Thread Rating:

Ayecarumba
Ayecarumba
Joined: Nov 17, 2009
  • Threads: 232
  • Posts: 6592
October 26th, 2010 at 11:02:16 AM permalink
I'm confused by the two proposed calculations. Which is intended for the final language?

This is described as a "proposed change" in your original post:

Quote: MichaelBluejay


1. I will wager my $30,000 against your $3,000.... A "lifetime" is 60 rounds per hour X 12 hours per weekend X 20 weekends per year X 35 years....




But this is described as the "overhaul":

Quote: MichaelBluejay



(3) ... So I'm thinking of saying that we run the sim for the same number of rounds you could play in a "lifetime" -- 10 hours (for a weekend trip) x 20 weeks a year x 30 years....



Is the final language intended as: "60 rounds/hour X 12 hours/weekend X 20 weekends/year X 35 years" or, "60 rph X 10 hpw X 20 wpy X 30 years"?
Simplicity is the ultimate sophistication - Leonardo da Vinci
SOOPOO
SOOPOO
Joined: Aug 8, 2010
  • Threads: 103
  • Posts: 6016
October 26th, 2010 at 11:05:35 AM permalink
Since you know that no system for the 3 games you mention can win over the number of required bets I would leave the 'win' by your opponent as any win, even a dollar. The 'concept' of a winning system does not imply a certain per hour rate, just that it will win. I truly hope someone takes you up on your offer, because even though you don't necessarily agree, I think you have earned your 'fee' already from the CSI (Craps strategy investment) debacle...
atlarson
atlarson
Joined: Oct 13, 2010
  • Threads: 0
  • Posts: 9
October 26th, 2010 at 11:55:20 PM permalink
Quote: MichaelBluejay

@larson, all your questions are answered specifically in my original post.


I don't see where you explain how you chose the duration of a billion bets resolved (whether those are considered as belonging to a single person over a long session, or to multiple people over many shorter sessions). How did you choose the billion? Why not ten billion, or a hundred million? (Obviously, because you want to get enough "into the long run" for the system to lose. But how do you define that?)

Essentially, how do you convince yourself that the bet you're offering is to your advantage? For a thousand bets, it obviously isn't, and for a billion bets, you're fairly confident that it is; but how did you calculate that?
MichaelBluejay
MichaelBluejay
Joined: Sep 17, 2010
  • Threads: 29
  • Posts: 793
October 27th, 2010 at 4:48:55 AM permalink
@Soopoo: Okay, you and others have convinced me that I'll count any profit as a win. Thanks.

@Ayecarumba. I don't think it matters much whether a weekend lasts 10 hours or 12 hours. That's a minor detail which I can decide on later.

@atlarson: Your question is answered in my original post, 2nd paragraph, 2nd sentence.

@everyone: Before I make this live, I'd like to hammer on it to make sure I'm not missing something that would leave it vulnerable to exploit. So, I'm making a special offer...

Bluejay's $1000 reward!

I'll pay $1000 to the first person who comes up with a betting system that can beat my proposed rules #1-3 in my original post. (i.e., A >10% chance of winning.) You don't have to put up any money; it's free. Just submit a system that can beat my rules, and I'll pay you $1000.

Note that this wouldn't be a real "winning betting system", since there's no such thing, but rather something that exploits an oversight in my rules. If anyone had a real winning betting system then of course they wouldn't share it for $1000.

This offer expires November 3, 2010, or sooner if I get overwhelmed with submissions.
coolhandnuke
coolhandnuke
Joined: Sep 18, 2010
  • Threads: 1
  • Posts: 6
October 27th, 2010 at 5:41:42 AM permalink
Quote:

1. I will wager my $30,000 against your $3,000 (or my $10,000 against your $1,000, if you prefer) that your betting system cannot show an average lifetime profit of even $1 in a game of roulette (single- or double-zero), baccarat, or craps (3-4-5 odds), as the player, using common Vegas rules, in a computer simulation, as per the additional terms below. You win the challenge if your system shows any profit at the end of the simulation, I win if it does not. A "lifetime" is 60 rounds per hour X 10 hours per weekend X 20 weekends per year X 35 years. In craps, a "round" is any roll in which at least one bet is resolved. Since you could show a profit for one lifetime out of pure luck, we'll simulate 1,000 average lifetimes and take the average profit or loss for all lifetimes. This assumes your system places a bet every round; for systems that involve "sitting out" some rounds in order to wait for presence or absence of streaks, we'll simulate additional lifetimes proportional to the number of rounds that were sat out.

2. Your system must be one that could actually be used in a typical Vegas casino, since the point of your challenge should be to prove that you can win in a casino environment. So, for example, methods that attempt to exploit the computer's random number generator are specifically disallowed. Any system that wins by screwing with the RNG cannot win in a real casino. Similarly, any system that cannot be actually be employed by an average person with nothing more than a pen and paper is disallowed, since if a system can't be used in the real world then it's worthless. Also, card-counting systems in baccarat are disallowed, because the whole point of this challenge is to disprove traditional betting systems, where bet size and placement is based on the outcome of previous rounds.



