Challenge: What's the best way to win 1 BTC with 1 BTC? - page 2.

hua_hui

legendary

Activity: 1386

Merit: 1016

Quote from: Aemon on October 21, 2015, 08:47:58 AM

So basically you are saying that strategy is good and the best thing you can use no matter what? No matter the amount of large rolls you do you should always be doing this is what I am gathering, is that correct?

nope, what we are trying to show u is that no matter how many different stragety we can come out with, the chance of getting it right is always less than 50%. so in another words, it is more likely you lost it all than to gain. so best is dont bet.

ndnh

legendary

Activity: 1302

Merit: 1005

New Decentralized Nuclear Hobbit

Quote from: Aemon on October 21, 2015, 08:47:58 AM

...you do you should always be doing this is what I am gathering, is that correct?

Not really. If you are concerned about increasing your EV and beating the edge, you shouldn't play. Whichever strategy you pick, you can only minimize the impact of house edge on your EV.

Maximize impact of HE on EV : bet at 98% win chance. Grin

Minimize impact of house edge : follow the winning strategies of Dooglus and the other winner.

Nullify impact of house edge : Don't play.
Beat the house edge : Not possible.

Aemon

sr. member

Activity: 308

Merit: 250

So basically you are saying that strategy is good and the best thing you can use no matter what? No matter the amount of large rolls you do you should always be doing this is what I am gathering, is that correct?

hua_hui

legendary

Activity: 1386

Merit: 1016

Quote from: Rammus on October 20, 2015, 08:58:15 PM

Quote from: leex1528 on October 19, 2015, 07:45:12 PM

Doog, how did you learn all of this? I would love to try and figure out about calculations and what actually gives you the best odds. Did you take a lot of stats classes in college or what? Any thoughts?

I`m pretty sure it comes from either majoring in math..

That alone or picking up and building the dice itself and studying the game which lead to his site? who knows lol.

for my country(in asia), probability is learn at y7 to y10, where you need to learn how to calculate when you throw up to 3 dices or pick 3 balls. of course, for good sch, they have to learn on how to calculate for n throws where n can be large. for advance y11 to y12, we learn for large n, what probability for specific k times of winning. but i believe dooglus even learn up to university standard as he mention that he know about L'Hôpital's rule which is only taught when your major has things to do with math where you have to look at the big picture.

Rammus

full member

Activity: 138

Merit: 100

Quote from: leex1528 on October 19, 2015, 07:45:12 PM

Doog, how did you learn all of this? I would love to try and figure out about calculations and what actually gives you the best odds. Did you take a lot of stats classes in college or what? Any thoughts?

I`m pretty sure it comes from either majoring in math..

That alone or picking up and building the dice itself and studying the game which lead to his site? who knows lol.

hua_hui

legendary

Activity: 1386

Merit: 1016

Quote from: leex1528 on October 19, 2015, 07:45:12 PM

Doog, how did you learn all of this? I would love to try and figure out about calculations and what actually gives you the best odds. Did you take a lot of stats classes in college or what? Any thoughts?

well it is actually quite simple mathematical calculation. you can start using it when you learn probability with tree diagram. but instead of picking 2 or 3 people or throwing 2/3 dice, the same occuring series can be applied to larger number of trials.

on higher level(my country use terms diff from you so i may use the wrong term) in pre-u school, you can learn even how to estimate your probability by approximating the binominal distribute to other form of distribute that allow more easy of calculation.

And to be honest, this is the first time i ever encountered after learning L'Hôpital's rule in uni. lol. finally what i learn got some use to it!

knowhow

hero member

Activity: 560

Merit: 500

You can do all calculations but only one thing will allow to earn 1 btc playing with 1 btc that is lucky the 49% means nothing besides you can loose as win anytime.

leex1528

hero member

Activity: 784

Merit: 1000

Doog, how did you learn all of this? I would love to try and figure out about calculations and what actually gives you the best odds. Did you take a lot of stats classes in college or what? Any thoughts?

Coef

hero member

Activity: 882

Merit: 1000

Exhausted

Quote from: dooglus on October 18, 2015, 09:21:02 PM

Are you able to beat 49.6522222%, or do you think that is a hard limit?

After going through post #198 carefully, I agree the number 1 - 2**-0.99 you get should be the limit under perfect conditions (the bet size and multiplier can always be set to make the balance after winning a bet to be exactly 2 btc). Under actual conditions (bet size cannot contain a fraction of satoshi, bet win chance % can only contain 2 decimal place) it cannot be done and so the "best" way would be slightly lower than that.

BTW, I have not heard of the word "L'Hopital's rule" for like a decade lol.

Quote from: dooglus on October 18, 2015, 09:21:02 PM

My recent post shows strategies that give a 49.6522222% percent chance of winning, so your post (which came after) doesn't beat mine.

