1
| |
yo
if you win when your roll of 50 is greater than the result of [ ( random1-100) + random1-100 ) ]or [ (random2-200) ], are you doing the same thing in both cases? ie do you win ~25% of the time in both situations?
just checkin'
thx
|
|
|
1
|
TalentedTom   Canada. Oct 23 2011 19:09. Posts 20070 | | |
|
Our deepest fear is not that we are inadequate. Our deepest fear is that we are powerful beyond measure. It is our light not our darkness that most frightens us and as we let our own lights shine we unconsciously give other people permision to do the same | |
|
|
1
|
taco   Iceland. Oct 23 2011 19:20. Posts 1793 | | |
In 2-200 you're only excluding 1 number (1) out of 200,
in (1-100)+(1-100) you're excluding 1 number (1) out of 200 of being possible.
(100/199*49%) win chance in 2-200
and (100/199*49)% win chance in (1-100)+(1-100)
Same calculation, same probability.
+ Show Spoiler +
Unless 1-100 = excludes 1 and 100 from the draw, in that case they are different
since that would be (100/197*49)% vs (100/196*49)% |
|
|
1
|
palak   United States. Oct 23 2011 19:41. Posts 4601 | | |
NO...this is a binomial distribution problem if I'm reading what ur asking correctly.
if random 1-100 r independent then no it's NOT the same distribution of random 2-200..think about it this way, odds of random 2-200 getting 100 or 200 exactly are both 1/199...odds of 2 random 1-100 trials adding to exactly 200 is 1/10000 since it's (1/100*1/100) whereas 2 random 1-100 adding to 100 could be done by 30+70, 60+40,55+45,etc each specific outcome has a probability of 2/10000 (1/10000+1/10000 since 30+70 is 1/10000 and 70+30 is 1/10000)..this follows a binomial distribution w/ 10k states which is a royal bitch to calculate right now for me.
But as a classic easier example would be 2 D6 vs 1 D12 w/ the 1 removed from the die so only 2-12 are possible
If you roll a 4 then the odds ur roll is greater then the sum of 2 D6 die you get .028+.056 or 1/36+2/36 (odds of getting 2+odds of getting 3) for a total probability of .084, 8.4%
For a die of 2-12 contains 12-2+1 sides so 11 sides, your odds of rolling below 4 on one roll of the die is .091+.091 or 1/11+1/11 or 2/11 which is 18%.
http://hyperphysics.phy-astr.gsu.edu/hbase/math/dice.html#c2
|
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | Last edit: 23/10/2011 19:58 |
|
|
4
|
Bigbobm   United States. Oct 23 2011 20:24. Posts 5511 | | |
at a glance i have to disagree with the above.
if the dice are assumed independent of each other, there should be no difference.
im pretty sure this still forms a pretty uniform distribution that mimics the first, which is all were worried about |
|
Its time to stop thinking like a bitch and think smart like a poker player - ket | Last edit: 23/10/2011 20:29 |
|
|
1
|
palak   United States. Oct 23 2011 20:33. Posts 4601 | | |
it does not form a uniform distribution it forms a standard distribution |
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | |
|
|
4
|
Bigbobm   United States. Oct 23 2011 20:42. Posts 5511 | | |
ah you're right.
the 2-200 is uniform, 2d 1-100 is standard.
|
|
Its time to stop thinking like a bitch and think smart like a poker player - ket | |
|
|
1
|
palak   United States. Oct 23 2011 21:12. Posts 4601 | | |
decided to figure it out for op,....here's an easyish way to do it.
Go again to rolling 2 six sided die. If you look at the distribution on the link I posted a patter occurs which is. Odds of getting a 2=1/36 (1+1) 3=2/36 (2+1 or 1+2), odds of getting a 4=3/36 (3+1,1+3,2+2) so odds of getting a 5=4/36 (4+1,1+4,2+3,3+2). See the pattern? Odds of getting x is equal to (x-1)/(sides^2)..this pattern follows up to the midway point. For values above the midway point the formula becomes probably of getting y is found by (2*median value-y)=x x willl be less then the median and will be the number u should use.. Since that was poorly worded the odds of getting an 8 (on 2 6 sided die) would be found by 2*7-8=6.the probabily of getting 6 is 5/36 (1+5,5+1,4+2,2+4,3+3) same as getting an 8 (4+4,5+3,3+5,6+2,2+6) and 10 would be 7*2-10=4, probability of a 4 is 3/36 which is the same as a 10.
