| Variation - bell shaped curve | |
|
+10RobertMac PatB Grassfarmer Tom D Mean Spirit EddieM Kent Powell Hilly MKeeney df 14 posters |
|
Author | Message |
---|
df
Posts : 521 Join date : 2010-09-28
| Subject: Variation - bell shaped curve Thu Jun 23, 2011 8:00 pm | |
| In statistics, the variance is a measure of variation. The square root of variance is the standard deviation. As the number of observations increases, the variance goes DOWN as the middle of a bell-shaped curve gets filled in. A small number of observations (typically) results in a large variance. | |
|
| |
MKeeney Admin
Posts : 3797 Join date : 2010-09-21
| Subject: Re: Variation - bell shaped curve Thu Jun 23, 2011 9:03 pm | |
| - df wrote:
- In statistics, the variance is a measure of variation. The square root of variance is the standard deviation. As the number of observations increases, the variance goes DOWN as the middle of a bell-shaped curve gets filled in. A small number of observations (typically) results in a large variance.
Scientist Leonhardt determined that all variation is actually a sphere...just ask Hilly for the explanation... the variance may go down technically, but the range of variation will almost certainly increase with more observations...thus, the old saying, "you ain`t seen nothing yet" | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 10:20 am | |
| Does the sphere illustrate one trait at a time or is it all traits of interest at once? | |
|
| |
Hilly
Posts : 368 Join date : 2010-09-24 Location : Sylvan Lake, Alberta
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 11:44 am | |
| - df wrote:
- Does the sphere illustrate one trait at a time or is it all traits of interest at once?
Not sure if I understand the “of interest" part, but my interpretation is that it represents all traits in the bovine species. And that isolated sub populations would differ by sequential reduction of peripheral distributions as they concentrate their purposeful function. Tom D can translate as I may have used the wrong words | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 12:23 pm | |
| My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? | |
|
| |
Kent Powell
Posts : 441 Join date : 2010-09-24 Location : SW Kansas
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 1:59 pm | |
| The model is on this page: http://www.keeneyscorner.com/t34p315-reflections-from-llThe page with the spherical distribution chart was given to me with the Tru-Line booklet and an old sale catalog on my one visit to Cowley, Wyoming and did more for my understanding of animal breeding and selection than 4-H, FFA, Judging Scholorships, reading all possible books and magazines, and 4 institutions of Higher Tuition combined. | |
|
| |
EddieM
Posts : 632 Join date : 2010-09-24 Location : South Carolina
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 3:09 pm | |
| Did I miss something or what prompted this thread? Just trying to figure out whether I need to act like I know something or not! | |
|
| |
MKeeney Admin
Posts : 3797 Join date : 2010-09-21
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 4:18 pm | |
| - df wrote:
- My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? how many traits are not related to other traits? if they are related, how can you study or graph just one?
Last edited by MKeeney on Fri Jun 24, 2011 4:46 pm; edited 1 time in total | |
|
| |
Hilly
Posts : 368 Join date : 2010-09-24 Location : Sylvan Lake, Alberta
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 4:41 pm | |
| | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 6:23 pm | |
| - MKeeney wrote:
- df wrote:
- My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? how many traits are not related to other traits? if they are related, how can you study or graph just one? In three ways; one is class variables such as sex, as we know males and females are different, the second is by fitting a covariate which is a continuous variable such as weaning wt to predict yearling wt and three, a combination of the two. The limitation in the past (and to a certain extent today) is computing power. The spheres only account for three traits, of which I believe they were phenotypes; is that an appropriate method? Are there not more traits of intesest? Are there any adjustments for age of cattle, year of birth, body condition score, etc? How do you determine if that point is "good" or "bad"? How do you know an individual animal won't produce offspring that land outside of the current sphere? How accurately can you predict the female will produce offspring on that same dot? The "you" is not directed at anybody or with malice. I have questions running through my head and then they appear on the screen. | |
|
| |
MKeeney Admin
Posts : 3797 Join date : 2010-09-21
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 6:32 pm | |
| - df wrote:
- MKeeney wrote:
- df wrote:
- My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? how many traits are not related to other traits? if they are related, how can you study or graph just one? In three ways; one is class variables such as sex, as we know males and females are different, the second is by fitting a covariate which is a continuous variable such as weaning wt to predict yearling wt and three, a combination of the two. The limitation in the past (and to a certain extent today) is computing power.
