In my last post, i showed you how to calculate the true odds for the bitcoin game satoshi mines
If you want to know how the game works, see this post. Or if you want to play the game go here
In that post, i showed you the true odds of the game (25 squares and 3 bombs) and in my previous post, i showed you the payout odds (for 25 squares and 3 bombs)
I thought it might be worth comparing them.... The table below shows true odds vs actual odds (for 25 squares 3 bombs)
I think the interesting thing here, is the cumulative edge isn't a constant. They offer a very low edge for your first bet and then the edge gets progressively worse until your 8th guess when the edge starts reducing. I don't have their payouts beyond guess 16. I might create a robot to get those numbers at some point. Let's look at the edge changes as a graph.
So what does this tell me.......
It says to me you really shouldn't play beyond your first guess... for 25 squares 3 bombs
As the cumulative edge (unto guess 17 doesn't get better than that first edge on that first guess..
You'd be better to cash out your money, and then just play another round. You'll have lower odds sure (as you'll always been playing with 22/25 probability) but you'll be playing with a lower edge, so you'll lose money over the long run.
It also tells me there maybe more value in the other variants of the game and we'll explore that in another post.
In true bitcoin tradition, if you found this article useful and want to donate me some spare bitcoins my wallet address is: 17QCjNaaXUWyG5WWaV3KVDQG7HdVkSDLs1
In the next article we'll look at some of the grids and options you have
Sunday, 11 January 2015
Calculating the true probability of Satoshi Mines (25 squares, 3 bombs)
In my last post, i explained how to play a bitcoin game called satoshi mines
In this post, i'm gonna explain how you can calculate the probabilities of the game. I'll start with 25 square, 3 bomb grid.
If you need a refresher on calculating probabilities, i'll do one here but you can always refer to this post.
If you want to play the game, you can play satoshi mines here
In this post, i'm gonna explain how you can calculate the probabilities of the game. I'll start with 25 square, 3 bomb grid.
If you need a refresher on calculating probabilities, i'll do one here but you can always refer to this post.
If you want to play the game, you can play satoshi mines here
Calculating the probability
It's pretty simple,. There are 25 squares, 3 of which have bombs.
First guess
That means the probability of me NOT HITTING a square on my first guess is....
22/25 or 88.50% (or decimal odds of (1.136)
Yes there are 22 squares out of 25 squares that are empty. So I have an 88.5% (22/25) chance of not hitting a bomb in my first attempt.
If I had an original stake of 100 bits, i'd need to win over 13.6 bits for me to be able to beat the game, as you could see from my previous post. The actual payout is 13 bits. So the house has an edge of about 0.5% (which is actually pretty low for a game)
Second guess
So it's a little harder for me now. Now there are 21 squares (i clicked a square remember) that are empty and 3 squares which have bombs. So I a smaller chance of hitting an empty square. So the probability of me not hitting a square in my second attempt is.....
21/24 or 87.5% (or decimal odds of 1.143 roughly)
So now that I am staking 113 bits (initial stake of 100 plus my 13 bits from last win), i really need to win around 16 bits for this to be fair to bring my cumulative wins to 129 bits. Well you've probably figured that isn't gonna happen. The win is 15 bits bringing my stake / win unto 128 bits
So how is cumulative probability calculated.
Basically for each guess, it is
probability of current guess * probability of previous guesses
so for 2 guesses, it'd be
(22/25) * (21/24)
for 3 guesses, it's be
(22/25) * (21/24) * (20/23)
etc etc
Anyways, hopefully now when you're playing Satoshi mines, you'll be able to calculate the true probability. i've put a handy chart below for you to see the true probabilities of the
25 square 3 bombs game true probabilities
In the next posts, i'll look at the probabilities of the various combinations And look at giving you, your best strategy.
In true bitcoin tradition, if you found this article useful and want to donate me some spare bitcoins my wallet address is: 17QCjNaaXUWyG5WWaV3KVDQG7HdVkSDLs1
In true bitcoin tradition, if you found this article useful and want to donate me some spare bitcoins my wallet address is: 17QCjNaaXUWyG5WWaV3KVDQG7HdVkSDLs1
Hope this helps
How does Satoshi Mines work?
So I've spent some of time over the last few days looking at some of the bitcoin gambling sites. Wow there is a lot to talk about in this subject. Anyways I found myself playing a fun little game called Satoshi Mines, it's kinda like Minesweeper.
It's Bitcoin only so if you want to gamble, you will need a bitcoin wallet (a subject for another day) or you can play in a practise mode. I was playing in practise mode and recommend you do the same because I'll prove in my next post, that you can't really win.
How does the game work?
You start with a square (default is 25 squares) with a hidden number of bombs (default is 3). You can change the number of bombs if you want (and i recommend you do). Here is a screenshot.
You bet an initial stake (let's say 100 bits) and everytime you click on a square and miss a bomb, you win some bits. You can either cash out your bits or risk those bits you just won against the chance of wining some more bits. Obviously if you hit a bomb, you lose all the bits won (and your original stake).
What's the payouts?
I played the game for a little bit and captured the following payouts for a 25 square game with 3 bombs
As you can see the more squares you uncover as empty, your cumulative odds get better and you can win a good multiplier of your original stake. As you can see above, the best I did was to manage 16 squares where i could win 23 times my stakes (if I won and cashed out).
So in the above example.
Okay, looks fun, can I win? In the short term, sure. In the long term, no. Why? Probability! And that's what i'll explain in my next post
However the game does have some pretty low house edges (as i'll explain in my next post)
Warning: This is a not a casino, not licensed and you're playing with Bitcoins. I really do recommend you only play in practise mode.
Hope this helps
It's Bitcoin only so if you want to gamble, you will need a bitcoin wallet (a subject for another day) or you can play in a practise mode. I was playing in practise mode and recommend you do the same because I'll prove in my next post, that you can't really win.
How does the game work?
You start with a square (default is 25 squares) with a hidden number of bombs (default is 3). You can change the number of bombs if you want (and i recommend you do). Here is a screenshot.
