Stories
Slash Boxes
Comments

SoylentNews is people

posted by cmn32480 on Wednesday August 05 2015, @03:37PM   Printer-friendly
from the corrupt-random-number-algorithm dept.

Aleksandar Vulovic, czar of the Serbian state lottery, has resigned from his position after winning numbers were mysteriously broadcast on television before they had been drawn.

Allegations of fraud abound on every hand following the incident, reports the Associated Press.

In what was presented to the public as a live broadcast on Tuesday evening, one of the numbers in the winning combination was revealed before the number appeared to be drawn.

"That sparked accusations that the numbers had been chosen in advance," according to AP.

Vulovic has denied there was any dodginess and explained that the incident was due to a "technical mistake".

He added that he was stepping down from his position, not out of guilt, but out of "moral obligation".


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 1, Touché) by Anonymous Coward on Wednesday August 05 2015, @04:37PM

    by Anonymous Coward on Wednesday August 05 2015, @04:37PM (#218628)

    So you're saying the probability was greater than 1?

    Something doesn't seem to be right with your calculation.

    Starting Score:    0  points
    Moderation   +1  
       Touché=1, Total=1
    Extra 'Touché' Modifier   0  

    Total Score:   1  
  • (Score: 2, Insightful) by wantkitteh on Wednesday August 05 2015, @05:14PM

    by wantkitteh (3362) on Wednesday August 05 2015, @05:14PM (#218649) Homepage Journal

    Try reading it again, makes perfect sense to me. And anyone else with basic literacy/numeracy.

    • (Score: 3, Informative) by Immerman on Wednesday August 05 2015, @10:07PM

      by Immerman (3985) on Wednesday August 05 2015, @10:07PM (#218805)

      It may make sense to you, but the formula described is mathematically invalid. Let's look at some totally made up numbers:
      Odds of reading error E= 0.001
      Number of balls B = 50
      Number of lotteries drawn using the technology L= 1,000,000

      Using the formula offered the odds of this "predictive" error occurring at least once would be would be approximately E/B * L
      0.001 / 50 * 1,000,000 = 20: a 2000% chance!

      And as anyone familiar with probability can tell you, if it's possible to get a probability greater than 1 out of a formula, you've done something very wrong.

      In this case they started out okay: The probability of a a single reading error matching the next ball read would be equal to the chance of a read error occurring (E) times the chance of that the next ball would match the read error (1/B). Where they went wrong is in combining the probabilities of independent events (lotteries drawn using the technology). Rather than a simple multiplication, which can easily exceed 1, you need to look at the probability that the event *won't* occur in a given sample size. (or so I remember from stats class)

      Basically if P is the probability that an event will occur, then the probability that the event will have occurred at least once in a sample of N is 1-(1-P)^N. Or in english: the odds that it WON'T happen that (it won't happen the first time * it won't happen the second time * ... * it won't happen the Nth time).

      So using my made-up numbers again we'd get 1 - (1 - 0.001/50)^1,000,000 ~= 0.999999998.

      • (Score: 2) by wantkitteh on Thursday August 06 2015, @12:14AM

        by wantkitteh (3362) on Thursday August 06 2015, @12:14AM (#218859) Homepage Journal

        My maths is a little rusty, but even I can see where you've made a huge mistake there. Of course you can come up with a probably of higher than 1 - you're not looking at a single event. Duh!

        • (Score: 3, Informative) by Immerman on Thursday August 06 2015, @01:24AM

          by Immerman (3985) on Thursday August 06 2015, @01:24AM (#218898)

          No, you're confusing probability with expected outcome.

          1 is the absolute highest probability possible - it means this absolutely will happen, you can't get any more likely than that. What would it even mean to say that there's a 2000% chance that something would happen? It may be very, very likely to occur more than once in a given sample size, but that probability will still be no greater than 100%.

          There's other math you can do to figure out how many times an event is likely to occur in a given sample size, the likely standard deviation of that number, etc. But those are not probabilities, that's using probability to determine expected outcomes - a thing which is back in the realm of countable numbers rather than probabilities, and thus very well may be greater than one.

          Consider flipping a coin with a 50% chance of landing heads up. Flip it ten times and you can expect it to come up heads roughly 0.50*10 = 5 times. That's the expected outcome. But there is a small chance that you'll get all tails ( (1-0.50)^10 = 0.098% ) And thus the odds of getting at least one head head must be will be less than 1 by that amount (only 99.9%). Flip it a billion, billion times and there's *still* a chance that you'll get all tails - infinitesimally small, but not quite zero, and thus the odds of getting at least one head is *still* ever so slightly less than one.

          • (Score: 2) by wantkitteh on Thursday August 06 2015, @10:26AM

            by wantkitteh (3362) on Thursday August 06 2015, @10:26AM (#219025) Homepage Journal

            What are you, some sort of maths troll?

            Firstly, the poster you originally replied to made no suggestion that the probability of this happening was greater than one, merely that they saw "no reason why that should be considered vanishingly small." Your own maths bear that out, you set the probability using your example figures at 0.999999998.

            Secondly, the poster described this as the "chance" this would happen, not the probability.

            Thirdly, you've mis-constructed your probability formula from the original poster's words. There are two sentences there - see if you can tell which one is the "probability formula" you're speaking of and which is the extension where it's turned into the expected outcomes you seem to know the difference between so well, yet completely failed to differentiate between earlier.

      • (Score: 0) by Anonymous Coward on Thursday August 06 2015, @10:54AM

        by Anonymous Coward on Thursday August 06 2015, @10:54AM (#219029)

        And as anyone familiar with probability can tell you, if it's possible to get a probability greater than 1 out of a formula, you've done something very wrong.

        ..or quantum mechanics ;)