[worried_citizen]

I recently purchased a OneRNG trng and decided to give it a chance at guessing the right number with its quantum-noise super powers. Reading from it overnight produced 24 million guesses and no winner. Though several were just one number off. Too bad so sad.

I'm really curious what RNG the registers use when people say "let the computer guess for me". I'm pretty sure they don't all have a TRNG installed on the circuit...

[dsl]

"Quick Pick" numbers are not generated at the store. Regardless of if you pick the numbers or let the computer do it, your request is sent in real time back to a lotto office where the real ticket is securely generated (with a TRNG if you so choose). The local terminal then receives a bitmap to print out.

[worried_citizen]

Do you have a source for this? I'm extremely curious. Random number generation is a fun problem.

[dsl]

It is based on my reverse engineering of a lottery terminal. I can't vouch for the quality of the RNG, as I alluded to in my previous comment, but the local RNG isn't used in any meaningful fashion.

I can point you in the direction of the MontaVista Linux distribution, which is what most of the terminals run.

[s73v3r]

Funny that they allow multiple people to pick the same numbers, then.

[PhantomGremlin]

Funny that they allow multiple people to pick the same numbers, then.

Why shouldn't they? It's to the lottery's benefit. Two reasons:

1) People are attached to "their" numbers, and would be very upset if you didn't let them pick exactly the ones they want. Alienated players equals lower participation.

2) There are more overlaps when people pick than when the computer picks. Which means that, each drawing, there is a lower probability of someone hitting the jackpot than if all numbers are distinct. Which means the money rolls over and the next drawing has a larger prize. Eventually the prize is $800 million and people get excited and buy even more tickets. It's even being discussed on HN.

[mulmen]

Is there reason to believe that a TRNG would somehow select numbers with a better chance of winning than any other method?

I can see the method impacting the number of dollars you win but surely any valid set of numbers has the same chance to win as any other.

It's not like your TRNG is actually emulating the behavior of the balls in the machine.

[kod]

It would impact the likelihood of someone else winning as well, which does affect the amount of money you get.

[andrewaylett]

What it might do is avoid the bias of selecting numbers that other people are also likely to choose. So you're no more likely to win, but you may be less likely to have to share the jackpot.

[worried_citizen]

I can't say if a TRNG has a better chance, but I don't think I'm leaping too far to say a broken, predictable RNG has a _worse_ chance of winning.

[mbreese]

Maybe worse over the long term, but wouldn't you have the same odds for any given draw.

[s73v3r]

No, but I don't have to blame myself for having stupid numbers when I lose :p