You bet an initial stake (let's say 100 bits) and everytime you click on a square and miss a bomb, you win some bits. You can either cash out your bits or risk those bits you just won against the chance of wining some more bits. Obviously if you hit a bomb, you lose all the bits won (and your original stake).
What's the payouts?
I played the game for a little bit and captured the following payouts for a 25 square game with 3 bombs
As you can see the more squares you uncover as empty, your cumulative odds get better and you can win a good multiplier of your original stake. As you can see above, the best I did was to manage 16 squares where i could win 23 times my stakes (if I won and cashed out).
So in the above example.
- My initial stake was 100 bits
- If i hit a mine, i lose all 100 bits.
- If I've guessed 10 times (i.e. picked 10 squares that don't have bombs in them), i'll have 466 bits (366 bits won plus my original 100 bits)
- If i want to guess another square (11th guess), i'll either win another 111 bits, making a total of 577 bits or i'll lose everything including my initial 100 bits stake. Or I can cash out and take my 466 bits as winnings.
Okay, looks fun, can I win? In the short term, sure. In the long term, no. Why? Probability! And that's what i'll explain in my next post
However the game does have some pretty low house edges (as i'll explain in my next post)
Warning: This is a not a casino, not licensed and you're playing with Bitcoins. I really do recommend you only play in practise mode.
Hope this helps
Friday, 9 January 2015
Roulette excel isn't the same as mechanical wheels
I've spent a lot of time so far giving you the tools on how to calculate probabilities and check if staking plans work in excel.
The game of roulette isn't played in Excel, it's played on a mechanical wheel (unless you are playing online in a non-live roulette game and then you may as well be playing in Excel).
So what's my point:
If your number based system (martingale, reverse martingale etc) doesn't hold up in Excel, it won't hold up in the real world either.
Ahhhhh, but you said in the real world, it's mechanical wheels, so there are wheel biases, dealer signatures etc, so maybe my system does hold up in the real world and not in Excel?
sure maybe, but that's blind luck. does every dealer have the signature, every wheel have the same bias? if that's what you think then you're crazy.
If you want to look at real world advantage play then we'll get to that but if your number based system doesn't hold up in excel it won't hold up with mechanical wheels. if you want something to hold up with mechanical wheels then you need a different strategy. and we'll get to that.
hope this helps
The game of roulette isn't played in Excel, it's played on a mechanical wheel (unless you are playing online in a non-live roulette game and then you may as well be playing in Excel).
So what's my point:
If your number based system (martingale, reverse martingale etc) doesn't hold up in Excel, it won't hold up in the real world either.
Ahhhhh, but you said in the real world, it's mechanical wheels, so there are wheel biases, dealer signatures etc, so maybe my system does hold up in the real world and not in Excel?
sure maybe, but that's blind luck. does every dealer have the signature, every wheel have the same bias? if that's what you think then you're crazy.
If you want to look at real world advantage play then we'll get to that but if your number based system doesn't hold up in excel it won't hold up with mechanical wheels. if you want something to hold up with mechanical wheels then you need a different strategy. and we'll get to that.
hope this helps
Labels:
Excel,
Martingale,
Randomness,
Roulette,
Staking Plans
Martingale stake formula
I've spent a few posts now showing why employing Martingale against fair odds (or in the case of red or black in roulette less than fair odds), that you're always gonna lose. If you want to read a bit more about it read my post on martingale in roulette begin very bad.
If you've forgotten what martingale is, it basically means everytime you lose, double up your previous bet
Calculating your martingale stake
In one of my earlier posts, i showed you this table, where for a losing streak, how much you'll have to bet to recover
I thought it might be useful to provide a formula that calculate the stake required to cover your losses during a martingale
And that is S = I * (2^(n+1))
I = Initial Stake
n = length of streak
S = Stake Required
Which means Initial Stake * (2 to the Power of the current length of the streak)
So in the scenario where our initial stake is £10 and we've lost 10 times and we need to make our eleventh bet. Our new stake would be:
£10 * (2 to the power of 10) = £10 * 1024 = £10,240
Obviously you'd be better throwing your money down the toilet but there ya go. But if you want to calculate it, that's how you do it.
In the cell B4 (stake column row 4) just enter
Obviously you can also calculate it in excel, you can just double up the previous stake. but the point is to show you how to do when you just know the initial stake and the bet number.
Please don't do martingale with roulette kids, not worth it.
If you've forgotten what martingale is, it basically means everytime you lose, double up your previous bet
Calculating your martingale stake
In one of my earlier posts, i showed you this table, where for a losing streak, how much you'll have to bet to recover
I thought it might be useful to provide a formula that calculate the stake required to cover your losses during a martingale
And that is S = I * (2^(n+1))
I = Initial Stake
n = length of streak
S = Stake Required
Which means Initial Stake * (2 to the Power of the current length of the streak)
So in the scenario where our initial stake is £10 and we've lost 10 times and we need to make our eleventh bet. Our new stake would be:
£10 * (2 to the power of 10) = £10 * 1024 = £10,240
Obviously you'd be better throwing your money down the toilet but there ya go. But if you want to calculate it, that's how you do it.
And if you wanna do it in Excel
To do this in Excel with a table like the following
=POWER(2,A3-1)*$B$3
Please don't do martingale with roulette kids, not worth it.
High-Odd-Black vs High-Odd-Red in Roulette
So I read this interesting article at
http://www.roulette30.com/2013/08/an-observation-on-even-chances.html
Which basically says you have a higher probability of losing if you bet High-Odd-Black than you do if you bet High-Odd-Red.
So the idea is that if I place a bet on High Numbers, Odd Numbers and Red Colors then there is only 5 numbers that can make me lose:
2, 4, 8, 6 and 10
But if I bet on High, Odd and Black then there is theoretically only 4 numbers that can make me lose
12, 14, 16 and 18
And then the article did some math wizardry showed that with each bet being evens then the "probability" of losing all your bets would be 1/8 (1/2 * 1/2 * 1/2). BTW, the article ignores zeroes. So i will too.