Anyhow that's the pattern. So how to do less then. Odds of getting less then 6 is just odds of 5+odds of 4+odds of 3+odds of 2. which is 4/36+3/36+2/36+1/36. So odds of less then x=(sum from 1 to x-2)/(total combinations) odds of below a 7 would be (5+4+3+2+1)/36 this breaks and must be altered for above the median but thats irrelevant since 50 is below the median value.
For your problem the odds are below a 50 on 2 100 sided die. So follow the pattern (48+47+...1)/10000 or 1176/10000 or 11.76%. Easy way to add consecutive numbers 1 to x is x*(x+1)/2 so 48 to 1 added together is 48*49/2 or 1176.
fwiw the odds of getting below a 50 on the 2-200 is 49/199 or 24.6% not even close to the same probability. |
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | Last edit: 23/10/2011 21:16 |
|
|
1
|
nixxxbg   Bulgaria. Oct 23 2011 22:59. Posts 436 | | |
What are random1 and random2?
Is random1 uniform on the discrete set {1,2,3,...,100}?
Is random2 uniform on the set {1,2,3,...,200}? |
|
|
1
|
palak   United States. Oct 23 2011 23:05. Posts 4601 | | |
random 1 discrete set, natural numbers 1-100
random 2 discrete set, integers 2-200
least that's my assumption |
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | |
|
|
1
|
nixxxbg   Bulgaria. Oct 23 2011 23:31. Posts 436 | | |
Oh I see. So we have X is uniform on {1,...,100} and Y is uniform on {2,...,200}. Then:
P(X+X <= 50) = P(X <= 25) = 25/100 = 0.25
P(Y <= 50) = 49/199 ~= 0.246
Or are we assuming that we draw X again, independently, i.e. Z is uniform on {1,...,100} and we want P(X+Z <= 50)? |
|
| Last edit: 23/10/2011 23:38 |
|
|
1
|
1
|
palak   United States. Oct 23 2011 23:37. Posts 4601 | | |
@nixxxbg that doesn't work. Both numbers on random1 don't have to be below 25. You could have x1=47, x2=1 so X=48 or x1=35 x2=12 ->X=47...both those satisfy that. Also you could have for example x1=10 x2=40 making X=50 and not meeting the requirement.
|
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | Last edit: 23/10/2011 23:38 |
|
|
1
|
nixxxbg   Bulgaria. Oct 23 2011 23:54. Posts 436 | | |
Sorry, I was confused.
Let X~Uniform{1,...,100} and Z~Uniform{1,...,100} and suppose they are independent. Then:
P(X+Z <= 50) = sum_j_2_50 P(X+Z = j) = sum_j_2_50 sum_k_1_100 P(X=k)*P(Z = j-k) = 1/100 * (1/100 + 2/100+ ... + 49/100) = 1225/10000 = 0.1225
I guess you had the right answer already. Sorry for the confusion. |
|
| Last edit: 23/10/2011 23:55 |
|
|
1
|
palak   United States. Oct 24 2011 00:10. Posts 4601 | | |
^u only go from 1-48, not 49...49/100 doesn't work since that would mean the sum would total at least 50.
U can either add 1-48 or 2-49 depending if u want to add single die value or sum. The answer is obv the same but u can not add 1-49 |
|
dont tap the glass...im about ready to take a fucking hammer to the aquarium | Last edit: 24/10/2011 01:14 |
|
|
1
|
Highcard   Canada. Oct 24 2011 03:37. Posts 5428 | | |
thanks palak for the write up |
|
I have learned from poker that being at the table is not a grind, the grind is living and poker is how I pass the time | |
|
|
1
| |
|
Luck just can´t explain it | |
|
|
1
|
longple   Sweden. Oct 24 2011 10:18. Posts 4472 | | |
|
|
1
|
longple   Sweden. Oct 24 2011 10:20. Posts 4472 | | |
am i stupid for not understanding anything of anything here? |
|
|
1
| |
| On October 23 2011 18:09 TalentedTom wrote:
yes |
''Hi im TalentedTom its a coin flip so ill answer and hope to be right so ill look smart''
xD |
|
|
|