As mentioned in the OP, the betting sequence has to be well-defined and subject to the constraints of a dice site such as Bitdice or PRC. IMHO, the numbers you get in that post does not seem to qualify. Tongue

Quote from: RHavar on October 18, 2015, 04:38:33 PM

Quote from: Coef on October 18, 2015, 03:28:30 PM

I notice the word "them" in the quote. Do I get 0.1 btc for beating dooglus' solution? Tongue

Damn it. Guess that serves me right for trying to use gender-neutral pronouns. What's your MP or BaB username?

(Give me a day or two to verify the solution, and after this, no more prize money will be given out Grin

)

Received the 0.1 btc. Thanks Ryan. Grin

knowhow

hero member

Activity: 560

Merit: 500

On the begining for sure bet 1 btc worthing 1 dollars weren nothing nowadays is worthing 250dollars soo why the hell i should try with 1 btc to double since i know and most of us know that the lucky usually stays with the owner of the project being casino,dice or any other way.

Xenoph0bia

hero member

Activity: 966

Merit: 501

Quote from: wearepoor on October 18, 2015, 04:23:17 PM

I think its possible but at the same time it is little difficult too. I believe that the best way to earn 1 btc is sports betting only If you have good knowledge about sports and if you are able to predict the result. So as per me you can take a chance with sports betting as there are many chances to make some good profit.
Avoid the other ways of gambling where you can loose the money.

Agreed with you that it is possible but at the same time its difficult too. People can loose the money while trying to earn one btc in gambling, they should follow some techniques and other strategies to minimize the loss or else they will loose their 1 btc instead of earning 1 btc.

dooglus

legendary

Activity: 2940

Merit: 1330

Quote from: dooglus on October 17, 2015, 02:02:41 PM

I'd love an explanation of why it tops out at 0.496522222!

It turns out that number doesn't end with recurring 2's, but is in fact:

>>> 1 - 2**-0.99
0.49652222497164056

I calculated it using L'Hôpital's rule, which I vaguely remember learning about in high school.

dooglus

legendary

Activity: 2940

Merit: 1330

Quote from: Coef on October 18, 2015, 03:28:30 PM

Chance of winning = 1 - ((1 - 0.02%) ^ 3430 * (1 - 0.01%)) = 49.649851332%.

Chance of winning = 1 - ((1 - 0.03%) ^ 2287) = 49.651579392%

I notice the word "them" in the quote. Do I get 0.1 btc for beating dooglus' solution? Tongue

My recent post shows strategies that give a 49.6522222% percent chance of winning, so your post (which came after) doesn't beat mine.

Are you able to beat 49.6522222%, or do you think that is a hard limit?

Here's a chart showing your two strategies, and where they fit on the curve of possible martingale strategies. The higher the payout multiplier you play with, the closer you get to the 49.652222% chance:

I wonder if flat-betting would be better?

Like we could divide the 1 BTC into N=10 0.1 bets.
Bet the first 0.1 at 11x. If we win, we're 1 BTC up, so we stop.
Bet the 2nd 0.1 12x. If we win, we're 1 BTC up, so we stop.
Etc.
Bet the Nth at (10+N)x. We're always 1 BTC up if we win.

Turns out it isn't better. What's funny is it is exactly the same.

Here's a Python script that calculates the chance of doubling up using this flat-betting strategy for various N:

Quote

for m in range(0, 6):
for N in range(10**m, 10**(m+1), 10**m):
p = 1.0
for i in range (1, N+1):
p *= (1 - 0.99 / (N+i))

print "%6d %.8f" % (N, 1 - p)

And here's its output:

Quote

1 0.49500000
2 0.49582500
3 0.49607333
4 0.49619171
5 0.49626081
6 0.49630606
7 0.49633797
8 0.49636168
9 0.49637999
10 0.49639455
20 0.49645915
30 0.49648035
40 0.49649088
50 0.49649718
60 0.49650137
70 0.49650436
80 0.49650660
90 0.49650834
100 0.49650973
200 0.49651599
300 0.49651807
400 0.49651911
500 0.49651973
600 0.49652015
700 0.49652044
800 0.49652067
900 0.49652084
1000 0.49652098
2000 0.49652160
3000 0.49652181
4000 0.49652191
5000 0.49652198
6000 0.49652202
7000 0.49652205
8000 0.49652207
9000 0.49652209
10000 0.49652210
20000 0.49652216
30000 0.49652218
40000 0.49652219
50000 0.49652220
60000 0.49652220
70000 0.49652221
80000 0.49652221
90000 0.49652221
100000 0.49652221
200000 0.49652222
300000 0.49652222
400000 0.49652222
500000 0.49652222
600000 0.49652222
700000 0.49652222
800000 0.49652222
900000 0.49652222

It converges on 49.6522222% chance of doubling up again!

RHavar

legendary

Activity: 2557

Merit: 1886

Quote from: Coef on October 18, 2015, 03:28:30 PM

I notice the word "them" in the quote. Do I get 0.1 btc for beating dooglus' solution? Tongue

Damn it. Guess that serves me right for trying to use gender-neutral pronouns. What's your MP or BaB username?