However, he points out if we were to bet on High-Odd-Black the probability of losing all your bets is
4/36 (ignoring zeroes) which simplifies to 1/9 but..... if we bet on High-Odd-Red the probability of losing all your bets is... 5/36 or 1/7.2
That's not the bet. The bet is 3 separate bets whose results are paid out in isolation This is not a trixie bet, this is 3 individual bets. And sometimes you will win more than once for a certain number. What we really should be looking is how many times can I win across these bets.
Well, here we go....
Oh, look, we have 54 possible wins. Shall we see how that breaks down
And we still have 54 possible wins. And shall we see how that breaks down also
It just means, that if we bet High-Odd-Black then sure we have a greater coverage of winning numbers than High-Odd-Red but we'll win less.
It all evens out. We're betting individual bets. And as shown they have a different distribution. But at the end of the day. There are only 54 possible winning bets..... regardless of which way you bet
And just in-case you wanna check out the math. here is a spreadsheet with my breakdown.
https://www.dropbox.com/s/fwcgk8eb0171lus/Kavouras%20HighOddBlack.xlsx?dl=0
Hope this helps
http://www.roulette30.com/2013/08/an-observation-on-even-chances.html
Which basically says you have a higher probability of losing if you bet High-Odd-Black than you do if you bet High-Odd-Red.
So the idea is that if I place a bet on High Numbers, Odd Numbers and Red Colors then there is only 5 numbers that can make me lose:
2, 4, 8, 6 and 10
But if I bet on High, Odd and Black then there is theoretically only 4 numbers that can make me lose
12, 14, 16 and 18
And then the article did some math wizardry showed that with each bet being evens then the "probability" of losing all your bets would be 1/8 (1/2 * 1/2 * 1/2). BTW, the article ignores zeroes. So i will too.
However, he points out if we were to bet on High-Odd-Black the probability of losing all your bets is
4/36 (ignoring zeroes) which simplifies to 1/9 but..... if we bet on High-Odd-Red the probability of losing all your bets is... 5/36 or 1/7.2
So should I bet on High-Odd-Black instead of High-Odd-Red?
Errrr, no...... Because you don't win bets on how many unique numbers lose. These are not combinator bets. If I was offered odds of 1/8 for all my bets to lose then sure this is a winning strategy and I'd say take the house to the cleaners. As 1/9 will yield profit on 1/8 you'd have your own edge.That's not the bet. The bet is 3 separate bets whose results are paid out in isolation This is not a trixie bet, this is 3 individual bets. And sometimes you will win more than once for a certain number. What we really should be looking is how many times can I win across these bets.
Well, here we go....
High-Odd-Black
Oh, look, we have 54 possible wins. Shall we see how that breaks down
- 14 numbers offer as a single win (only one bet wins (High, Odd or Black)
- 14 numbers offer as a win on two bets (High and Odd, High and Black or Odd and Black)
- 4 numbers off a win on all threes bets (High, Odd and Black)
High-Odd-Red
- 13 numbers offer as a single win (only one bet wins (High, Odd or Red)
- 13 numbers offer as a win on two bets (High and Odd, High and Black or Odd and Red)
- 5 numbers off a win on all threes bets (High, Odd and Red)
What does this mean?
Don't get me wrong, it's an interesting observation but it's not gonna help you win more. the odds of winning with High-Odd-Black are still the same as High-Odd-Red (But i applaud the observation skills)It just means, that if we bet High-Odd-Black then sure we have a greater coverage of winning numbers than High-Odd-Red but we'll win less.
It all evens out. We're betting individual bets. And as shown they have a different distribution. But at the end of the day. There are only 54 possible winning bets..... regardless of which way you bet
And just in-case you wanna check out the math. here is a spreadsheet with my breakdown.
https://www.dropbox.com/s/fwcgk8eb0171lus/Kavouras%20HighOddBlack.xlsx?dl=0
Hope this helps
Thursday, 8 January 2015
Hot numbers in roulette
I'm not gonna name the site but I was looking at yet another person peddling their roulette strategy based on tracking numbers (no genuine advantage play).
Basically, you stick in the last 6 numbers and voila, it will tell you where to bet. If the wheel spins 3, 4, 5, 14, 15 and 17. You should probably bet on 19-36, cause those numbers are due....
Just for fun, let's see how that looks in the Excel Simulator
The Random Numbers
This is what my Excel Random Number generator produced (feel free to replicate this experiment yourself)
3, 14, 9, 9, 3, 21, 35, 3, 29, 0, 19, 36, 4, 9, 5
The distribution
Oh look the number 3 came up 3 times (surely that's a hot number)???? BTW, that was sarcasm. We know enough about random numbers now to know that that can easily happen. However this could be an interesting blog post in the future. Hot numbers.
What might be more interesting is the ranges, which is what they were betting on in this system.
Their system was either avoid betting on 1 to 12 because it's hot. but it might have been bet on 1 to 12 because it's hot.
There is no hot.... It's just some random numbers. nothing is due or not due....
The kicker
And here was the kicker. Don't worry, if you lose, just martingale and it'll recover your losses.
And we know how that ends...
Anyways, you know my thought on this sort of thing already. I could type up their system into Excel and formally disprove it. But you can always do that yourself if you want.
We know enough about random numbers, streaky numbers and martingale to know this is a silly strategy.
For more info on streaks, read my post on streaky coin tosses
For generating roulette numbers in excel, read my post (or look at the martingale one)
Why playing Red or Black in Roulette using Martingale is a bad idea
I've been leading unto this post for a bit now.
One of the most popular techniques in roulette is to use a progressive betting pattern such as
One of the most popular techniques in roulette is to use a progressive betting pattern such as
- Martingale
- Reverse Martingale
- d'Alembert
- etc
In all honesty, these are all a pile of rubbish. If the math was corrupt, do you really think the casino's would let you play the game? So the idea of this post is to let you see how much of a pile of rubbish they are.
So what is Martingale
The basic idea is that everytime you lose a bet, you double up your previous bet until you recoup your losses. Here is an example with a £10 stake and an initial bank of £10,000
Oh, wow, how amazing. this must be free money. Except, if you hit a long streak of losses the next stake is gonna exponentially grow.
And this is how much you'll have to bet to recover your losses during a losing streak
Oh wow after a losing streak of 11 losses, I'm betting £10,000 to recover my money. Or 15 losses, I'm betting £163,000. Get's pretty uncomfortable pretty quickly.
