4 ms·

you cant' improve the probability in roulette with a pattern

robomartin 14y ago
I didn't say I improved the probability of the game. I improved the probability of my betting decisions producing favorable outcomes. Huge difference. Think of it as betting by considering the physics of the game vs. in complete ignorance. Betting single numbers, black, white, dozens, columns and rows is absolutely moronic once you stop to think about how the game actually works.
- eru 14y ago
  So what physics did you take into account?
  - ww520 14y ago
    The roulette is unbalanced and tilted one way?
    - robomartin 14y ago
      No. Read my reply to "eru".
      - eru 14y ago
        No need to scare-quote the nicknames here.
        robomartin 14y ago
        Not intended that way at all. I didn't even know the term "scare quote" existed.
  - ghshephard 14y ago
    We're being trolled by robomartin, which is not normally acceptable behavior on HN.
    - sophacles 14y ago
      No you aren't, roulette is not an RNG, it is a PRNG. If you take into account the initial marble velocity and rate of rotation of the wheel, you can seriously reduce the probability space of where the ball will land. Further, this can be modelled statistically pretty simply. To win at betting with roulette you only need to bet on a set of numbers, where the probability of the marble landing on that set is pretty good, and the payout of winning over time covers the losses from being wrong. So say you can accurately guess which 1/4 of the wheel will capture the marble. (and there are no greens... the real math is the same but uglier a bit). You put down equal bets on the 9 numbers represented by that. Each beting round will net you a 4:1 payout. Now say you can only guess that 1/4 of the wheel, 50% of the time. You will still average a 2:1 payout. Guessing the 1/4 of the board 50% of the time is not that hard, you just need to calculate the linear drift of the time the marble passes a fixed point, and know the frequency of "loops" that is too slow for the marble to stay on the track. Accounting for the rotation of the board is straight-forward, it is almost constant for a given play. The 50% covers the wierd bounces. Just calculate board position for the time that the marble will get below the threshold, and the marble position as well, and you've won. A person can't do it well, but an arduino is overkill for it. Edit: further notes on this- casinos will be able to detect this easily. All they have to do is look for people on hot streaks that are "too hot", then compare betting placement with wheel positions. And they generally kick people out who are winning too much anyway, even if they can't prove 'chaeting'.
      - seanx 14y ago
        There was a book on this a few years back: http://www.amazon.com/exec/obidos/ASIN/0140167129/jimloyshomepage http://www.amazon.com/exec/obidos/ASIN/0140167129/jimloyshom...
      - tomfakes 14y ago
        And this is why shoe computers are banned in casinos! Google Glass will make this hack so simple, the casino's will have to ban them from being anywhere close to a roulette table to reduce team cheating possibilities.
  - robomartin 14y ago
    I won't spell it out here because the fun is in working out these problems, at least it is for me. I'll shove you in the right direction though. Betting on a single number is foolish. The mathematical probability of hitting any single number is only 2.7%. In other words, the house absolutely loves you. Dozens and Columns are better, 32.4%. Still, the house advantage is huge. Even/Odd or Red/Black get you up to 48.6%. Hey, that's a deal, right? Well, no. The problem with all of these calculations you'll find around the internet is that they are equivalent to picking a random number (or 12 or whatever) out of 38. The problem is that this is NOT how the roulette works. The roulette has a spinning wheel with numbers separated by fences. The ball spins as well. Along the path of the ball there are also little bumpers that can alter the path of the ball. Yes, in the long term it is just like picking random numbers. However, if you look at what makes you win or loose at the "micro" level it is a little simpler. If you bet on a single number the ball only has to move one slot for you to loose. As it lands on the number field the ball has a bunch of excess energy. The surfaces are hard and it bounces around with great efficiency. It can go over one or more number fences very easily. In fact, if you study every single common bet, single numbers, rows, columns, dozens, color, even/odd, etc. they all share one trait: If the ball jumps over just ONE fence you loose. If you've played roulette enough you'll eventually run into this stereotype: Big guy. Smells and looks like money. Shows up with piles of cash, gold watch and gold chain around his neck. And starts to place piles of cash on numbers he is pulling out of his anal orifice. Most of the time these guys loose a ton of money. Why? The "one fence" problem. OK, I think that sets it up. Actually, it is probably beyond obvious at this point. Take a look at a roulette wheel and think about how you should bet to solve the problem that makes every single common bet nearly worthless. The genius of the roulette is in the wheel.
    - tiziano88 14y ago
      I don't get it, surely if I bet on only one number and win, my return must be higher than if I bet on even/odds. it's not the probability that counts, rather the expected return.
      - robomartin 14y ago
        Look at the table. Find two numbers too bet on. Now find those numbers on the wheel. If I drop the ball directly on one of your numbers from, say, 1 inch above the number. Wheel not spinning. How far would the ball have to bounce for you to loose? Bet on the dozens, columns, rows, black/red and other common ( and convenient) bets. Repeat the exercise.

