Notices
Non Scooby Related Anything Non-Scooby related

Any probability experts here?

Thread Tools
 
Search this Thread
 
Old 08 January 2004, 02:31 PM
  #1  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

Hi,

I've recently been playing roulette online using a demo account. The wheel they use is a 'French' wheel meaning it only has one zero on it, which slightly improves the odds.

Anyway, I've been playing the 'Martingale' system and have done pretty well from it, but I realise how it can go wrong. For those that don't know the system, this is it:

You put $1 chip down on red, for example, then spin the wheel. If red comes up you get back $2 meaning a profit of $1. Then you do the same again and so on. If however, black comes up then you lose your chip. So this time you put down $2 on red - if red comes up you get back $4 giving a profit on that spin of $2 and clearing the loss on the black that came up. If black comes up again you put down $4 looking to make $8, and so on. Basically you continue doubling up as the spins go against you until red comes up. The sequence is $1, $2, $4, $8, $16, $32, $64, $128

This leads me to the problem. The table limit per spot is $100. Therefore, as soon as you get the 'wrong' colour 7 times in a row you're screwed and you've lost $127 (1+2+4+8+16+32+64).

So, to you experts out there - how often is this situation likely to occur? Seven of the same colour in a row? (don't forget the zero too!) Is it rare enough to just continue through it?

I made $470 the other day (yes, it took ages ) - too good to be true? Almost certainly!
Old 08 January 2004, 02:43 PM
  #2  
matt85
Scooby Regular
 
matt85's Avatar
 
Join Date: Nov 2003
Posts: 527
Likes: 0
Received 0 Likes on 0 Posts
Post

im not an expert and this is off the top of my head but would it be somnething like this... .47x.47x.47x.47x.47x.47x.47 that leaves 0.005ish so i dont know if thats right to be honest.

(roughly .47 as the 0 would be roughly 3% chance of occuring)
Old 08 January 2004, 02:45 PM
  #3  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

I should have said more about the wheel. There are 37 spaces on the wheel - 18 red, 18 black and one zero.

Old 08 January 2004, 03:54 PM
  #4  
marky1
Scooby Regular
 
marky1's Avatar
 
Join Date: Sep 2003
Posts: 549
Likes: 0
Received 0 Likes on 0 Posts
Post

Open another account so you can keep going on another name.

Just be careful, if you get 20 in a row against you, your next stake needs to be $520,000. It's called averaging! Unlikely to happen, but certainly possible

M
Old 08 January 2004, 04:03 PM
  #5  
milo
Scooby Regular
 
milo's Avatar
 
Join Date: Nov 2001
Posts: 2,043
Likes: 0
Received 0 Likes on 0 Posts
Post

0.0094 is the prob. that you'll get 7 against you in a row... so less than 1% chance.

however, play it enough times...
Old 08 January 2004, 04:05 PM
  #6  
ajm
Scooby Regular
 
ajm's Avatar
 
Join Date: Sep 2002
Location: The biosphere
Posts: 7,824
Likes: 0
Received 0 Likes on 0 Posts
Post

I am assuming the zero is neither black nor red?


if so:

chances of 7 reds in a row = (18/37)^7 = 0.006449046

7 blacks is the same

chances of 7 zero's in a row = (1/37)^7 = 1.05339E-11
Old 08 January 2004, 04:07 PM
  #7  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

Milo,

Yes, eventually it happens. But the thing is, if you've taken in more than $127 in profits then you can just keep going. Do you see what I mean? If it only happens every so often then overall you're going to profit from it.
Old 08 January 2004, 04:11 PM
  #8  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

ajm,

Yes, the zero is neither black nor red and you lose if the ball lands on it.

How do you express your results in a percentage form? As in, during the course of 1,000 spins how likely is it that you'll get seven of the same colour, or seven of the same colour including the zero, in a row?
Old 08 January 2004, 04:16 PM
  #9  
milo
Scooby Regular
 
milo's Avatar
 
Join Date: Nov 2001
Posts: 2,043
Likes: 0
Received 0 Likes on 0 Posts
Post

yes - i know the system well

with your odds tho, you stand to only gain $1 per run of bets. so if u think about it on a smaller scale, u will be making $32 and $64 bets to win a slightly less than 50% chance of getting $1.

on the grand scheme of things, u could turn a profit on this. bear in mind that u need to have 7 in a row NOT coming up a minimum of 127 times (it will be a lot more in reality, because you're not going to win first time every time - statistically you're probably hoping the 7 in a row does NOT come up several hundred times or even thousand times). given that there is a near 1% chance of it coming up, it may well come up in a few hundreds "turns" (where a turn is a run of bets in which u hope to turn the $1 profit).
Old 08 January 2004, 04:16 PM
  #10  
ajm
Scooby Regular
 
ajm's Avatar
 
Join Date: Sep 2002
Location: The biosphere
Posts: 7,824
Likes: 0
Received 0 Likes on 0 Posts
Post

multiply by 100 to get percentage

If you include the zero as a loss then the results are slightly different.

