Page 2 of 3

Re: UOSA cAmillions testing

Posted: Wed Jul 14, 2010 6:48 pm
by nightshark
There is a problem with doing it by 2 dice.

The reason is that the chance of rolling a 2 or 12 is not the same as the chance of rolling a 6 or 7.

EG:
Chance of rolling a 2 = Rolling (1,1) (1 possibility)
Chance of rolling a 3 = Rolling (1,2),(2,1) (2 possibilities)
Chance of rolling a 4 = Rolling (1,3), (3,1), (2,2) (3 possibilities)
Chance of rolling a 5 = Rolling (1,4),(4,1),(2,3),(3,2) (4 possibilities)
Chance of rolling a 6 = Rolling (1,5), (5,1), (2,4), (4,2), (3,3) (5 possibilities)
Chance of rolling a 7 = Rolling (1,6), (6,1), (2,5), (5,2), (3,4), (4,3) (6 possibilities)
Chance of rolling an 8 = Rolling (2,6), (6,2), (3,5), (5,3), (4,4) (5 possibilities)
Chance of rolling a 9 = Rolling (6,3),(3,6),(5,4),(4,5) (4 possibilities)
Chance of rolling a 10 = Rolling (6,4),(4,6),(5,5) (3 possibilities)
Chance of rolling an 11 = Rolling (6,5), (5,6) (2 possibilities)
Chance of rolling a 12 = Rolling 6,6 (1 possibility)

There are a total of 36 different dice combinations, 6 of those giving you a 7. That means you have a 1/6 chance of rolling a 7. So the best possible ticket is 7-7-7-7-7, followed by 7-7-7-7-(6 or 8 ).. etc. Every number you deviate from center is lowering your odds.

Edit:

So yeah, I enter tickets
7-7-7-7-7
7-7-7-7-6
7-7-7-7-8

Re: UOSA cAmillions testing

Posted: Wed Jul 14, 2010 7:09 pm
by Pacifico
interesting night haha, only problem is the UO dice roll two and you add them up

Re: UOSA cAmillions testing

Posted: Wed Jul 14, 2010 7:12 pm
by nightshark
Pacifico wrote:interesting night haha, only problem is the UO dice roll two and you add them up
Yeah, I imagine whatever algorithm it would use would mimic RL dice anyhow.

The fact that UO dice roll 2 and you have to add them up only strengthens that theory, as it cannot be a simple RNG (random number generator) from 2-12.

Try it out with a sample of 10 or 100, the odds are heavily in favour of rolling 4-9. 60% of the numbers are rolled 83% of the time :P

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 3:42 am
by Chim
What about a raffle? Each person buys the same book for 10k or 20k or whatever. They put there name in a book drop it in a box and then someone axes the box and the book that is on top is the winner. That is simpler

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 6:25 am
by Pacifico
Chim wrote:What about a raffle? Each person buys the same book for 10k or 20k or whatever. They put there name in a book drop it in a box and then someone axes the box and the book that is on top is the winner. That is simpler

Not a bad idea but legitimacy is the big thing with lotteries. Atleast with this you have the numbers, the picked numbers, and the prize/current pot all on the forums. So many things can go wrong with a book raffle

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 12:10 pm
by fooka03
Chim wrote:What about a raffle? Each person buys the same book for 10k or 20k or whatever. They put there name in a book drop it in a box and then someone axes the box and the book that is on top is the winner. That is simpler
It's also not random, it's either the first one in, the last one in, or the last one 'touched' will be the one on top I just can't remember which.

Instead of just saying that the real life probability is implemented here, why don't we actually test it? Take 110 dice rolls and record the results. If each number has 10 hits (+-1) then it's even distribution and the lotto results won't skew towards the middle numbers.

However, even if it does skew towards the middle numbers, with 5 dice rolls there's still enough randomness (artificial in this case) that the outcome will be spread across all numbers. You're just more likely to win with a 5-6-7-8-9 pick rather than a 2-2-2-12-12 pick.

Edit: Nightshark, your logic is a little flawed since for each additional 7 the probability changes. 1/6, 1/36, 1/216, 1/1296, 1/7776 <- your actual chance of winning. Granted this is the best chance of winning but still there's enough other possibilities out there that still make it "fair"

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 5:06 pm
by nightshark
fooka, the logic isn't flawed, that was my whole point. That ticket has the best chance to win. Do some analysis and you'll see it's chance to win is a lot higher than other tickets. Someone else entered 3-5-9-12-2 for example. I calculated my ticket is 243 times more likely to win than this ticket.

I'm sure there are other variables I'm not taking into account, like I haven't considered the fact that the die are unordered. My stats knowledge has faded over the years and it's stuff like this that almost makes me want to crack my uni books open and do some problems for fun. :shock:

IMO, roll 6 1 sided die. The chances of picking the right numbers are still very very low, enough so that you might never see a jackpot winner.

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 5:46 pm
by Pacifico
The drawing will be on top of Glow bank at 6:00 Eastern (14 minutes)

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 5:59 pm
by Pacifico
Officially Closed, Will draw in 3 minutes on top of glow bank

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 6:00 pm
by Pirul
56778
66778
67788
Damn just missed it

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 6:05 pm
by Pacifico
9-5-9-7-5


Marmalade was a witness, and actually was the closest picking 3 correct numbers

7152010.jpg
7152010.jpg (102.49 KiB) Viewed 2190 times

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 7:08 pm
by nightshark
Grats marm.

How does this work?

Do you have to get all 5 numbers correct to win? Or is it the "best tickets"?

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 7:14 pm
by Pacifico
have to get all of the numbers right, might add in if you get 4 right you get like 25% of the pot or something.. like i said its still testing. I think that this rate if it builds up 20k a week just by testing, implementing some kind of funds from players it could get to a million gold after a few weeks of no winners which would be pretty cool

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 7:20 pm
by nightshark
Charge an entry fee.. or even an entry fee for extra tickets.

Like anyone can enter 3 tickets for free, but for 10k gold you get an extra 5 tickets or something. Leave books on a vendor for 10k each with serial numbers that you use to enter. If this becomes a regular thing I can imagine the pot getting large and people gunning for it buying more tickets. Just an idea :)

I'd also love to play some texas holdem, how would you implement something like that though?

Re: UOSA cAmillions testing

Posted: Thu Jul 15, 2010 7:53 pm
by Pacifico
My idea for the lottery:

You post in a thread your numbers (can pick as many sets of 5 numbers as you want) for 100gold per pick
I put a book with the players name as the title for the amount of gold spent. The gold stays on the vendor until someone wins, in which case the gold is removed safely and given to the winner. The book itself doesn't need to be turned it or anything, its just a filler to get the gold from the person to the pot without me having to be on and trading each time.

As for hold em, I found this java based private tournament host thing, so that is an option to have tournaments/sit n gos but I really want to find a real private server where we can set up cash games etc with UOSA gold. I found one alright looking client but it was like 90 dollars a month or a year I dont remember