I didn't say I improved the probability of the game. I improved the probability of my betting decisions producing favorable outcomes. Huge difference. Think of it as betting by considering the physics of the game vs. in complete ignorance. Betting single numbers, black, white, dozens, columns and rows is absolutely moronic once you stop to think about how the game actually works.

eru 14y ago
So what physics did you take into account?
- ww520 14y ago
  The roulette is unbalanced and tilted one way?
  - robomartin 14y ago
    No. Read my reply to "eru".
    - eru 14y ago
      No need to scare-quote the nicknames here.
      - robomartin 14y ago
        Not intended that way at all. I didn't even know the term "scare quote" existed.
- ghshephard 14y ago
  We're being trolled by robomartin, which is not normally acceptable behavior on HN.
  - sophacles 14y ago
    No you aren't, roulette is not an RNG, it is a PRNG. If you take into account the initial marble velocity and rate of rotation of the wheel, you can seriously reduce the probability space of where the ball will land. Further, this can be modelled statistically pretty simply. To win at betting with roulette you only need to bet on a set of numbers, where the probability of the marble landing on that set is pretty good, and the payout of winning over time covers the losses from being wrong. So say you can accurately guess which 1/4 of the wheel will capture the marble. (and there are no greens... the real math is the same but uglier a bit). You put down equal bets on the 9 numbers represented by that. Each beting round will net you a 4:1 payout. Now say you can only guess that 1/4 of the wheel, 50% of the time. You will still average a 2:1 payout. Guessing the 1/4 of the board 50% of the time is not that hard, you just need to calculate the linear drift of the time the marble passes a fixed point, and know the frequency of "loops" that is too slow for the marble to stay on the track. Accounting for the rotation of the board is straight-forward, it is almost constant for a given play. The 50% covers the wierd bounces. Just calculate board position for the time that the marble will get below the threshold, and the marble position as well, and you've won. A person can't do it well, but an arduino is overkill for it. Edit: further notes on this- casinos will be able to detect this easily. All they have to do is look for people on hot streaks that are "too hot", then compare betting placement with wheel positions. And they generally kick people out who are winning too much anyway, even if they can't prove 'chaeting'.
    - seanx 14y ago
      There was a book on this a few years back: http://www.amazon.com/exec/obidos/ASIN/0140167129/jimloyshomepage http://www.amazon.com/exec/obidos/ASIN/0140167129/jimloyshom...
    - tomfakes 14y ago
      And this is why shoe computers are banned in casinos! Google Glass will make this hack so simple, the casino's will have to ban them from being anywhere close to a roulette table to reduce team cheating possibilities.
- robomartin 14y ago
  I won't spell it out here because the fun is in working out these problems, at least it is for me. I'll shove you in the right direction though. Betting on a single number is foolish. The mathematical probability of hitting any single number is only 2.7%. In other words, the house absolutely loves you. Dozens and Columns are better, 32.4%. Still, the house advantage is huge. Even/Odd or Red/Black get you up to 48.6%. Hey, that's a deal, right? Well, no. The problem with all of these calculations you'll find around the internet is that they are equivalent to picking a random number (or 12 or whatever) out of 38. The problem is that this is NOT how the roulette works. The roulette has a spinning wheel with numbers separated by fences. The ball spins as well. Along the path of the ball there are also little bumpers that can alter the path of the ball. Yes, in the long term it is just like picking random numbers. However, if you look at what makes you win or loose at the "micro" level it is a little simpler. If you bet on a single number the ball only has to move one slot for you to loose. As it lands on the number field the ball has a bunch of excess energy. The surfaces are hard and it bounces around with great efficiency. It can go over one or more number fences very easily. In fact, if you study every single common bet, single numbers, rows, columns, dozens, color, even/odd, etc. they all share one trait: If the ball jumps over just ONE fence you loose. If you've played roulette enough you'll eventually run into this stereotype: Big guy. Smells and looks like money. Shows up with piles of cash, gold watch and gold chain around his neck. And starts to place piles of cash on numbers he is pulling out of his anal orifice. Most of the time these guys loose a ton of money. Why? The "one fence" problem. OK, I think that sets it up. Actually, it is probably beyond obvious at this point. Take a look at a roulette wheel and think about how you should bet to solve the problem that makes every single common bet nearly worthless. The genius of the roulette is in the wheel.
  - tiziano88 14y ago
    I don't get it, surely if I bet on only one number and win, my return must be higher than if I bet on even/odds. it's not the probability that counts, rather the expected return.
    - robomartin 14y ago
      Look at the table. Find two numbers too bet on. Now find those numbers on the wheel. If I drop the ball directly on one of your numbers from, say, 1 inch above the number. Wheel not spinning. How far would the ball have to bounce for you to loose? Bet on the dozens, columns, rows, black/red and other common ( and convenient) bets. Repeat the exercise.