## Help create a real actual computer simulated wheel (oxymoron)

Started by gizmotron, September 13, 2008, 03:45:15 AM

0 Members and 1 Guest are viewing this topic.

#### gizmotron

I'm seeking advice for conditions that make up the attributes of a real ball being sent around a track and then dropping into a slot.

List of variables:

random ball release speeds (within a range) -- ball travels from 125 to 175 feet

random speeds of wheel speed (within a range) -- changes distance ball rolls

random ball size and weight -- changes distance ball rolls

circumference of track -- allows ball to drop into or miss bumper diamonds

circumference of slot boundary -- allows ball to hit along a fixed position of slots on the wheel

random hitting of bumper diamonds -- creates a random bounding bounce

random strikes of slot channels -- makes the ball bounce, rattle/bounce, or hit

random conditional bounces -- wild card random bounce

calculated slot hits based on distances from center of slot channels

secondary local bounces do to near misses of channel flange strikes

simulated gravity

simulated motion

With all those conditional variables and possible actual simulated physical ball strikings I would have to say that that would have to be more random than the noise created by the universe as a basis for charting randomness and simulated actual spins.

Any ideas welcome

P.S.

Are these the proper order for slots on 0 / 00 wheels?

0,26,3,35,12,28,7,29,18,22,9,31,14,20,1,33,16,24,5,10,23,8,30,11,36,13,27,6,34,17,25,2,21,4,19,15,32,

00,27,10,25,29,12,8,19,31,18,6,21,33,16,4,23,35,14,2,0,28,9,26,30,11,7,20,32,17,5,22,34,15,3,24,36,13,1,

Golden Spiral

Might be useful.

TTT

#### gizmotron

Yeah, that looks great. Funny too. It looks like our money going down the drain. No Wonder it's golden. What is the name of that place, "Golden Casino?"

With this it will only need to be calculated down to about "E."

Thanks

#### winkel

Do you know that this is based upon the Fibonacci-Numbers

#### gizmotron

Quote from: winkel on September 13, 2008, 01:34:44 PM
Do you know that this is based upon the Fibonacci-Numbers

Nope, I knew it looked like some kind of geometry. Good old Leonardo of Pisa. Actually, I'm not a math guy. I was looking at the girls and day dreaming.

If I can assemble these attributes and variables into a multi-faceted formula then we can have a widely accepted testing wheel. I just want help with the math. In fact I want it peer reviewed by the experts here step-by-step, and until we all will have contributed to its build. It's supposed to be free and cross-platform, unless someone can give a reason to charge for it.

Here's one. Like this: (2 x 3.14 x Radius) gets me the length around the track. But I need a general consensus how many times the dealers send the balls around the track before they drop out of these tracks.

Any experts are welcome to advise all along the way.

#### winkel

In Germany it is rule that the ball should be that fast, that it is able to spin more than ten circles before it leveas the outest way

#### gizmotron

Quote from: winkel on September 13, 2008, 02:48:13 PM
In Germany it is rule that the ball should be that fast, that it is able to spin more than ten circles before it leveas the outest way

So on a 24 inch wheel I get a radius of 11.5 inches for a length of track of 72.22 inches. That's basically 6 feet rounded. If the minimum is 10 revolutions then 60 feet is the lowest number. Double that for the high number of the range, I guess? (60 to 120 feet)

Interesting.

#### bliss

Gizmo, rather than re-invent the wheel (ha ha) you might like to try a physics engine: nolinks://en.wikipedia.org/wiki/Physics_engine

Some of the external links at the foot of the page might be worth checking out.

#### gizmotron

I have a radius of 7 inches for the outside edge of the channels for the slots. So there is a 4.5 inch drop from the track to the slots. I need a range for this so that different sized/weighted balls can have a randomness to them falling out of that track and striking some point along the circumference of the slot boundary.  I will then use a simple arithmetic to calculate if a ball hits a bumper diamond or if it travels free and clear down to the slots. I will also know if the ball hits a slot dead center or on the edge of a slot channel.

I'm going to add the diameter of a ball too.

#### gizmotron

Bliss, that's too complicated. I only need ranges of values and apply a randomness to select different values for each step along the way. I'm not trying to create the actual physics involved but I'm trying to produce a truly random result that is far beyond the techniques for most RNGs.

Like 100 different strike points at the top of a single slot to determine a bounce. Like 1000 different bounces & directions. It doesn't need to be that real.

Perhaps it wouldn't be accepted if it were not real enough? Is there a way to test it's trueness to randomness?

#### bliss

Gizmo, why are you trying to create something more random than an RNG? I'm not even sure that it makes any sense. If you wanted to model the physics of a wheel to test for stuff like wheel tilt and dealer signature that would be understandable and probably more interesting, but the perfect RNG would be meaningless for testing roulette systems, because it could never exist.

Anyway, here is a well-known RNG test suite: nolinks://nolinks.stat.fsu.edu/pub/diehard/

#### gizmotron

Quote from: bliss on September 13, 2008, 03:47:34 PM
Gizmo, why are you trying to create something more random than an RNG? I'm not even sure that it makes any sense. If you wanted to model the physics of a wheel to test for stuff like wheel tilt and dealer signature that would be understandable and probably more interesting, but the perfect RNG would be meaningless for testing roulette systems, because it could never exist.

Your point is well made. I started talking about this just to see what people would accept. What bothers me is that many people accept this: "RANDOM.ORG offers true random numbers to anyone on the Internet."  -- for a fee!

I wanted a program that is as good as this, based on roulette, and spits out numbers good enough to be considered truly random. It would also spit out lists of 1,000s of spins for text save. It would be wonderful for sim development.

It could be accessed from a CGI too. It could be a peer to peer test engine for gambling competitions. It could automatically be FTPed to a download link for timed spins continuously spitting out numbers every 2 minutes. It could spit out an encrypted list of spins that can only be decrypted one spin at a time every 1 to 2 minutes apart. There could be a competition test platform that allows each spin/bet combo to be saved and encrypted as you go along. I have bet/result data savable and loadable in Roulette Ride 1.1 now. That data could be encrypted. Testing could go to a live action format. We could do live competitions. I could pass all this through a chat engine.

Add this too. It could have dealer change. The new dealer could have a different common release point and speed of launch.

But! ...not if it's dead on arrival.

#### TwoCatSam

Mark

I can barely count with my shoes on, but if I can help I want to.

How can I?

Sam

#### bliss

I thought Victor was working on a similar idea for competitions and challenges, not sure if he's actually started on it yet, but I've heard it mentioned a couple of times.

Random.org's spins are generated by a TRNG (from atmospheric noise), which are supposedly superior to anything a software RNG can produce. The best pseudo-rng is reckoned to be the "mersenne twister", which is the standard RNG algorithm in a lot of modern programming languages - maybe runrev too.

#### bliss

If I remember rightly Victor was planning to fetch the spins from Wiesbaden casino in real time, using some kind of screen-scraper.