Indiana, RNGs & Allegations: Time to Test

Reply #28

Todd — Sat, 18 Jun 2005 02:39:21 GMT

I know that robust testing of RNGs (or any complex data set for that matter) would require massive amounts of data, but since the lotteries do not give the public much information, that's the only thing we have to rely on. Another reason to NOT use computers to conduct drawings. With lottery balls, you just need to know that they have been carefully weighed and measured (which they do), and the randomness takes care of itself. Any weirdness in the drawings can then be considered the strangen... [ More ]

Reply #27

MathWizard — Sat, 18 Jun 2005 02:34:39 GMT

I absolutely agree that there could be a programming bug. And, a bug is unlikely to be discovered without a huge number of samples. DIEHARD (a program for testing RNG algorithms) generally requires a minimum of 3 million data points to make a decision about the randomness of an algorithm. Basic mathematical tests like chi-squared could show a tendency towards non-randomness with perhaps 25,000. It would not be a proof, but if there is a problem in the sample, then further investigation wou... [ More ]

Reply #26

LOTTOMIKE — Sat, 18 Jun 2005 02:21:56 GMT

lol,thats pretty bad a million dollar corporation like indiana lottery using a free download off the internet just to save a few pennies.i don't understand why these people don't see that when you change from ball drawings to computers most of the time sales do suffer...........

Reply #25

Todd — Sat, 18 Jun 2005 02:11:15 GMT

MathWizard, Sometimes, like in the case of California's Daily Derby, an unintended problem exists that the lottery does not know about. In that game, it made certain combinations of numbers impossible to win, while making it easier to win for others. There was no one fired and no one ordered it up at the CA Lottery, but it existed all the same. Denying that such a problem could exist in the Indiana Lottery would be denying that computer programs have bugs. As a computer programmer, yo... [ More ]

Reply #24

MathWizard — Sat, 18 Jun 2005 02:03:50 GMT

To prove this one way or the other mathematically would require tens of thousands of draw results and that would only work if the draw corruption was extremely blatant.Logically, however, the facts that triples result in large payouts, and that triples are appearing at an expected frequency, tends to discredit the concept that there is an algorithmic procedure manipulating the draw results. Such a program, were it to be used, would have to be intelligent enough to know when high payout results

Reply #23

RJOh — Sun, 05 Jun 2005 02:02:57 GMT

It's almost possible with ball drawings too.MEGAMILLIONS - 09/20/02 - 01 02 04 05 46 +08RJOh

Reply #22

Todd — Sat, 04 Jun 2005 20:49:27 GMT

Reply #21

LOTTOMIKE — Sat, 04 Jun 2005 20:48:05 GMT

with computerized drawings you never know 1-2-3-4-5-6 is a real possibility.

Reply #20

time*treat — Sat, 04 Jun 2005 19:58:23 GMT

mathematical proof:without knowing how much is being played on each number, each day; as long as the overall number outcome is in line with expected stats, you can't prove rigging.If you can choose from A-to-Z, as long as I make sure each letter shows up NEARLY 1/26th of the time, I just pick from among the less popular (for a particular draw) choices.This sounds worse than it is. Imagine how many (VERY UNHAPPY) winners there would be if something like 1-2-3-4-5-6 ever came out in a lotto.Jackpo... [ More ]

Reply #19

LOTTOMIKE — Wed, 01 Jun 2005 19:31:22 GMT

the only thing you have a descent shot at in indiana is powerball because that drawing is not computerized and they can't mess with it....

Reply #18

LOTTOMIKE — Sat, 14 May 2005 06:40:00 GMT

this is one subject i've thought long and hard about and i still think its rigged somewhere,just my opinion.....

Reply #17

Rick G — Sat, 30 Apr 2005 15:00:11 GMT

Aye',I don't think it can be proven to be illegitimate mathematically is my point. The results and classification percentages are right in line with other states, whether they are computerized or not.I don't know much about RNG testing, but have used the inspector 3 tool extensively on many different states using many different histories. The longer the history the more in line with normal expectations. Indiana does not appear to be out of line statistically.I DO, however, have a problem with... [ More ]

Reply #16

LOTTOMIKE — Fri, 29 Apr 2005 09:18:26 GMT

i'm sure losingjeff would agree....

Reply #15

LOTTOMIKE — Fri, 29 Apr 2005 09:17:56 GMT

i still think indiana is rigged.....

Reply #14

ayenowitall — Sun, 24 Apr 2005 21:59:37 GMT

Quote: Originally posted by Rick G on April 24, 2005The figures CalifDude provided for the last year of IN midday draws are right in line with what you should expect (using the Inspector 3 tool) and with the normal variations of any state's lottery results.However, I conducted an experiment in the past here at L.P. and using the inspector 3, inspected the all-states predictors' picks. Using their overall highest percentage in the different inspector 3 categories and then using the deflate tool... [ More ]

Reply #13

JAP69 — Sun, 24 Apr 2005 16:50:02 GMT

Reply #12

Rick G — Sun, 24 Apr 2005 16:37:15 GMT

The figures CalifDude provided for the last year of IN midday draws are right in line with what you should expect (using the Inspector 3 tool) and with the normal variations of any state's lottery results.However, I conducted an experiment in the past here at L.P. and using the inspector 3, inspected the all-states predictors' picks. Using their overall highest percentage in the different inspector 3 categories and then using the deflate tool eliminating all of their picks I was still able to c... [ More ]

Reply #11

JAP69 — Sun, 24 Apr 2005 16:23:20 GMT

Quote: Originally posted by JAP69 on April 24, 2005 __________pick 3______________type_______combos___% drawn no match______120 _______72% doubles________90________27% triples_________10_________1%_________pick 4_____________ __type________combos____% drawn no match_______210________50.4% doubles________360________43.2% triples__________90_________3.6% double/double___45_________2.7% quads__________10_________.1%pick 4 comparison last year of draws in IndianaNumber Types: ABCD37751.6%8065 38... [ More ]

Reply #10

JAP69 — Sun, 24 Apr 2005 16:04:11 GMT

__________pick 3______________type_______combos___% drawnno match______120 _______72%doubles________90________27%triples_________10_________1%_________pick 4_______________type________combos____% drawnno match_______210________50.4%doubles________360________43.2%triples__________90_________3.6%double/double___45_________2.7%quads__________10_________.1

Reply #9

ayenowitall — Sun, 24 Apr 2005 15:36:13 GMT

It took me a few days to get to this, but I finally discovered Inspector 3...LOL. (I'm not a Pick 3 player.) Well, I inspected the full 1000 Pick 3 possibilities and placed the percentages for each number classification beside CaliDude's inspection of 365 Indiana Midday results. The percentages for the full 1000 possibilities are in red. The positive or negative difference of the sample percentage as compared to the full 1000 possibilities is shown on the extreme right.Considering the sample siz... [ More ]

Reply #8

ayenowitall — Wed, 20 Apr 2005 08:36:02 GMT

four4me,I think we're thinking along the same lines. I suspect that, for whatever reason, the daily games just aren't very popular in Indiana. Maybe the Hoosier Lottery just isn't doing a very good job of promoting. I've noticed that the vast majority of their advertising seems to be geared toward scratch-off games and their TV game show. There's some advertising for the PowerBall game also, but that's about the extent of it.I'm going to try to compare those number classifications that CalifDude... [ More ]

Reply #7

ayenowitall — Wed, 20 Apr 2005 08:24:16 GMT

CalifDude,Thanks for supplying all those numbers. I'm having a little trouble with comparing the 365 actual draw numbers to those 220 unique numbers. It seems like those percentages wouldn't really be comparable. For example, there are ten possible triples. They account for about 4.5% of the 220 unique numbers, but they only account for 1% of the total 1000 possible pick three numbers. Out of 365 numbers drawn, three were triples. That figures out to 0.82%, which seems pretty much in line with t... [ More ]

Reply #6

CalifDude — Tue, 19 Apr 2005 18:59:44 GMT

That was for the last year, by the way.

Reply #5

CalifDude — Tue, 19 Apr 2005 18:59:15 GMT

Here are the Midday results for Indiana. Check the percentages against what I show for the actual 220 numbers in my Blog.Numbers Inspected: 365ClassificationMatchesPercentMatching Numbers High/Low: HHH4512.3%979 989 766 768 698 988 555 995 988 868 567 658 559 596 988 555 896 956 755 978 588 595 767 796 976 975 998 558 787 686 995 977 699 566 678 985 696 666 688 958 687 696 575 779 577 HHL14238.9%872 466 487 097 851 525 580 584 992 289 871 519 187 289 086 584 845 497 636 479 864 056 299 507 837

Reply #4

four4me — Tue, 19 Apr 2005 18:44:52 GMT

ayenowitall test away i was just giving you the examples that were handed me. When i kept informing the posters that we hit the indiana's lottery every day almost pick3/4 games on the prediction pages they replied. Come here and play your numbers for real and you wouln't win. If i ever hit the lottery for a substancial amount i will do that. I will go there and play for a week maybe two and if i dont hit both games straight i will write an apology. I'm also going to try and find out the percenta... [ More ]

Reply #3

ayenowitall — Tue, 19 Apr 2005 08:37:28 GMT

four4me,I think I understand the allegations, but my thinking was that Indiana's draw numbers wouldn't stand up to testing for randomness if they were selectively avoiding the numbers which would cost them most to pay off. Maybe my basic premise was incorrect. As RJOh pointed out in his response, the accepted tests for randomness cannot so easily be applied to draw results. Perhaps its easier to test RNG systems than it is to test a given series of numbers that they might produce. Frankly, the w... [ More ]

Reply #2

four4me — Sun, 17 Apr 2005 20:10:11 GMT

aye What the guys from indians are saying is you can pick the number that might win on paper and it will win. They say the problem is if they actually play the number it woln't win. people on lottery post predictions get the numbers right for indiana all the time. But that just for fun. If you live in indiana and bet 500 pick 3 numbers tonight then they will draw the number you didn't pick. Or so thy say.

Reply #1

RJOh — Sun, 17 Apr 2005 20:00:08 GMT

With 1000 possible combinations in the pick3 game, if one was trying to avoid the 10-20 most popular combinations for any particular drawing and pick only from the remaining 980, a .2% variance couldn't be detected in those type of tests unless those 10-20 combinations were the same for all the games over a period of time. Even in games that use balls, it's normal for certain box combinations not to come up in months and straight combinations not to up in years.RJOh

Indiana, RNGs & Allegations: Time to Test

ayenowitall — Sun, 17 Apr 2005 19:18:30 GMT

We've all read the numerous posts about the Hoosier Lottery's use of RNGs. Most people seem to think that there's something crooked going on. I'll have to admit that I at first thought all the charges of fraud were altogether groundless. After looking at the payouts on Indiana's daily games, I began to think that there was perhaps some substance to accusations made against the Hoosier Lottery. Now, I'm thinking that the lottery games in Indiana simply aren't very popular. That would very easily... [ More ]