What else, can go wrong? Well, even if you are stupid enough to try and bet £20,000 to recover your losses, the casino probably won't let you. They have table limits (i.e. a maximum bet size). And the chances are it's somewhere between £500 and £10,000.
Yeah but i'll never get a streak that long
What's the chances of hitting a streak of 10 losses in a row and busting the table limits. Errrr, pretty high. See my post on streaky coin tosses (more favourable than red or black on roulette)
As you can see from the above graph (which is 20,000 simulated coin tosses). You're gonna hit a streak of 10 or more, 18 times in 20,000 coin tosses. In roulette this will work out basically around 1 in every 1,000 games. That's pretty frequent.
Still don't believe me.
Try it yourself. I've created a martingale simulator in excel. Which uses my excel roulette simulator and an enhancement of my level staking plan simulators
The simulator allows you to simulate a large number of random roulette runs. You can put in whatever table limits you want (it handles it) and you can generate new runs.
And as you can see in the above run, i bust out after about 1,000 games. Sure i had a maximum bank of 3350 but i eventually bust out.
In-fact, you always bust out.
Some people will say, ahhh but you play it until you win and then go home. Sure but what happens the next night.....? It's a small amount of games until you bust out. You can't win.
Try it, play it in my simulator. You can't win
Download my martingale simulator
You can download it from my dropbox for free.
Play with it until your heart is content. (please note the simulator doesn't handle half-bet returns and doesn't need to the point is to show why martingale doesn't work)
You can even modify if you wan to try out your own wacky staking plans
So, next time you're tempted by playing martingale, just stop and think about it before you throw away your money
Hope this helps
Wednesday, 7 January 2015
Roulette Red or Black Level Staking Plan Simulator
In my previous post, I showed you how you can simulate roulette in Excel and track how much you'd win lose against a level staking plan (bet the same amount each time).
As you can see, i've add the following features to the simulator
The idea of this is to prove that over time, you can only lose in Roulette playing Red or Black. You can use the simulator to play 10, 100, 10,000, 40,000, in-fact as many games as you want.
You should see over time, that your losses will match the house edge. i.e. you will lose about 2.7% to the house.
I promised in that post that'd I'd enhance the version to allow you to summarise your bank. I have now done that. And it now looks like this
- Win Percentage
- Max Bet
- Minimum Bank
- Maximum Bank
- Final Bank
- Percentage of Bank
- Game Over (when you run out of money)
Download
You can now download the Roulette Level Staking Plan simulator from my dropbox
https://www.dropbox.com/s/d2508drcv73rn8z/Roulette%20Level%20Staking%20Plan%20Simulator.xlsx?dl=0
(please note the simulator doesn't handle half-bet returns)
(please note the simulator doesn't handle half-bet returns)
Hope this helps
How to track your wins, losses and bank in Excel for Roulette
In my last post, I showed you how to create a roulette simulator in excel.
In this post, I'll show you how to do a simple of track your wins, losses and bank for a level staking plan (bet the same flat amount every time). In a future post we'll make it a little more advanced.
The following screenshot shows you what we're looking to generate.
This allows us to see the wins, losses and bank for a Red/Black betting strategy against our roulette simulator (btw, betting red/black blindly is a good way of losing money, see my post on true odds of roulette).
1) The first 2 columns are generated by our roulette simulator, please see how to create a roulette simulator in excel to show you how to generate those two columns. Alternatively you can manually type in your own values (or copy and paste them in) if you want to track your bank against actual results instead of randomly generated ones.
2) The 3rd Column (Bet) - this is your prediction, just enter in Red or Black (i've bet red for every row). Enter Red in cell C3.
3) The 4th column (stake) - this is a level staking plan, e.g. bet $10 every time. Enter 10 in cell D3
4) The 5th column (win) - this tells you if your prediction won or not. type in the following formula in Cell E3
Hope you have fun playing with this. Feel free to drag a million rows :). See what happens to your bank. BTW, you will see that the casino always has the edge (should correlate with the true odds after a while)
UPDATE
I've now included a copy of the Roulette Level Staking Plan simulator for Excel in my dropbox here:
hope this helps
In this post, I'll show you how to do a simple of track your wins, losses and bank for a level staking plan (bet the same flat amount every time). In a future post we'll make it a little more advanced.
The following screenshot shows you what we're looking to generate.
This allows us to see the wins, losses and bank for a Red/Black betting strategy against our roulette simulator (btw, betting red/black blindly is a good way of losing money, see my post on true odds of roulette).
1) The first 2 columns are generated by our roulette simulator, please see how to create a roulette simulator in excel to show you how to generate those two columns. Alternatively you can manually type in your own values (or copy and paste them in) if you want to track your bank against actual results instead of randomly generated ones.
2) The 3rd Column (Bet) - this is your prediction, just enter in Red or Black (i've bet red for every row). Enter Red in cell C3.
3) The 4th column (stake) - this is a level staking plan, e.g. bet $10 every time. Enter 10 in cell D3
4) The 5th column (win) - this tells you if your prediction won or not. type in the following formula in Cell E3
=IF(C3=B3, "Win", "Loss")
B2 has the result of the random number generator, C3 has your prediction. (The letter is the column, the number is the row)
5) The 6th column (bank) - is your bank. In cell F2, enter your start bank (1000 in my case) And for cell F3, enter the following formula
=IF(E3="Win",F2+D3,F2-D3)
This basically means, if the result in E3 (Win column) is Win, then you add your winnings to your previous bank (D3 is stake, F2 is previous bank). if it's not a win, then minus your stake (D3) from your previous bank F2.
6) And now do the draggy copy with each row down the excel sheet, to do this for multiple rows. Excel will automatically update the formula cells as you drag and copy.
This is pretty simple. At the moment only works for red/black. but can be easily modified for other betting strategies.
In a future posts
- will show you how to use excel to track your current bank and final bank etc.
