Any statistics fien...
 

  You don't need to be an 'investor' to invest in Singletrack: 6 days left: 95% of target - Find out more

[Closed] Any statistics fiends in the house?

42 Posts
16 Users
0 Reactions
56 Views
 DrJ
Posts: 13416
Full Member
Topic starter
 

I have a statistics problem that is probably simple when you know the answer, but ... I don't.

I have a bag with 7 white balls and one black ball. I draw a ball from it at random and then put it back in the bag. I do this a total of 8 times. What is the probability that I draw the black ball exactly once?

Thanks for saving my sanity!!


 
Posted : 03/12/2015 10:34 pm
Posts: 31056
Free Member
 

114%


 
Posted : 03/12/2015 10:37 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

That's no better than my answer DD 🙂


 
Posted : 03/12/2015 10:40 pm
Posts: 5559
Free Member
 

100%- basically you should draw each ball once in a random trial

That said this wont happen and what will happen is that in say 8000 picks you will pick it 1000 times and sometimes you will pick it once and sometimes more than this and sometimes you wont pick it

I think its going to be one of those that depends on exactly how you word the question


 
Posted : 03/12/2015 10:46 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

You guys are hopeless !! 🙂


 
Posted : 03/12/2015 10:48 pm
Posts: 36
Free Member
 

4.9%

I reckon it's 1/8 * (7/8)^7


 
Posted : 03/12/2015 10:49 pm
Posts: 1593
Full Member
 

Approx 5%...

Edit: See what Stoner said above...


 
Posted : 03/12/2015 10:50 pm
Posts: 5559
Free Member
 

says the one who has to ask the internet 😉


 
Posted : 03/12/2015 10:50 pm
Posts: 0
Free Member
 

39% for all 8 ways of doing it.
4.9% for it drawing black the first time.


 
Posted : 03/12/2015 10:53 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

I reckon it's 1/8 * (7/8)^7

I though that but isn't it the prob of black on the first draw and then white?


 
Posted : 03/12/2015 10:54 pm
Posts: 36
Free Member
 

quite right. Stoat's on the money, I forgot to *8 for all the permutations.

If it's any defence, my scribblings here do have the *8 on it.


 
Posted : 03/12/2015 10:56 pm
Posts: 1593
Full Member
 

Although... on second thoughts...

If I draw the balls in order, then I can get (b for black, w for white)

bwwwwwww
wbwwwwww
wwbwwwww
wwwbwwww
wwwwbwww
wwwwwbww
wwwwwwbw
wwwwwwwb

Now, each of those probabilities is (1/8) * (7/8)^7... but there are 8 of them..

So, to revise my original answer... about 40%

I am possibly very wrong though...


 
Posted : 03/12/2015 10:56 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

Part 2

If there are 4 black balls and 4 white, what is the prob of drawing 4 white and 4 black.


 
Posted : 03/12/2015 10:57 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

Now, each of those probabilities is (1/8) * (7/8)^7... but there are 8 of them..

So, to revise my original answer... about 40%

I followed that logic but doesn't it seem odd to you that this rather special outcome would be almost 50:50?


 
Posted : 03/12/2015 11:05 pm
Posts: 0
Free Member
 

Part 2

If there are 4 black balls and 4 white, what is the prob of drawing 4 white and 4 black.

(0.5)^8 x 64 = 0.25


 
Posted : 03/12/2015 11:06 pm
Posts: 0
Free Member
 

27%
70 combinations

But this time, that's cheating with an online combinations calculator and probable BS
36 yrs since a level. Too old.


 
Posted : 03/12/2015 11:13 pm
Posts: 0
Free Member
 

[quote=DrJ ]I followed that logic but doesn't it seem odd to you that this rather special outcome would be almost 50:50?

The thing is, it's not that special - actually picking the black ball once is the "expected" outcome, ie the one you'd expect to happen most often.

Far too easy a problem, can't we have something more difficult, like the probability of getting a clear winner from several candidates in an election? 😉


 
Posted : 03/12/2015 11:13 pm
Posts: 0
Free Member
 

[quote=Stoatsbrother ]27%
70 combinations
But that's cheating with an online combinations calculator and probable BS

Seems right - without thinking about it too hard I'd expect the formula to be:
(4/8)^8 * 8! / (4! * 4!)
the latter part gives 70 combinations.

Similarly for the first question, number of combinations = 8! / (7! * 1!)


 
Posted : 03/12/2015 11:16 pm
Posts: 0
Free Member
 

Hopeless 😀


 
Posted : 03/12/2015 11:19 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

OK smart guys ...

3 red balls, one black, 4 white

Prob of 3 red balls and 5 white
Prob of 1 black 7 white

🙂


 
Posted : 03/12/2015 11:20 pm
Posts: 15068
Full Member
 

Your talking about probability, not statistics. Just sayin!


 
Posted : 03/12/2015 11:20 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

The thing is, it's not that special - actually picking the black ball once is the "expected" outcome, ie the one you'd expect to happen most often.

Yeah - just checked with 2000 tries on Excel and it's about right. Just felt odd ....


 
Posted : 03/12/2015 11:21 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

Your talking about probability, not statistics. Just sayin!