The spheres only account for three traits, of which I believe they were phenotypes; is that an appropriate method? Are there not more traits of intesest? Are there any adjustments for age of cattle, year of birth, body condition score, etc? How do you determine if that point is "good" or "bad"? How do you know an individual animal won't produce offspring that land outside of the current sphere? How accurately can you predict the female will produce offspring on that same dot?
The "you" is not directed at anybody or with malice. I have questions running through my head and then they appear on the screen. good stuff df...can`t answer off the cuff | |
|
| |
EddieM
Posts : 632 Join date : 2010-09-24 Location : South Carolina
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 7:28 pm | |
| - Quote :
- Come on Eddie give us a hand... its hard work making the simple seem complicated.
Hilly, no hard work on my part, it comes naturally with ease! I've looked at all of the pages of a book on the subject of the human genone and if folks don't know me very well I tell them I "read" it! But there is quite a bit about genetic "noise" and constant DNA flaws and generational mutations. I would think that these types of genetic data transfer flaws would keep the base width of the curve intact and possiply would broaden it as the number of generations increase. The book did not discuss phenotypic damage done with sorting posts! If just a portion of the outliers are from a number of new mutations then that would be reason enough to move back to the middle for stablity. | |
|
| |
Hilly
Posts : 368 Join date : 2010-09-24 Location : Sylvan Lake, Alberta
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 9:44 pm | |
| - df wrote:
- MKeeney wrote:
- df wrote:
- My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? how many traits are not related to other traits? if they are related, how can you study or graph just one? In three ways; one is class variables such as sex, as we know males and females are different, the second is by fitting a covariate which is a continuous variable such as weaning wt to predict yearling wt and three, a combination of the two. The limitation in the past (and to a certain extent today) is computing power.
The spheres only account for three traits, of which I believe they were phenotypes; is that an appropriate method? Are there not more traits of intesest? Are there any adjustments for age of cattle, year of birth, body condition score, etc? How do you determine if that point is "good" or "bad"? How do you know an individual animal won't produce offspring that land outside of the current sphere? How accurately can you predict the female will produce offspring on that same dot?
The "you" is not directed at anybody or with malice. I have questions running through my head and then they appear on the screen. I would think, as they are spheres and we only see a flat circle that there could be as many antagonistic traits as you can come up with at different locations on or in the sphere, but whether or not there are more or less is irrelevant to me as it was simply a way to help illustrate basic under lying principals. Regardless of what you want to measure with the bell curve, we don’t have to look far to see the damage measuring can cause without understanding the underlying principals, however we can see benefits to these same measurements once even limited understanding of the more basic principles is achieved. "Give a man a fish; feed him for a day. Teach a man to fish; feed him for a lifetime” ....Lao Tzu It seems to me there is a fair bit of teaching that needs to be done before we can utilize more numbers properly. As a newbie and a second class commercial producer, when I read a majority of the cattle “breeder” forums I have been surprised in many ways at the shallowness of understanding, not that I’m smarter or that I can do it better, I just don’t feel dumber anymore | |
|
| |
Mean Spirit
Posts : 288 Join date : 2010-09-26
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 10:08 pm | |
| - EddieM wrote:
- Did I miss something or what prompted this thread? Just trying to figure out whether I need to act like I know something or not!
Never hurts.... just jump in. Fake it til you make it. | |
|
| |
Tom D Admin
Posts : 443 Join date : 2010-09-25 Age : 45 Location : Michigan
| Subject: Re: Variation - bell shaped curve Fri Jun 24, 2011 10:33 pm | |
| - Hilly wrote:
- As a newbie and a second class commercial producer, when I read a majority of the cattle “breeder” forums I have been surprised in many ways at the shallowness of understanding, not that I’m smarter or that I can do it better, I just don’t feel dumber anymore
Hell, I'll say it... you ARE smarter and you ARE doing it better. That's why you're HERE. | |
|
| |
Grassfarmer
Posts : 660 Join date : 2010-09-27 Location : Belmont, Manitoba, Canada
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 12:53 am | |
| - Hilly wrote:
- when I read a majority of the cattle “breeder” forums I have been surprised in many ways at the shallowness of understanding.....
Isn't that the truth. Many of the concepts and truths aired here are so far above the heads of many purebred "breeders", and even many of the multi-generational purebred guys I grew up with in the old country it's not even funny. Like you Hilly I feel i've still to prove this new knowledge I'm learning in practice but boy, I'm glad to be learning it. Thanks to all the participants here for creating this learning environment. | |
|
| |
MKeeney Admin
Posts : 3797 Join date : 2010-09-21
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 3:30 am | |
| - df wrote:
- MKeeney wrote:
- df wrote:
- My bell-shaped curve is for one trait.