The chances of it going against you would be (19/37)^7 = 0.009415928 (oops just noticed someone already worked this out above!)


Old 08 January 2004, 04:27 PM
  #11  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

OK, then that answers it. If the chances of it happening are 0.942% it's going to happen more often than taking in the required $127 to cover it through successful bets.

However, you can spin the wheel without placing a bet. Therefore, you could wait until the wheel has recorded, let's say, 3 of the same colour in a row before placing a chip down. By doing that you'd be saying that the wheel has to go against me 10 times in a row for me to lose.

Is that false logic? Every spin is completely independent of the last.
Old 08 January 2004, 04:35 PM
  #12  
dsmith
Scooby Regular
 
dsmith's Avatar
 
Join Date: Mar 1999
Posts: 4,518
Likes: 0
Received 0 Likes on 0 Posts
Post

Every spin is completely independent of the last.
Are you sure about that in an on-line simulation ?

Deano

Old 08 January 2004, 04:39 PM
  #13  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

Are you sure about that in an on-line simulation ?
It's exactly the same software as the real game you'd be playing with them if you were using money. It all seems pretty well regulated, and apparently the games are audited by Price Waterhouse Coopers.



Old 08 January 2004, 04:39 PM
  #14  
LG John
Scooby Regular
 
LG John's Avatar
 
Join Date: Mar 2002
Location: Bradford
Posts: 13,720
Likes: 0
Received 0 Likes on 0 Posts
Post

Ok, I researched this quite extensively when I tried to modify the system a little. Basically I concluded that you should try for runs in your favour. So if red was your b!tch for the night you'd play it and if you landed you'd keep the £2 on and try to double it up. Do this until you've got a tenner (or however big your ***** are) and then bank all or half of it as profit. If your run brakes before you bank you double up to recover your initial investment and start over. This worked really well and I tried it on an online table that had a play for fun feature.

However, like you I became worried about bad runs and sure enough they came along all too often and very often much much more than 7 in a row I also tested this probability with a 10p coin and flipped it over and over and over recording 'runs' This isn't even a proper reflection as there is no 'zero' on a coin. ***Before anyone starts when I flip a coin its spins like **** and I allowed it to bounce on the carpet. Chaos theory dictates randomness from here*** The result??? You guessed it, monster (and I mean MONSTER) runs against me of upwards of 15 in a row. Flip a coin long enough and they come

I gave up after that without putting real money into it. If I was you I'd cash up now that you are ahead because probability dictates that if you play it long enough you'll get done Of course if you want to play with 1% of luck or so against you do it online where you can get 10-20 times the spin rate of real life and thus if you go bust you'll have not made a massive time committment to it
Old 08 January 2004, 04:41 PM
  #15  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

Saxo Boy,

I've walked away from it already. I was just curious about the maths behind it.
Old 08 January 2004, 04:42 PM
  #16  
LG John
Scooby Regular
 
LG John's Avatar
 
Join Date: Mar 2002
Location: Bradford
Posts: 13,720
Likes: 0
Received 0 Likes on 0 Posts
Post

Is that false logic? Every spin is completely independent of the last.
In a word, yes! There is no reason why it couldn't have 100 or 1000 black in a row. It not very likely but it could just keep going and going and going. You'd probably find letting 3 of one colour come it would give you emotional security but it really doesn't change much. The probability on each individual spin remains fixed ....or should that be
Old 08 January 2004, 05:35 PM
  #17  
Scooby_simon
Scooby Regular
 
Scooby_simon's Avatar
 
Join Date: Nov 2003
Location: Hitchin
Posts: 102
Likes: 0
Received 0 Likes on 0 Posts
Post

You will always come out worse in the end.....

18 red
18 black
1 green

Prob of red or black is 18/37 (ie 0.486486) on every go....

so you will on average get back around 97% of your stake over time....

Don't play it to make money, you will not do it. Play for fun or not at all.....