- will test stupid plans such as martingale and reverse martingale
UPDATE
I've now included a copy of the Roulette Level Staking Plan simulator for Excel in my dropbox here:
https://www.dropbox.com/s/d2508drcv73rn8z/Roulette%20Level%20Staking%20Plan%20Simulator.xlsx?dl=0
(please note the simulator doesn't handle half bet returns on zero and only does red or black. the point of the simulator is to disprove strategies hence why i haven't put time onto this yet).hope this helps
Simulating Roulette in Excel
In one of my earlier posts, i showed you how to simulate a coin toss in excel (or google docs)
In this article, we will look at extending that technique to allow us to simulate roulette. We're obviously just gonna randomly throw a roulette number and not simulate physics or mechanics (for now ;). The idea is that you will be able to use this simulator to test out any crazy betting systems you may come up with.
This technique will work for both European and American wheels. So let's get to it.
So to set it up, do the following
1) Open Excel
2) Rename the tab to "European Wheel"
3) In the first row enter the headers (as shown in the screenshot): Number, Actual Number Color
4) In column A enter in each row the numbers 0 to 36
5) Enter the corresponding number in the roulette wheel in order in Column B
6) Enter the color of the actual number in Column C
In this article, we will look at extending that technique to allow us to simulate roulette. We're obviously just gonna randomly throw a roulette number and not simulate physics or mechanics (for now ;). The idea is that you will be able to use this simulator to test out any crazy betting systems you may come up with.
This technique will work for both European and American wheels. So let's get to it.
Setting up our wheel
First we want to do is setup our roulette wheel in Excel, and it should look something like this...- First column is a list of numbers from 0 to 36 (37 if you want to do an American Wheel)
- Second column is the corresponding number on the roulette wheel (following actual layout)
- Third column is the color of the number
So to set it up, do the following
1) Open Excel
2) Rename the tab to "European Wheel"
3) In the first row enter the headers (as shown in the screenshot): Number, Actual Number Color
4) In column A enter in each row the numbers 0 to 36
5) Enter the corresponding number in the roulette wheel in order in Column B
6) Enter the color of the actual number in Column C
European Wheel
The following table gives the layout of a European Wheel, you can just copy and past it into excelNumber | Actual Number | Color |
0 | 0 | Green |
1 | 32 | Red |
2 | 15 | Black |
3 | 19 | Red |
4 | 4 | Black |
5 | 21 | Red |
6 | 2 | Black |
7 | 25 | Red |
8 | 17 | Black |
9 | 34 | Red |
10 | 6 | Black |
11 | 27 | Red |
12 | 13 | Black |
13 | 36 | Red |
14 | 11 | Black |
15 | 30 | Red |
16 | 8 | Black |
17 | 23 | Red |
18 | 10 | Black |
19 | 5 | Red |
20 | 24 | Black |
21 | 16 | Red |
22 | 33 | Black |
23 | 1 | Red |
24 | 20 | Black |
25 | 14 | Red |
26 | 31 | Black |
27 | 9 | Red |
28 | 22 | Black |
29 | 18 | Red |
30 | 29 | Black |
31 | 7 | Red |
32 | 28 | Black |
33 | 12 | Red |
34 | 35 | Black |
35 | 3 | Red |
36 | 26 | Black |
American Wheel
Setting up an American wheel is also pretty easy. Just create a new table and name it "American Wheel" and set it up the same as a European wheel (0..37 though) and paste in the following layoutNumber | Actual Number | Color |
0 | 0 | Green |
1 | 28 | Black |
2 | 9 | Red |
3 | 26 | Black |
4 | 30 | Red |
5 | 11 | Black |
6 | 7 | Red |
7 | 20 | Black |
8 | 32 | Red |
9 | 17 | Black |
10 | 5 | Red |
11 | 22 | Black |
12 | 34 | Red |
13 | 15 | Black |
14 | 3 | Red |
15 | 24 | Black |
16 | 36 | Red |
17 | 13 | Black |
18 | 1 | Red |
19 | 00 | Green |
20 | 27 | Red |
21 | 10 | Black |
22 | 25 | Red |
23 | 29 | Black |
24 | 12 | Red |
25 | 8 | Black |
26 | 19 | Red |
27 | 31 | Black |
28 | 18 | Red |
29 | 6 | Black |
30 | 21 | Red |
31 | 33 | Black |
32 | 16 | Red |
33 | 4 | Black |
34 | 23 | Red |
35 | 35 | Black |
36 | 14 | Red |
37 | 2 | Black |
Setting up the random number generator
So we have our wheels setup, so we just need to setup our random number generator. Which we want to look something like this
In this diagram:
- Column A shows our randomly roulette number
- Column B shows the color of the number corresponding to it's color on the European Wheel
This is pretty easy to setup
- Manually enter the headers (Winning Number, Winning Color)
- In row 2, enter the following formula in Column A
=VLOOKUP(RANDBETWEEN(0,36),'European Wheel'!$A$2:$A$38,1)
This basically generates a random number between 0 and 36 (RANDBETWEEN(0,36). And then looks up the corresponding number on our European Wheel tab. If we want to do this for an American Wheel, we would change the tab to American Wheel and set RANDBETWEEN to generate numbers between 0 and 37,
- Enter the following formula in Row 2 of Column B
=VLOOKUP(A3,'European Wheel'!$A$2:$C$38,3)
This will look up the color of the generated number in our European Wheel (again if you want to simulate American Wheels then switch the tab to American Wheel.
- Do the draggy extends paste thing to copy it to multiple rows
Voila, one very simple roulette simulator in Excel.
This is very simple and easily allows you test betting plans etc. I use it for disproving systems such as martingale. It usually proves it to fail pretty quickly. I obviously have more advanced versions which can do billions of rows of testing but Excel isn't so suitable for that. However this should be pretty suitable for your needs.
In a future post, i'll show you how to test staking plans such as Martingale and Reverse Martingale with the simulator. So you can see for yourself how ridiculous these staking plans are.
Hope this helps
UPDATE
You can download a copy of my simulator from my dropbox
https://www.dropbox.com/s/3btgpurbtsuof50/Roulette%20Simulator.xlsx?dl=0
https://www.dropbox.com/s/3btgpurbtsuof50/Roulette%20Simulator.xlsx?dl=0
Tuesday, 6 January 2015
True odds vs Real Odds for Roulette
In my last post, we discussed how you calculate the probability of a single event.