Under the sphere model, there is selection for many traits, but a concentration on a small subset of the population. Is this true? how many traits are not related to other traits? if they are related, how can you study or graph just one? In three ways; one is class variables such as sex, as we know males and females are different, the second is by fitting a covariate which is a continuous variable such as weaning wt to predict yearling wt and three, a combination of the two. The limitation in the past (and to a certain extent today) is computing power.
The spheres only account for three traits, of which I believe they were phenotypes; is that an appropriate method? Are there not more traits of intesest? Are there any adjustments for age of cattle, year of birth, body condition score, etc? How do you determine if that point is "good" or "bad"? How do you know an individual animal won't produce offspring that land outside of the current sphere? How accurately can you predict the female will produce offspring on that same dot?
The "you" is not directed at anybody or with malice. I have questions running through my head and then they appear on the screen. the sphere was an eye opener for me ; like to never have visulized it...hard to see a sphere on flat paper...and not really a neccessity now, because it`s firmly etched in my head...a semi-sphere | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 9:39 am | |
| I see the sphere as an index, where each animal is within that sphere is a combination of (genetic or phenotypic) values of all traits. If economic values were added, it could be a bioeconomic index. For some, the sphere is easier to see than a list of numbers in a sire summary, but it serves the same purpose, IMO.
The same visual can be made on a slide presentation as technology allows the rotation of a 3-D image.
| |
|
| |
df
Posts : 521 Join date : 2010-09-28
| |
| |
Kent Powell
Posts : 441 Join date : 2010-09-24 Location : SW Kansas
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 10:29 am | |
| - df wrote:
- I see the sphere as an index, where each animal is within that sphere is a combination of (genetic or phenotypic) values of all traits. If economic values were added, it could be a bioeconomic index. For some, the sphere is easier to see than a list of numbers in a sire summary, but it serves the same purpose, IMO.
The same visual can be made on a slide presentation as technology allows the rotation of a 3-D image.
Human nature being what it is- to maximize and increase, where more is not better, how can economic values or indexes be of value for the purposes at hand? Why have, do, or measure what is unnecessary for our purposes? | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 10:42 am | |
| - Keystone wrote:
- df wrote:
- I see the sphere as an index, where each animal is within that sphere is a combination of (genetic or phenotypic) values of all traits. If economic values were added, it could be a bioeconomic index. For some, the sphere is easier to see than a list of numbers in a sire summary, but it serves the same purpose, IMO.
The same visual can be made on a slide presentation as technology allows the rotation of a 3-D image.
Human nature being what it is- to maximize and increase, where more is not better, how can economic values or indexes be of value for the purposes at hand? Why have, do, or measure what is unnecessary for our purposes? Keystone, A very good question. What do you and your customers value? Can it be measured and if so, how is it measured? I am willing to bet all breeders practice multiple-trait selection. Indexes increase multiple-trait selection and by default, reduce single-trait selection, a practice that has proven to not work in poultry, swine and cattle. Indexes are very important to poultry and swine breeding. | |
|
| |
Mean Spirit
Posts : 288 Join date : 2010-09-26
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 11:13 am | |
| It still seems that indices, to be truly useful, must somehow account for the tendency to go for higher, bigger, more. Can't some smart math guy make an index in which the optimums are more highly valued than the maximums? Or do you think including enough traits in the index will just naturally balance things out? I guess I could imagine an index that included enough traits that do not correlate being somewhat more balanced. | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 11:34 am | |
| Which trait(s) are you concerned about that are simply maximized and not optimized? | |
|
| |
PatB
Posts : 334 Join date : 2010-09-25 Age : 60 Location : Turner, Maine
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 11:35 am | |
| - Mean Spirit wrote:
- It still seems that indices, to be truly useful, must somehow account for the tendency to go for higher, bigger, more. Can't some smart math guy make an index in which the optimums are more highly valued than the maximums? Or do you think including enough traits in the index will just naturally balance things out? I guess I could imagine an index that included enough traits that do not correlate being somewhat more balanced.
Wouldn't the optimums be different in different regions? The available resources would dictate what the opitimum is for that operation. | |
|
| |
df
Posts : 521 Join date : 2010-09-28
| Subject: Re: Variation - bell shaped curve Sat Jun 25, 2011 11:40 am | |
| Are there any traits that are priorities, regardless of environment, for profit? | |
|
| |
Sponsored content
| Subject: Re: Variation - bell shaped curve | |
| |
|
| |
| Variation - bell shaped curve | |
|