Old 08 January 2004, 05:35 PM
  #18  
Dan B
Scooby Regular
 
Dan B's Avatar
 
Join Date: Feb 2002
Posts: 1,377
Likes: 0
Received 0 Likes on 0 Posts
Thumbs up

Technical term: "Discrete Random Variable."

Oh, and if you go to a real casino and try this, they will give you a swift kick in the ***** and throw you out the door.
Old 08 January 2004, 07:05 PM
  #19  
Scumbag
Scooby Regular
 
Scumbag's Avatar
 
Join Date: Aug 2001
Posts: 429
Likes: 0
Received 0 Likes on 0 Posts
Post

Thinking that what has gone before has the slightest influence on the next colour out is known as the gamblers fallacy...it is plain wrong.

IIRC, if you cant card count at blacjack(and very few people can do it properly, and even then you can be spotted, and asked to leave), the best chance of winning money is to take everything you can afford to lose, and plce it on black or red once, then walk out. Still likely to lose, mind!!!
Old 08 January 2004, 08:22 PM
  #20  
matt85
Scooby Regular
 
matt85's Avatar
 
Join Date: Nov 2003
Posts: 527
Likes: 0
Received 0 Likes on 0 Posts
Post

whats 'card counting' if i may ask?
Old 08 January 2004, 08:38 PM
  #21  
carl
Scooby Regular
 
carl's Avatar
 
Join Date: May 1999
Posts: 7,901
Likes: 0
Received 0 Likes on 0 Posts
Post

It's exactly the same software as the real game you'd be playing with them if you were using money
Surely if it was a real game there would be no software -- just a bloke spinning a roulette wheel?
Old 08 January 2004, 08:42 PM
  #22  
cletterridge
Scooby Regular
 
cletterridge's Avatar
 
Join Date: Nov 2001
Posts: 246
Likes: 0
Received 0 Likes on 0 Posts
Post

This reminds me of the old 2:1 horse trick that me and my friends used to talk about (but never tested, so I'm not offering advice here!). Actually, since I'm not into the whole betting thing, I'm not sure if it's 2:1 or 3:1, whichever is the one that gives you three times your money on a win. I'll call it 2:1 here....

Anyway, the idea is you start by betting £1 on a 2:1 horse.

It loses. You lose your £1.

So, on the next 2:1 horse you see, you bet £2.

It loses. You lose that £2 too.

On the next step you bet £4, then £8, then £16 and as many times as you have to, doubling your money each time, until a horse finally comes in - let's say it doesn't come in until the eighth horse on which you've bet £128. Here's how it adds up:

You bet £128, you get £128 x 3 = £384.

BUT you've only lost (on all the other horses) £1 + 2 + 4 + 8 + 16 + 32 + 64 + 128 = £255.

Leaving you with a profit of £129.

Mathematically speaking, so long as you continue to double your outlay each time, you always recoup more than you have bet at this odds ratio. The main problem is that if you get up to, say, 25 horses and none have won yet, you're next bet is about £33 million pounds!

Spend it on beer instead!
Old 08 January 2004, 08:42 PM
  #23  
Butkus
Scooby Regular
Thread Starter
 
Butkus's Avatar
 
Join Date: Feb 2003
Posts: 1,120
Likes: 0
Received 0 Likes on 0 Posts
Post

Surely if it was a real game there would be no software -- just a bloke spinning a roulette wheel?
From the Casino-on-Net website...

Fair Play

One of the main concerns of Online gaming operations is to guarantee fair gaming. Casino-on-Net understands the importance of explaining how this is actually done:

The Casino-on-Net "dealer" is really a computer. A Random Number Generator (RNG) is utilized to ensure that performance is truly random. Casino-on-Net uses the trusted MD5 RNG, which provides absolutely random results. The system has been rigorously tested by running millions of rounds and examining results. The system is also systematically tested by Casino-on-Net.

Casino-on-Net provides our Members with an easy to use Online log of their game results. This information may be reviewed any time, by pressing the "My Game Log" button, and includes amounts wagered, game results, as well as Deposit and Cash Out records.

Casino-on-Net has retained the services of PricewaterhouseCoopers Inc., an internationally respected and distinguished accounting firm, in order to review the payout percentage.

The review is conducted on a monthly basis, and includes the Payout Percentage of each Casino-on-Net game, as well as the total average Payout Percentage of Casino-on-Net.





[Edited by Butkus - 1/8/2004 8:45:00 PM]
Old 08 January 2004, 09:43 PM
  #24  
carl
Scooby Regular
 
carl's Avatar
 
Join Date: May 1999
Posts: 7,901
Likes: 0
Received 0 Likes on 0 Posts
Post

Casino-on-Net uses the trusted MD5 RNG, which provides absolutely random results.
I don't know the details of the MD5 algorithm, but it's clearly just a PRNG (pseudo-random number generator). Computers cannot generate random numbers. It could be truly random if it was linked to, say, an alpha radioactive decay source. Otherwise it's just a sequence of numbers that repeats after a large number of iterations.
Old 08 January 2004, 10:04 PM
  #25  
ajm
Scooby Regular
 
ajm's Avatar
 
Join Date: Sep 2002
Location: The biosphere
Posts: 7,824
Likes: 0
Received 0 Likes on 0 Posts
Post

MD5 is used in encryption, such as in PGP etc.
Old 08 January 2004, 10:09 PM
  #26  
Scumbag
Scooby Regular
 
Scumbag's Avatar
 
Join Date: Aug 2001
Posts: 429
Likes: 0
Received 0 Likes on 0 Posts
Post

You can beat blackjack by counting at the number of picture cards left in the deck compared to those that have been already dealt. You bet differently depending on where the majority of picture cards lie (high bets one way, low the other). This gives you the mathematical advantage BUT it is VERY difficult to do (bearing in mind there are multiple decks in the 'shoe'). Lots of people try it, but as everything moves so fast, if you get slightly out, you are betting the wrong way round and giving the house a massive advantage.
If a casino spots anyone doing this (and winning), they will throw you out on your ear. Dealers say it is very easy to spot.

Various machines have been built to try and give gamblers the edge in this counting business, so they know whether to bet high or low...these I think are actually illegal (fraud, I guess)

If you do a google on "card counting" you'll see lots and lots of books advertising the "easyiest" way to do it...if it was so easy they wouldnt have written books 'cus they'd be millionaires by now!
Old 08 January 2004, 10:13 PM
  #27  
ProperCharlie
Scooby Regular
 
ProperCharlie's Avatar
 
Join Date: Oct 2004
Location: London
Posts: 4,797
Likes: 0
Received 0 Likes on 0 Posts
Post

i thought the practice of putting several shuffled packs of card into the big holder and then drawing off a block from the top of this, to deal with during the game, prevented the practice of card counting as you cannot know how many picture cards are in the block of cards being used. and then after the game they draw a differnt block, so you can't use information gained from counting the cards in the previous game?

[Edited by ProperCharlie - 1/8/2004 10:14:58 PM]
Old 09 January 2004, 11:54 AM
  #28  
Dan B
Scooby Regular
 
Dan B's Avatar
 
Join Date: Feb 2002
Posts: 1,377
Likes: 0
Received 0 Likes on 0 Posts
Post

It can be a bit more simple than that: Lets say you are at a table with 5 other players, that means (in BJ) that one hand will use at least 12 cards, all of which every player gets to see. If a "10" hasn't come out for all those 12 cards you can be sure one is about to come out, as the odds of a 10 coming out is 4/13. Its not a "sure thing", but you can make an educated guess...
Old 09 January 2004, 12:46 PM
  #29  
Old_Fart
Scooby Regular
 
Old_Fart's Avatar
 
Join Date: Feb 2003
Posts: 1,717
Likes: 0
Received 0 Likes on 0 Posts
Post

Cleterridge, the horse racing scheme doesn't work because the 2for1 odds aren't going to be the correct odds in all cases..ie in 'fair' roulette (no greens) you are exactly 50/50 to win betting on a colour. If you play one millions times, always betting on one colour using the doubling technique when you lose you will almost certainly break even..assuming the payout is 2 for 1. On the horses you can't cant on that mathmatical certainty..many horses rated 2 for 1 should in reality be longer odds, and ultimately bets made on them will reflect that with their payoff.
Roulette is a suckers game...statistically the best you can hope for is just shy of breaking even(ie being a loser long term)...unless the wheel isn't exactly fair and you know what number/s is/are favoured.
Rgds
Cman
Related Topics
Thread
Thread Starter
Forum
Replies
Last Post
JimBowen
ICE
5
02 July 2023 01:54 PM
InTurbo
Other Marques
20
08 October 2015 08:59 PM
bluebullet29
General Technical
9
05 October 2015 02:17 PM
Lillyart14
ScoobyNet General
24
01 October 2015 01:29 AM
InTurbo
ScoobyNet General
21
30 September 2015 08:59 PM



Quick Reply: Any probability experts here?



All times are GMT +1. The time now is 11:09 AM.