In this post we'll show you how to convert that to odds and then compare the real odds to the probable odds. In this example we'll look at Roulette.
In this post we'll show you how to convert that to odds and then compare the real odds to the probable odds. In this example we'll look at Roulette.
Red or Black Probability
For now we'll focus on European Roulette (although the theory works just as well for American Roulette). In the last post we calculated the probability for Red or Black to be 18/37. In this example we'll bet Red.
There are 18 outcomes that can win for us (i.e. 18 reds), specifically the numbers
1, 3, 5, 7, 9, 12, 14, 16, 18, 19, 21, 23, 25, 27, 30, 32, 34, 35
And 19 outcomes that will lose for us
0, 2, 4, 6, 8, 10, 11, 13, 15, 17, 20, 22, 24, 26, 28, 29, 31, 33, 35
How to convert probability to Decimal Odds
The true probability for Red or Black is 18/37.
To convert this to odds we need to do the following conversion
1 / probability
so for Red or Black this would be
1 / (18 / 37)
which would make the real odds
2.0556
The casino is offering odds of 2/1 which is equivalent in decimal odds of
2.0 (it's 2/1 ;)
What does this mean?
This means the casino is offering odds lower than the true probability of winning. And the casino will profit on the game on average by 2.7% (I'm ignoring "en prison" for now). Remember over a long period of time the actual results will be close enough to the real probability (see other posts). So the casino can expect to make 2.7% on average for every game.
So if that roulette table stakes $100,000,000 per year. Then that roulette table will be profiting by over $2,700,000
And this is how casinos make money.
The bad news is.......
If you play $100,000 worth of games on red or black. you'll probably lose just over $2,700
We play 10,000 games with $10 stakes.
We win 4864 games (10,000 * (18/37)
We lose 5136 games (10,000 - 4864)
We win 4864 * 10 = $48,640
We lose 5136 * 10 = $51,360
In total we lose (£2,720), which the casinos bank as profit
We play 10,000 games with $10 stakes.
We win 4864 games (10,000 * (18/37)
We lose 5136 games (10,000 - 4864)
We win 4864 * 10 = $48,640
We lose 5136 * 10 = $51,360
In total we lose (£2,720), which the casinos bank as profit
The odds are in the casinos favor...
In future posts, we'll show:
- Your bankroll degrading using an excel roulette simulator and how the casino's always win
- A full post on odd conversions
- How to manage books / be a bookie / be a casino
- How fancy betting systems don't help when there is no advantage (Martingale, Reverse Martingale, Fibonacci, James Bond) - they'll all lose you money and we'll show it.
Hope this helps
Calculating Probability for single events
In the last post we figured out how to calculate the number of possible outcomes for coin tosses (or any game). Now we'll look at the probability.
We can calculate the number of possible outcomes using the following formula
We can calculate the number of possible outcomes using the following formula
number of combinations ^ (number of tosses)
otherwise known as
number of combinations to the power of number of tosses.
So to calculate the probability of an event, we are basically saying the following
the number of times the desired outcome can occur / the number of possible outcomes
Coin Toss (2 sided coin)
So for a 2 sided coin (heads or tails). there are 2 possible outcomes.
H, T
So there is 1 in 2 chance of flipping a head (as shown in blue below)
H, T
1 instance of the desired outcome within a set of 2 possible outcomes. Giving a 1/2 chance (50% chance),
Coin Toss (3 sided coin)
So for a 2 sided coin (heads, tails or body). there are 3 possible outcomes.
H, T, B
So there is 1 in 3 chance of flipping a head (as shown in blue below)
H, T, B
1 instance of the desired outcome within a set of 3 possible outcomes. Giving a 1/3 chance (33.3333333% chance),
Dice Throw (6 sided dice)
So for a 6 sided dice throw there are 6 possible outcomes.
1,2,3,4,5,6
So there is 1 in 6 chance of rolling a 3 (as shown in blue below)
1,2,3,4,5,6
There is also a 1 in 6 (16.6666666%) chance of hitting any specific number
And there is a 3 in 6 (otherwise known as 1 in 2 (or 1/2 or 50%)) chance of hitting an even number
1,2,3,4,5,6
And the same for hitting an odd number
1,2,3,4,5,6
Specific Number in European and American Roulette
Now that we know how to calculate probability. Let's look at roulette (we will have bigger discussions on this later).
However in European Roulette, we have the following numbers (note the number zero)
0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
So in European Roulette, the probability of predicting a specific number is 1 in 37, i.e. 1 desired outcome in 37 possible outcomes (2.7% chance)
Specific Number in American Roulette
And in American Roulette, we have the following numbers (not the extra number, the double zero)
0,00,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36
So in American Roulette, the probability of predicting a specific number is 1 in 38, i.e. 1 desired outcome 38 possible outcomes (2.63% chance)
Red or Black in European Roulette
In European Roulette, the zero number is green. And every other number (1 to 36) is evenly split between red and black (will explain the exact wheel in another post). That means there are:
18 reds in 37 possible outcomes
18 blacks in 37 possible outcomes
1 zero (green) in 37 possible outcomes
Which means if you were to bet red or black, you would have an 18 in 37 (18/37 or 48.6% chance)
Red or Black in American Roulette
In American Roulette, the zero and double zero numbers are green. And every other number (1 to 36) is evenly split between red and black (will explain the exact wheel in another post). That means there are:
18 reds in 38 possible outcomes
18 blacks in 38 possible outcomes
2 zero numbers (green) in 38 possible outcomes
Which means if you were to bet red or black, you would have an 18 in 38 (18/38 or 47.36% chance)
Okay, so now you know how to calculate probabilities in single events.
In future posts, I'll discuss the following
- Calculating Odds from Probabilities
- How to make a book and calculating edge and over round
- Calculating probabilities for sequenced events and multiple outcomes (double dice throws, lottery numbers etc).
Calculating the number of possible outcomes of a Coin Toss
In the previous posts, we've been concentrating a lot on coin tosses.
I thought it might be useful for us to look at how we can calculate the possible number of outcomes.
In order to calculate the probability of an event to occur mathematically (or to be able to effectively analyze what happened, we need to be able to calculate all possible outcomes).
So in the case of a coin toss. There are always two possible outcomes in a coin toss. You will either flip heads or tails. So let's look at how this breaks down for multiple coin tosses.
1 coin = 2 outcomes = H,T
2 coins = 4 outcomes = HH,HT,
TH,TT
3 coins = 8 outcomes = HHH,HHT,HTH,HTT,
THH,THT,TTH,TTT
4 coins = 16 outcomes = HHHH,HHHT,HHTH,HHTT,HTHT,HTTH,HTTT,
THHH,THHT,THTH,THTT,TTHT,TTTH,TTTT
And so on and so forth. And it grows quite quickly. Here's how the growth looks charted in Excel.
Hey doesn't the graph look like the streaky graph we did the other day. Sure it's a typical binomial distribution (we'll discuss that another day). For now we'll concentrate on how to calculate on the number of outcomes for higher number ranges.
In our run of 39 coin tosses, it would probably take us a long time to figure out how many outcomes there were if we were to calculate each and every outcome. So we wanna take a shortcut.
3) Do the cell draggy thing to generate the table.
And if you want, you can generate a chart like I did.
BTW, the answer is there is 549,755,813,888 (around 550 billion) possible outcomes for 39 coin tosses.
And if you don't believe me, try it yourself, you know how to check it now
I thought it might be useful for us to look at how we can calculate the possible number of outcomes.
In order to calculate the probability of an event to occur mathematically (or to be able to effectively analyze what happened, we need to be able to calculate all possible outcomes).
So in the case of a coin toss. There are always two possible outcomes in a coin toss. You will either flip heads or tails. So let's look at how this breaks down for multiple coin tosses.
1 coin = 2 outcomes = H,T
2 coins = 4 outcomes = HH,HT,
TH,TT
3 coins = 8 outcomes = HHH,HHT,HTH,HTT,
THH,THT,TTH,TTT
4 coins = 16 outcomes = HHHH,HHHT,HHTH,HHTT,HTHT,HTTH,HTTT,
THHH,THHT,THTH,THTT,TTHT,TTTH,TTTT
And so on and so forth. And it grows quite quickly. Here's how the growth looks charted in Excel.
Hey doesn't the graph look like the streaky graph we did the other day. Sure it's a typical binomial distribution (we'll discuss that another day). For now we'll concentrate on how to calculate on the number of outcomes for higher number ranges.
In our run of 39 coin tosses, it would probably take us a long time to figure out how many outcomes there were if we were to calculate each and every outcome. So we wanna take a shortcut.
Inferring the number of outcomes for a coin toss
Just by looking at the coin toss data, we can probably infer that the number of outcomes doubles every time we add a coin. 2 outcomes, 4 outcomes, 8 outcomes, 16 outcomes...
And you'll probably notice the pattern as well. If you look at 3 coin tosses, you can see that the first line is the exact same sequence as 2 coin tosses except the first coin toss is a head. And the second line is again the same sequence as 2 coin tosses but with a tail as the first coin toss. So sure, it makes sense that it doubles up all the time.
So we are really number of combinations (heads or tails) * number of combinations for previous number of coins
1 coin = 2 outcomes
2 coins = 4 outcomes = 2 outcomes * 2 combinations
3 coins = 8 outcomes = 4 outcomes * 2 combinations
4 coins = 16 outcomes = 8 outcomes * 2 combinations
What would happen if we had a 3 sided coin
Let's say for a second our coin has 3 sides, (Heads, Tails and errrrr Body). What would happen?
1 coin = 3 outcomes = H, T, B
2 coins = 9 outcomes = HH, HT, HB,
TH, TT, TB,
BH, BT, BB
3 coins = 27 outcomes = HHH,HHT,HHB,HTH,HTT,HTB,HBH,HBT,HBB,
THH,THT,THB,TTH,TTT,TTB,TBH,TBT,TBB,
BHH,BHT,BHB,BTH,BTT,BTB,BBH,BBT,BBB
Okay, so now we seem to be tripling up from the previous set of outcomes.
And we're following the same pattern as before.
First line is the exact same sequence as 2 coin tosses except the first coin toss is a head. And the second line is again the same sequence as 2 coin tosses but with a tail as the first coin toss. And the third line is again the same sequence as 2 coin tosses but with a body as the first coin toss. So sure, it makes sense that it triples up all the time.
So we are still number of combinations (heads, tails, body) * number of combinations for previous number of coins
1 coin = 3 outcomes
2 coins = 9 outcomes = 3 outcomes * 3 combinations
3 coins = 27 outcomes = 9 outcomes * 3 combinations
4 coins = 81 outcomes = 27 outcomes * 3 combinations
This is cool but would still be a pain to calculate 39 combinations. What's the formula for doing this?
Quick way for calculating number of outcomes
The quick way of doing this is calculating this is
number of combinations ^ (number of tosses)
otherwise known as
number of combinations to the power of number of tosses.
So for a 3 sided coin with 2 tosses, we would calculate it as
1 coin = 3^1 = 3 = 3 outcomes
2 coins = 3^2 = 3 * 3 = 9 outcomes
3 coins = 3^3 = 3 * 3 * 3 = 27 outcomes
4 coins = 3^4 = 3 * 3 * 3 * 3 = 81 outcomes
And for our regular 2 sided coin
1 coin = 2^1 = 2 = 2 outcomes
2 coins = 2^2 = 2 * 2 = 4 outcomes
3 coins = 2^3 = 2 * 2 * 2 = 8 outcomes
4 coins = 3^4 = 2 * 2 * 2 * 2 = 16 outcomes
How would we do this in Excel (or google docs)?
Pretty easy really. Lets generate a table of coin tosses (2 sided coin) with their number of outcomes. Something like this
To generate this in excel, we just do the following
1) In column A, just put the number of coin tosses to generate.
2) In column B, just put in the following formula
=POWER(2,A2)
And if you want, you can generate a chart like I did.
BTW, the answer is there is 549,755,813,888 (around 550 billion) possible outcomes for 39 coin tosses.
And if you don't believe me, try it yourself, you know how to check it now
Monday, 5 January 2015
a lot of talk of coin tosses
just so you know, i didn't really intend to talk about coin tosses...
But as I started writing this blog and started writing up my thoughts, i found it easier to talk about coin tosses. it goes a bit like this, much of what applies to roulette, sports betting, blackjack, poker etc also applies to coin tosses.
so let's explain this stuff in the simplest possible context (i.e. coin tosses) and then we get to roulette etc. we've got most of the basic concepts.
i really hope this helps. and you don't mind the coin tosses for now.
i promise we'll get to the predicting roulette, physics of roulette, predicting blackjack, sports betting, betting patterns and techniques soon.
but for now, let's continue with the basics.
But as I started writing this blog and started writing up my thoughts, i found it easier to talk about coin tosses. it goes a bit like this, much of what applies to roulette, sports betting, blackjack, poker etc also applies to coin tosses.
so let's explain this stuff in the simplest possible context (i.e. coin tosses) and then we get to roulette etc. we've got most of the basic concepts.
i really hope this helps. and you don't mind the coin tosses for now.
i promise we'll get to the predicting roulette, physics of roulette, predicting blackjack, sports betting, betting patterns and techniques soon.
but for now, let's continue with the basics.
How to calculate streaks in excel
So in my last post, i showed you how to look at streaks for coin tosses in Excel
And then we can generate a chart against the distribution, i used a nice bar chart
Anyways, i hope this is useful. This technique is pretty useful for predicting monte-carlo predictions of betting patterns and proving how bad things like martingale are. And if you develop some advantage play system in blackjack, roulette, sports betting then this technique will allow you to calculate streaks.
Hope it helps.
In this post, i'll show you how i calculate the streaks in Excel (or google docs). You can use the same technique to calculate streaks in anything (i.e. roulette).
What is a streak?
For coin tosses would be the same coin be thrown X times in a row. e.g. 7 heads in a row would be a streak.
What does this look like in Excel?
First thing we need to do is track the length of a streak. This is how it looks for me in Excel.
1) In the left hand column is the result of the coin toss (heads or tails), you can generate this randomly using the technique described in my blog post, generating random coin tosses in excel
2) The right hand column tells the length of the streak. As you can see in the image above, every time the streak ends, the count gets reset to 1. If the latest coin toss is the same as the previous it increases
How do I calculate this in Excel (or google docs)?
I've used the following formula to calculate this. I put this in each cell in the B column
=IF(A2=A1,B1+1,1)
Column A is the column with the result of the coin toss
Column B is the length of the streak.
This formula basically says, if the current coin toss is the same as the previous coin toss then increment the streak count by 1. If the coin toss is different from the previous, reset the streak count to 1.
Summarizing the Data
Now that we've got the streaks. We want a summary of the data. This is how it looks in Excel
And this how we calculate it
1) Left column is a just count of the streaks (just enter 1 to 15)
2) Right column is calculated by the following formula
=COUNTIF(Sheet17!B2:B20001,A2)-B3
This formula counts up all the entries in the range that meet the IF condition. In our case, we look at our 20,000 coin tosses that we've held in Sheet 17 in cells B2 to B20001. The condition it needs to meet is the value in the A column (i.e. the length of the sequence, in this case A2). So for row 10, it will count up all the sequences of length 9. Now obviously a sequence of 9 will also count all the sequences of 10's. So we subtract that number that's next in the sequence. We don't want to count all the 10's in the 9 sequence count (that's the -B3 bit).
Anyways, i hope this is useful. This technique is pretty useful for predicting monte-carlo predictions of betting patterns and proving how bad things like martingale are. And if you develop some advantage play system in blackjack, roulette, sports betting then this technique will allow you to calculate streaks.
Hope it helps.
Streaky coin tosses
In the previous post we looked at how the data distribution gets closer to the real probability as the number of samples increase. And we showed this happened in our excel generated random coin tosses. What we didn't look at was how streaky the coin tosses were?
A streak is considered as the same coin toss in a row i.e. Heads, Heads, Heads, Heads. Would be a streak of 4 heads.
A streak is considered as the same coin toss in a row i.e. Heads, Heads, Heads, Heads. Would be a streak of 4 heads.
So what happened in our 20,000 coin tosses?
As we can see above, we get lots of long streaks. Actually we get 14 tosses of the same side in a row. That's quite a streak.
And that curve looks quite appropriate. It's a binomial distribution and one that looks like it should for a 50/50 probability.
At another time, we'll discuss binomial distributions and the probability of coin tosses. But for now, lets just say, a streak of 14 is quite normal. :)
Loss Recovery betting plans aren't looking so good now.
Do we really wanna talk about Martingale / loss recovery betting patterns? Really? Ummm ok, we can. Are you sure you're expecting a black just because the last color was a red? Are you sure you're owed a black. Are you sure 10 reds in a row is unusual? Looks pretty normal to me.
Labels:
Coin Toss,
Excel,
Hot Numbers,
Martingale,
Randomness,
Roulette
The more coin tosses we do, the closer we get to 50/50
So in the previous posts, we looked at how we can generate coin tosses in excel using a random number generator.
One of the things we noted is that although coin tosses have a 50/50 chance of being heads or tails, it doesn't work out that way in reality for small samples. The following shows the distribution for one of our excel simulations (39 coin tosses)
This is what happens as we increase the number of tosses, here is 1000 tosses
And 5000 tosses
10,000 tosses
And 20,000 tosses
As we can see, the more tosses we do the closer we get to 50%, but ultimately it's still random, so we can still get a decent difference.
This tells us a few things:
- Small samples aren't sufficient to predict trends or check randomness
- The more samples we do the more likely we will get something that fits the probability
- Random doesn't look random
- Our coin tosses over a large number of samples fits the probability
In future posts we will check some of these simulations for betting patterns such as martingale, check probabilities of roulette tables etc etc
And of course, you can check all of this yourself using the techniques described in the earlier post. You don't need to take my word for it.
Subscribe to:
Posts (Atom)