What perplexes me is you run a simulation generated by a computers algorithm (or even from true randomness from noise, radioactive decay, thermal changes, blah blah blahdy blah). Then you state any system submitted must be a system that can be employed in the real world in a real casino (like on a real roulette table?).

My question, in my mind is, you can't have SIMULATED data to find a systems REAL TRUE outcome on a real physical mechanical roulette wheel device combined with human interaction ball spinning dealer as you are not getting data from a GENUINE SOURCE.
winning at roulette is easy, knowing when to cash in and leave is not so easy
MichaelBluejay
MichaelBluejay
Joined: Sep 17, 2010
  • Threads: 29
  • Posts: 793
October 27th, 2010 at 5:57:23 AM permalink
You're right, it would be best to use real-world casino data. But that's impossible. So we make do with what we have. And it's really only a problem for people who think that casino results or a computer sim is less random than the other -- of which you seem to be one. But there's no way I can placate you, so sorry.
weaselman
weaselman
Joined: Jul 11, 2010
  • Threads: 20
  • Posts: 2349
October 27th, 2010 at 7:53:16 AM permalink
Quote: MichaelBluejay



I'll pay $1000 to the first person who comes up with a betting system that can beat my proposed rules #1-3 in my original post. You don't have to put up any money; it's free. Just submit a system that can beat my rules, and I'll pay you $1000.




What's the definition of "can beat"? If my betting system has negative expectation (obviously) but the probability of showing a profit after 1000 "lifetimes" under your rules is more than 10%, is it good enough?

What if it is less than 10%, but only slightly? Even if it is as low as 8.6%, it'd be a losing proposition for your challenger, but still a better bet than playing double-zero roulette.

What if it is even lower? Say, 5%, would not make sense mathematically for a challenger, but is 5% an acceptable risk level for you?
"When two people always agree one of them is unnecessary"
MathExtremist
MathExtremist
Joined: Aug 31, 2010
  • Threads: 88
  • Posts: 6526
October 27th, 2010 at 8:53:39 AM permalink
Quote: coolhandnuke

My question, in my mind is, you can't have SIMULATED data to find a systems REAL TRUE outcome on a real physical mechanical roulette wheel device combined with human interaction ball spinning dealer as you are not getting data from a GENUINE SOURCE.



That turns out to be wrong. Gaming regulatory agencies, the ones that test and approve new games, sometimes do so using computer-based random number generators and other software. They don't require "REAL TRUE" outcomes dealt from a deck of cards or spun on a slot machine. In most cases games can be analyzed and solved directly, but software-based simulation is accepted when required. And there's no fundamental difference between a slot machine reel, a roulette wheel, or a deck of cards when it comes to random simulation. The RNG can be the same in all cases.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
coolhandnuke
coolhandnuke
Joined: Sep 18, 2010
  • Threads: 1
  • Posts: 6
October 27th, 2010 at 9:00:34 AM permalink
Quote: MichaelBluejay

You're right, it would be best to use real-world casino data. But that's impossible. So we make do with what we have. And it's really only a problem for people who think that casino results or a computer sim is less random than the other -- of which you seem to be one. But there's no way I can placate you, so sorry.



There's no need to be sorry MBJ, I wasn't going to hand over to you my beautiful system for $1000, though I would reconsider if you offered me $5,000,000
winning at roulette is easy, knowing when to cash in and leave is not so easy
Ayecarumba
Ayecarumba
Joined: Nov 17, 2009
  • Threads: 232
  • Posts: 6592
October 27th, 2010 at 11:52:11 AM permalink
Quote: MichaelBluejay



Bluejay's $1000 challenge!

I'll pay $1000 to the first person who comes up with a betting system that can beat my proposed rules #1-3 in my original post.



Regarding the, "for systems that involve "sitting out" some rounds in order to wait for presence or absence of streaks, we'll simulate additional lifetimes proportional to the number of rounds that were sat out."

How exactly would that work, if for example my "magic" craps system requires waiting for three consecutive come out seven's, then betting it all on Don't Pass. If win, Stop. If lose, have to stop.

1 - Would the additional "lifetime" requriement be one lifetime for every 360,000 (or portion thereof) rolls that were "sat-out"?

2 - Is this determined before (based on a projection of the average number of "sit-outs") or after, using the actual sit-outs encountered in the initial 1k lifetime run?

3 - if using the actual 1k lifetime run sit-out count to determine additional lifetime requirement, do the sit-outs required for the additional lifetime(s) result in even more lifetimes, and so on, and so on?
Simplicity is the ultimate sophistication - Leonardo da Vinci

  • Jump to: