1. Welcome to VegasMessageBoard
    It appears you are visiting our community as a guest.
    In order to view full-size images, participate in discussions, vote in polls, etc, you will need to Log in or Register.

Slots Payout percentage and machine "randomness", can someone enlighten me?

Discussion in 'Slots' started by Jordan, Oct 2, 2014.

Thread Status:
Not open for further replies.
  1. Jordan

    Jordan Caveman

    Joined:
    Sep 20, 2013
    Messages:
    2,384
    Location:
    Halifax, Canada
    Trips to Las Vegas:
    25
    If a machine has a 98.5% payback % what timeframe is that based on?

    What I really want to know if that if a machine has said 98.5% payout and I sit on a machine for 3 hours hitting nothing, when does it have to give back the percentage? And if it knows that it has only given out 50% for a time and it still needs to pay back 48.5% more then how could it be random winnings?

    Does this even make sense?
     
  2. nostresshere

    nostresshere Mr. Anti Debit Card

    Joined:
    May 4, 2009
    Messages:
    23,220
    Location:
    TN
    Trips to Las Vegas:
    40
    The payout is statisical over a LONG period of time.

    Let's do a simple example. Let's say a machine pays out $1000 only. And, you play $1 a spin to try to win that $1,000. If the machine is a 98.5% payout, it has to (in theory) take in just over $1,015 for each time it pays out. So, even if you played $1014 in a PERFECT situation, it still would not pay you anything.

    Now, in reality, machines do lots of small pays on the way to the jackpot. So, you would have to play way more OVERALL to even get to that $1000 payout. And, since it is all random numbers, you could in theory hit $1,000 on each of your first three pulls. But, it is not going too do that. The little random number is spinning and coming up with possible combinations for each spin. Sometimes $0, sometimes $5, some times $100, sometimes $1000 - for example.

    But, it almost NEVER pays 98.5% over a period of time of hours or days. It does not look at recent payouts and determine it is time to adjust. The math in the system takes care of that.
     
  3. Chuck2009x

    Chuck2009x VIP Whale

    Joined:
    Apr 5, 2009
    Messages:
    14,199
    Location:
    Boston
    Trips to Las Vegas:
    30
    Every slot machine has a finite number of combinations of the number generated by the RNG vs the paytable, a lot win nothing, some win a little, some win more than a little, and one or two win big.

    If you had an unlimited amount of money, and could arrange to sit at that machine, and take the R (random) out of the RNG and have it progress sequentially through every possible combination, without ever repeating one, when you finished the last spin, it would have paid back 98.5% of the amount you put in.
     
  4. Ty

    Ty ?

    Joined:
    Sep 2, 2013
    Messages:
    8,796
    Location:
    Mid Ga
    Trips to Las Vegas:
    22
  5. Snidely

    Snidely VIP Whale

    Joined:
    Nov 2, 2012
    Messages:
    2,129
    Trips to Las Vegas:
    20
    Here is a simple thought experiment for you. Imagine a slot machine that pays out 100%. Instead of the many, many combinations in the typical slot machine, this slot machine only as 10 outcomes. 9 outcomes pay nothing and 1 outcome pays 9 units. Each outcome has an equal probability. It's easy to see that the game is 100% but randomness enters into reality. After 20 spins you might be up or down. As number of spins approaches infinity, the payout approaches the theoretical. Now, imagine another slot machine similar to the one just described. The one also pays out 100%, too, but it has 1 million possible outcomes. 1 outcome pays 999,999 units and the other 999,999 spins pay nothing. Both slot machines are 100% payback games. The difference is the variance of the probability distribution.

    Back to the original question, the 98.5% is not based on a number of spins. It's the theoretical. How close to to the theoretical payback you'll see in your time at the machine is based on the variability of the game. For example, Jacks or Better is a low variance game and you'll results will be closer to the theo in a shorter amount of time. Triple bonus is higher variability.

    When people speak of higher variance games, they'll say things like, "you'll either win or lose real quick in that game". In a low variance game people will say, "you don't win a lot but you don't lose a lot, either."
     
  6. Auggie

    Auggie Dovahkiin

    Joined:
    Jan 8, 2009
    Messages:
    5,820
    Location:
    Burnaby, BC
    Trips to Las Vegas:
    17
    Its not based on a time frame... the machine's 98.5% payback is just a theoretical number.

    If a machine as a payback of 98.5% then in the long run it will usually reach that payback, but technically it doesn't have to... it could be set to a payback of 98.5% but only ever pay back 90% over the long run.

    What payback means is: its how much a machine is expected to return to the player based on wager versus return versus the probability of all events on the machine happening.

    A really simplified example would be:
    Imagine there is a slot machine that has only one reel and on that reel its blank except for a single spot with a "$100" symbol.
    Player puts in $1 and hits "SPIN" and if the "$100" symbol comes up they win $100, anything else wins nothing.

    If the odds of the "$100" symbol coming up is 1 in 100 then that would be an example of a machine with 100% payback (0% house edge) - the odds of the event happening is equal to the payback (100/100 = 1.00 = 100% payback)

    If the odds of the "$100" symbol coming up is 1 in 105 then that would be an example of a machine with 95.2% payback (4.8% house edge) - the odds of the event happening is greater than the payback (100/105 = 0.952 = 95.2% payback)

    In the end that doesn't mean the machine has to pay back 95.2%, but that in theory it is expected to and most likely in the long run it will...
     
  7. nostresshere

    nostresshere Mr. Anti Debit Card

    Joined:
    May 4, 2009
    Messages:
    23,220
    Location:
    TN
    Trips to Las Vegas:
    40
    Some very good explanations - most better than mine.
     
  8. tringlomane

    tringlomane STP Addicted Beer Snob

    Joined:
    Jan 21, 2013
    Messages:
    31,303
    Location:
    Chicagoland
    Trips to Las Vegas:
    28
    Considering the data though and the text describing the graph, I'm pretty sure it's completely fabricated.

    And as others have said, 98.5% is what the machine is expected to payout in the long, long term. It doesn't have to do this at any given time though. But the "law of large numbers" shows the more the game is played with the same expected paybacks, the closer the average actual results will approach the theoretical expected average.

    http://en.wikipedia.org/wiki/Law_of_large_numbers

    The wiki page above has an excellent image in the upper right corner of the page showing how rolling one random die has the tendency to approach the average value of 3.5: (1+2+3+4+5+6) / 6 = 3.5

    Now for more volatile games like video poker, larger samples are needed to see this trend come to fruition.

    Return Percentage of 1000 samples of 1000 hands of 9/6 JoB (expected return 99.54%)

    [​IMG]

    Return Percentage of 1000 samples of 100,000 hands of 9/6 JoB (expected return 99.54%)

    Note the second graph is much closer to approaching the expected 99.54% return. Almost every sample is within 5% of the mean now.

    [​IMG]

    And this trend will continue the more hands machine plays out.
     
  9. Jordan

    Jordan Caveman

    Joined:
    Sep 20, 2013
    Messages:
    2,384
    Location:
    Halifax, Canada
    Trips to Las Vegas:
    25
    Thanks for the responses peeps, makes a little more sense.
     
  10. smartone

    smartone VIP Whale

    Joined:
    Jul 17, 2011
    Messages:
    9,892
    Location:
    Northern Nevada
    Trips to Las Vegas:
    248
    Yep... I've always wondered the same thing... this helps a lot!
     
  11. Buddha

    Buddha VIP Whale

    Joined:
    Apr 3, 2012
    Messages:
    1,470
    Location:
    Chicago 'burbs
    Trips to Las Vegas:
    60
    Basically, every machine is based on the probability of every resulting hand being dealt, the probability of all the various results of the re-draw, combined with the payback chart of every completed hand. The machine has no memory of what hands have been completed, and each hand is an independent trial of all the existing possibilities. The machines are not able to create hands solely for the purpose of making the payback percentage come out as expected. Only an INFINITE number of trials will approach that percentage.
     