(Give me a day or two to verify the solution, and after this, no more prize money will be given out Grin

)

wearepoor

hero member

Activity: 784

Merit: 500

I think its possible but at the same time it is little difficult too. I believe that the best way to earn 1 btc is sports betting only If you have good knowledge about sports and if you are able to predict the result. So as per me you can take a chance with sports betting as there are many chances to make some good profit.
Avoid the other ways of gambling where you can loose the money.

Coef

hero member

Activity: 882

Merit: 1000

Exhausted

Quote from: dooglus on October 17, 2015, 02:13:43 PM

Quote from: mexxer-2 on October 17, 2015, 02:08:39 PM

Hmm just thought about it, has OP yet paid the prize he has claimed to pay to the best method or is he still searching for more strats?

He already paid, and it's likely we have the best strategy already.

It is dangerous to jump to the conclusion quickly. Tongue

Quote from: dooglus on October 17, 2015, 02:21:05 PM

Quote from: blockage on January 29, 2015, 12:24:18 PM

leading to a 49.64716% chance of winning.

Quote from: blockage on January 29, 2015, 12:24:18 PM

you have a 49.64812% chance of winning.

Quote from: dooglus on February 20, 2015, 09:58:23 PM

chance of winning = 100 * (1 - ((100 - 1.0664) / 100) ** 64) = 49.649475445976265%

With 1 btc, I should be able to make a total of 3430 x4950 (ie 0.02%) bets, each with a bet size of roundup((2-balance)/4949) to the nearest satoshi that could bring my balance to slightly over 2 btc if the bet wins. After that, I will have 0.00027093 left and a final x9900 (ie 0.01%) bet can be made which could bring my balance to 2.682207 btc.
Check http://pastebin.com/TL7xLTx4 for the complete betting sequence in details in CSV format.

Chance of winning = 1 - ((1 - 0.02%) ^ 3430 * (1 - 0.01%)) = 49.649851332%.

Or, I should be able to make a total of 2287 x3300 (ie 0.03%) bets, each with a bet size of roundup((2-balance)/3299) to the nearest satoshi.
Check http://pastebin.com/Je5ujHH9 for the complete betting sequence in details in CSV format.

Chance of winning = 1 - ((1 - 0.03%) ^ 2287) = 49.651579392%

Quote from: RHavar on October 11, 2015, 09:47:34 PM

Quote from: dooglus on October 10, 2015, 09:39:46 PM

Yeah, but you also said:

Quote from: RHavar on January 29, 2015, 06:24:35 PM

added 0.1 to blockage's account. If someone can beat his solution, I'll give them 0.1 as well.

Ahh touché, forgot I wrote that.

I notice the word "them" in the quote. Do I get 0.1 btc for beating dooglus' solution? Tongue

knowhow

hero member

Activity: 560

Merit: 500

Well one funny way to get fast 1btc ,im sure you cant play with 1 btc but you can earn 600% or even 1000% playing a game ,and well only need lucky and a good conection im playing yesterday i made 40ksatoshis with 10k fee xD,i guess i cant post any ref link here soo send me pm.

dooglus

legendary

Activity: 2940

Merit: 1330

Quote from: mexxer-2 on October 17, 2015, 02:15:50 PM

Could you point to the post which won it? And did it really have lower chances of losing?(highest EV I mean)

These two posts:

Quote from: blockage on January 29, 2015, 12:24:18 PM

leading to a 49.64716% chance of winning.

Quote from: blockage on January 29, 2015, 12:24:18 PM

you have a 49.64812% chance of winning.

Quote from: dooglus on February 20, 2015, 09:58:23 PM

chance of winning = 100 * (1 - ((100 - 1.0664) / 100) ** 64) = 49.649475445976265%

And yes, they have a higher EV because on average they risk less. The EV is -1% of the amount you risk, so minimising the amount you risk maximises the EV.

The corollary to this of course is that the optimal strategy for maximising your EV is to risk 0 - ie. stop playing.

A strange game. The only winning move is not to play. How about a nice game of chess?

mexxer-2

hero member

Activity: 924

Merit: 1005

4 Mana 7/7

Quote from: dooglus on October 17, 2015, 02:13:43 PM

Quote from: mexxer-2 on October 17, 2015, 02:08:39 PM

Hmm just thought about it, has OP yet paid the prize he has claimed to pay to the best method or is he still searching for more strats?

He already paid, and it's likely we have the best strategy already.

Could you point to the post which won it? And did it really have lower chances of losing?(highest EV I mean)

dooglus

legendary

Activity: 2940

Merit: 1330

Quote from: mexxer-2 on October 17, 2015, 02:08:39 PM

Hmm just thought about it, has OP yet paid the prize he has claimed to pay to the best method or is he still searching for more strats?

He already paid, and it's likely we have the best strategy already.

Topic: Challenge: What's the best way to win 1 BTC with 1 BTC? - page 2. (Read 15897 times)