Who is good with determining odds (chances of getting a specific thing)
   

       The Horse Forum > Life Beyond Horses > General Off Topic Discussion

Who is good with determining odds (chances of getting a specific thing)

This is a discussion on Who is good with determining odds (chances of getting a specific thing) within the General Off Topic Discussion forums, part of the Life Beyond Horses category

     
    LinkBack Thread Tools
        06-26-2011, 10:32 PM
      #1
    Banned
    Who is good with determining odds (chances of getting a specific thing)

    There is an argument going on another site. Nothing to do with horses.

    The question relates to putting a ram with a pattered body that has been coded to passing this pattern on 50% of the time. The female has been coded to have no influence on the pattern of the lamb...just the colour.

    Now do not take real life genetics into the equation as these are cyber coded animals.

    Argument 1 is that you should get about 50% lambs with patterns and about 50% with no patterns.

    Argument 2 is that since you only have a 50/50 chance of getting a pattern each time you breed then the overall percentages can be anything...not an expected 50%....you could from this argument get as low as 20% with patterns or even 80%.
         
    Sponsored Links
    Advertisement
     
        06-27-2011, 12:08 AM
      #2
    Green Broke
    I'm just subscribing...
         
        06-27-2011, 12:12 AM
      #3
    Banned
    Quote:
    Originally Posted by lildonkey8    
    I'm just subscribing...

    LOL

    It is an exercise in one's luck I assume.

    Here is one person's results. Remember these are not real sheep.

    We bred the same pair of animals 12 times: the yellow ram with blue stars + violet ewe with white stars.

    Of the 12 offspring, 9 were plain. 3 were patterned: two were exact copies of the father (yellow ram with blue stars), and one was exact copy of the mother (violet with white stars).


    So the question is...should the results have been closer to 6 plain and 6 patterned?
         
        06-27-2011, 12:14 AM
      #4
    Green Broke
    Quote:
    Originally Posted by Spyder    
    LOL

    It is an exercise in one's luck I assume.

    Here is one person's results. Remember these are not real sheep.

    We bred the same pair of animals 12 times: the yellow ram with blue stars + violet ewe with white stars.

    Of the 12 offspring, 9 were plain. 3 were patterned: two were exact copies of the father (yellow ram with blue stars), and one was exact copy of the mother (violet with white stars).


    So the question is...should the results have been closer to 6 plain and 6 patterned?
    I like being quoted
         
        06-27-2011, 12:14 AM
      #5
    Showing
    50/50 patterned or not.
    I don't understand argument #2....
         
        06-27-2011, 01:05 AM
      #6
    Green Broke
    Eeek, *scratches head and tries to remember probability theory*

    If I read correctly you are talking about independant events so the result of one event doesn't change the odds of the next.

    Quote:
    Originally Posted by Spyder    
    Argument 1 is that you should get about 50% lambs with patterns and about 50% with no patterns.

    Argument 2 is that since you only have a 50/50 chance of getting a pattern each time you breed then the overall percentages can be anything...not an expected 50%....you could from this argument get as low as 20% with patterns or even 80%.
    Both statements can be true. On average, over time, you would expect the data to be distributed 50/50. But everything is a possibility, it just isn't probable. So it is true that the overall percentages could be everything but the probability of 80/20 is smaller than 50/50. But still possible.

    Quote:
    Originally Posted by Spyder    
    LOL

    It is an exercise in one's luck I assume.

    Here is one person's results. Remember these are not real sheep.

    We bred the same pair of animals 12 times: the yellow ram with blue stars + violet ewe with white stars.

    Of the 12 offspring, 9 were plain. 3 were patterned: two were exact copies of the father (yellow ram with blue stars), and one was exact copy of the mother (violet with white stars).


    So the question is...should the results have been closer to 6 plain and 6 patterned?
    Sample size isn't really big enough. If you toss a coin 10 times you might get 8 heads and 2 tails. You are far less likely to get 800 heads and 200 tails if you toss the coin 1000 times.

    Over time you would expect the data to settle closer to the 50/50 mark.
         
        06-27-2011, 01:06 AM
      #7
    Banned
    Quote:
    Originally Posted by JustDressageIt    
    50/50 patterned or not.
    I don't understand argument #2....
    LOL I can understand it if you simply had bad luck but here is argument 2...

    with a 50% probability of producing patterned EACH time you breed, you could even have a ratio of 0:10 with none coming out as patterned because what you get previous to next has no mathematical relationship.


    And...

    Each separate time you breed, you have a 50% chance of passing on the pattern. So, you could breed over 1,000 times and still get no good sheep because there's was only a 50% chance each time.
         
        06-27-2011, 01:11 AM
      #8
    Showing
    This, ladies and gentlemen, is why I hated my statistics classes with a bright, burning, evil passion.
    Strictly speaking... shouldn't each individual lamb be 50/50? Of course with a small sample size you might see a larger number of one over the other, but that doesn't skew the bare bones actual probability from 50/50 to 80/20... I don't think?
         
        06-27-2011, 01:12 AM
      #9
    Green Broke
    Quote:
    Originally Posted by Spyder    
    LOL I can understand it if you simply had bad luck but here is argument 2...

    with a 50% probability of producing patterned EACH time you breed, you could even have a ratio of 0:10 with none coming out as patterned because what you get previous to next has no mathematical relationship.


    And...

    Each separate time you breed, you have a 50% chance of passing on the pattern. So, you could breed over 1,000 times and still get no good sheep because there's was only a 50% chance each time.
    True. Possible but not probable. Too late for me to dig out the formulae though, I'm going to bed!
         
        06-27-2011, 06:41 AM
      #10
    Showing
    With the BIG number of bred lambs it'll be close to 50/50%. You won't go around statistics doesn't matter what unless other factors come in play (which is not according to the rules of the game). Now, if you speak of only 10 lambs (or even 20 or 30) statistically it's just NOTHING. So yes, you can get 20/80%, but it doesn't prove anything because of the statistical error (which is HUGE in such a small sample).
         

    Thread Tools

    Similar Threads
    Thread Thread Starter Forum Replies Last Post
    Too much of a good thing kevinshorses Horse Talk 36 03-17-2010 11:31 PM
    Daily Worming? Not a good thing?? 7Ponies Horse Health 3 07-28-2009 07:03 PM
    Chances Diet Good or bad? HorsesAreForever Horse Health 16 12-26-2008 10:29 PM



    All times are GMT -4. The time now is 03:16 PM.


    Powered by vBulletin® Version 3.8.5
    Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
    Search Engine Friendly URLs by vBSEO 3.6.0