  12. Snidely

    Snidely VIP Whale

    Joined:
    Nov 2, 2012
    Messages:
    2,129
    Trips to Las Vegas:
    20
    Along these lines, here is a simpler gambling game. A fair coin is flipped. If you bet $100 and win, you get $198 back. You win $98. If you lose, you lose your $100. The game has a 1% house advantage. You know intrinsically that there is no numbers of coin flips where you'd expect it come out exactly 50/50 heads/tails but you know the theoretical is 50/50 and the house is taking their cut.
     
  13. nostresshere

    nostresshere Mr. Anti Debit Card

    Joined:
    May 4, 2009
    Messages:
    23,220
    Location:
    TN
    Trips to Las Vegas:
    40
    I find this interesting.

    The original post only talked about a machine. No mention of Video Poker. Yet, some of the answers talk about Video Poker.

    Hmmmm
     
  14. tringlomane

    tringlomane STP Addicted Beer Snob

    Joined:
    Jan 21, 2013
    Messages:
    31,303
    Location:
    Chicagoland
    Trips to Las Vegas:
    28
    I related it to Video Poker because the input of my "ad hoc" simulator is based on video poker probabilities. But the same general idea applies to slots too. If you created a slot machine with the same expected probabilities and payouts as 9/6 JoB, you would get similar looking graphs that I posted (not exactly the same though because you would be taking a different random sample). Most jurisdictions lump video poker results in with slot machines, fwiw.
     
  15. bardolator

    bardolator Lifelong Low Roller

    Joined:
    Mar 30, 2005
    Messages:
    3,055
    Location:
    Gig Harbor, WA
    Trips to Las Vegas:
    999
    Can anybody point out a slot machine in Las Vegas that is known by players to have a 98.5% payout? A Video Poker game pretty much has to be the theoretical "machine".

    If you have lost ten million dollars a nickel at a time on a machine that pays 98.5%, your odds on the next pull have not changed.
     
  16. 3544quebec

    3544quebec High-Roller

    Joined:
    Mar 19, 2013
    Messages:
    881
    Trips to Las Vegas:
    5
    But the odds of you having arthritis in your index finger would be as close as possible to a sure thing :eek:
     
  17. Buddha

    Buddha VIP Whale

    Joined:
    Apr 3, 2012
    Messages:
    1,470
    Location:
    Chicago 'burbs
    Trips to Las Vegas:
    60
    A video poker machine is based on a fixed deck of 52 cards (except for joker poker), and all the probabilities and percentages can be precisely calculated. A slot machine is "programmed" by a computer chip that determines the "unknown" number of stop positions, and combinations of symbols. The casino is able to put in place different "chips" with different programming, and you would never know it.
    It's possible to walk up to a bank of "identical looking" slot machines, and they could possibly all have a different programming, and different payback percentages.
     
  18. Ty

    Ty ?

    Joined:
    Sep 2, 2013
    Messages:
    8,796
    Location:
    Mid Ga
    Trips to Las Vegas:
    22
    I mentioned VP since it's the only machine I know of with a stated return. I feel my post was a fair reply to the OP's question.
     
  19. bardolator

    bardolator Lifelong Low Roller

    Joined:
    Mar 30, 2005
    Messages:
    3,055
    Location:
    Gig Harbor, WA
    Trips to Las Vegas:
    999
    "But the odds of you having arthritis in your index finger would be as close as possible to a sure thing."

    Shit, I thought I got it from texting. :)

    I was just being wildly theoretical. The chance of losing ten mill at nickels probably depends more on heredity and eating habits than it does on the hold.
     
    Last edited: Oct 4, 2014
  20. tringlomane

    tringlomane STP Addicted Beer Snob

    Joined:
    Jan 21, 2013
    Messages:
    31,303
    Location:
    Chicagoland
    Trips to Las Vegas:
    28
    I dunno of any non-progressive games like that. With the progressive, Lion's Share was probably above 98.5%, but that can't really be proven either without the PAR sheet.
     
Thread Status:
Not open for further replies.