receive, I regret that I can't send a reply to every question. Also be advised that it may take several months for your question to appear in my column.