Same difference 🙂


 
Posted : 03/12/2015 11:22 pm
Posts: 31056
Free Member
 

23%


 
Posted : 03/12/2015 11:26 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

and a half


 
Posted : 03/12/2015 11:30 pm
Posts: 1593
Full Member
 

It's easier to see how it works if you just use 3 balls... 2 white 1 black..

www = (2/3)^3 = 29.6%
wwb = (2/3)^2 * 1/3 = 14.8%
wbw = 2/3 * 1/3 * 2/3 = 14.8%
wbb = 2/3 * (1/3)^2 = 7.4%
bww = 14.8%
bwb = 7.4%
bbw = 7.4%
bbb = 3.7%

So, for all white the probability is 29.8%, however for 2 white and 1 black it's approx 45%...


 
Posted : 03/12/2015 11:31 pm
Posts: 31056
Free Member
 

What's a half between friends.


 
Posted : 03/12/2015 11:32 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

My brain hurts. If you're wondering, the real problem is about drilling oil wells. If you plan to drill 8 wells, what is the chance you drill exactly one good one (black ball)


 
Posted : 03/12/2015 11:41 pm
Posts: 0
Free Member
 

So are the red balls oil wells which explode?


 
Posted : 03/12/2015 11:44 pm
Posts: 0
Free Member
 

[quote=DrJ ]OK smart guys ...
3 red balls, one black, 4 white
Prob of 3 red balls and 5 white
Prob of 1 black 7 white

a) (3/8)^3 * (4/8)^5 * 8! / (3! * 5!) = 9.2%
b) 1/8 * (4/8)^7 * 8! / (7! * 1!) = 0.78%

off the top of my head - I think that's right, but only about 95% certain 😉


 
Posted : 03/12/2015 11:47 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

Red balls explode your career


 
Posted : 03/12/2015 11:53 pm
 poly
Posts: 8699
Free Member
 

Dr J,

It was too late at night to do it the hardway so I just wrote half a dozen lines of code to simulate it and test it and the answer is about 39% of the time.

Now I'm not sure your drilling oil wells is the same? Is there a 1/8 chance of hitting the sweet spot? would you keep going to eight wells if you hit one of the early ones? it wouldn't be a bad thing if you got more than 1 "hit".


 
Posted : 03/12/2015 11:57 pm
Posts: 43345
Full Member
 

[quote=DrJ ]My brain hurts. If you're wondering, the real problem is about drilling oil wells. If you plan to drill 8 wells, what is the chance you drill exactly one good one (black ball)
Depends if there's a Paveway with your name on it


 
Posted : 04/12/2015 12:00 am
 Spin
Posts: 7655
Free Member
 

says the one who has to ask the internet

Says the one who clearly has no clue but thought he'd post an answer anyway.


 
Posted : 04/12/2015 7:38 am
Posts: 3729
Free Member
 

My brain hurts. If you're wondering, the real problem is about drilling oil wells. If you plan to drill 8 wells, what is the chance you drill exactly one good one (black ball)

Given that fact that you will have gotten your information from a geologist who in reality won't actually have a clue (despite what they might tell you) then there is no way of telling.


 
Posted : 04/12/2015 9:34 am
Posts: 0
Free Member
 

[pedantic semantics]
This is probability not statistics 😉
[/pedantic semantics]


 
Posted : 04/12/2015 9:38 am
 DrJ
Posts: 13416
Full Member
Topic starter
 

Given that fact that you will have gotten your information from a geologist who in reality won't actually have a clue (despite what they might tell you) then there is no way of telling.

True dat 🙂


 
Posted : 04/12/2015 3:31 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

aracer - you're my hero 🙂


 
Posted : 04/12/2015 3:32 pm
Posts: 0
Free Member
 

[quote=jambalaya ]This is probability not statistics

Yeah, but given input from a geologist, the results bear a lot of similarity with statistics.


 
Posted : 04/12/2015 3:59 pm
Posts: 7433
Free Member
 

I'll give you the correct answer, guaranteed, if you PPG me a tenner. I used to do these sort of things for fun (back when I had no bills to pay).


 
Posted : 04/12/2015 4:16 pm
 DrJ
Posts: 13416
Full Member
Topic starter
 

No problem captain. How much do you charge for advice on tyres?


 
Posted : 05/12/2015 11:42 am
 DrJ
Posts: 13416
Full Member
Topic starter
 

Now I'm not sure your drilling oil wells is the same? Is there a 1/8 chance of hitting the sweet spot? would you keep going to eight wells if you hit one of the early ones? it wouldn't be a bad thing if you got more than 1 "hit".

You're right, poly - the situation was that someone said that if 1/8 of the spots are sweet, drilling 8 wells would guarantee hitting one, which raised a red flag in my head. It seems to be true that hitting one and only one is the most likely outcome, but it's far from guaranteed.


 
Posted : 05/12/2015 11:45 am
Posts: 0
Free Member
 

[quote=DrJ ]You're right, poly - the situation was that someone said that if 1/8 of the spots are sweet, drilling 8 wells would guarantee hitting one, which raised a red flag in my head. It seems to be true that hitting one and only one is the most likely outcome, but it's far from guaranteed.

Ah, but that's a slightly different question. The probability of hitting at least one is 66% = 1 - (7/8)^8

Though your scepticism does show that you have a rather better grasp of probability than the person who said that (which to be honest is what I'd expect most people to think).


 
Posted : 05/12/2015 9:26 pm

6 DAYS LEFT
We are currently at 95% of our target!