Ever wondered how the production evaluations work?
It’s always been a bit of a mystery to me.
It’s always been a bit of a mystery to me.
12/7/22:
So I have been breeding/raising goats for almost 8 years and on milk test for the last 6 years. We have been milk testing YEAR ROUND since April 2017 (5 years.) I have studied hundreds of CDCB milk records, and followed the elite and top ten lists to understand how to breed for production. But a week ago, it finally clicked.
Let me preface by saying, I am sure I do not understand everything about production evaluations. I do not have a CDCB published source to back me up. I know I have a lot more to learn. But I wanted to share what I can see happening IN MY OWN HERD after 6 years of milk test. I also want to thank many breeders for teaching me all this and answering the same questions over and over again lol (Alexis Bachrach/Redstone Dairy Goats.)
A few basics:
Production evaluations take the milk test day raw data breeders put into the system and try to “rank” all the does on milk test in the nation. You will see a doe in the 17th percentile and another in the 97th percentile. 17th percentile means that out of 100 does, your doe should milk better than 16 does, but the other 83 does milk better than her. 97th percentile means that out of 100 does, your doe milks better than 96 of them.
To help us see (about) how many goats are on milk test with qualifying lactations, I counted how many does were on the elite list (that’s the top 5% of does.) There are 20 does per page and 24 pages and a last page of 3 does= 483 does on the elite list or top 5% of Nigerian dwarf does on test. Which means there should be about 9,660 Nigerian dwarf does with production evaluations. I also counted 82 does in the 99th percentile. So this is totally different from the “top ten” list where there are only 10 does on the list and if two tie, then there are only 9 places.
I would also like to note that when you look up the “Elite Doe Current” list on adgagenetics, it shows the does in an “order” which causes some people to mistakenly think that their doe (or buck) is #2 in the nation. As far as I understand, the 82 does in the 99th percentile are not ranked against each other and so that order changes all the time. Your doe may be #2 today and #40 tomorrow (someone correct me if I’m wrong here.)
This evaluation doesn’t just look at raw milk weights either. The system favors high butterfat and even more favors high protein. And this is not %, this is daily lbs of fat produced. A doe who milks 1 lb with 9% butterfat is actually not making as much fat per day as a doe doing 6 lbs with 6%. You can also look at total lbs of fat produced during a 305 day total: 65 lbs vs. 95 lbs of butterfat.
It also ranks them against their herd-mates to try to account for management. Maybe you feed grass hay and milk once a day with average temps during milking season being in the 100s, but a herd on the other side of the country feeds alfalfa hay, milks twice a day, and has a moderate climate. Obviously, the raw data for the 1st herd would be below the raw data in the 2nd herd. But in each herd, what if one doe milks way better than her herd-mates? And what if those does are actually close relatives? Genetics!
So the evaluations are trying to see past weather, feed, management, etc to see which genetics actually produce the most milk and components in the breed.
However, you can have a high producing doe with an amazing milk record that is ranked 70th percentile. And a doe with a crummy milk record who is 94%. What?! I saw this all the time and was so confused. Does it mean that the 94% doe is really better? I should use a buck out of her? Or that I should sell my doe ranked 70th?
This happens because the evaluations are not ranking each individual doe. They are ranking families. The system doesn’t just want a high producing doe. It wants a whole huge family of does who all milk well and better than their herd-mates in multiple herds all over the country.
The exciting part:
Basically, production evaluations are trying to identify “dynasties.” The definition of dynasty is “a succession of people from the same family who play a prominent role in politics, or their field.” Another definition said at least 4 generations.
Yes! The evaluations want 4+ generations of high producing does that are all related and consistently producing offspring who are producing well.
Okay, so the really exciting part of this is that it happened in my herd! I made my own baby dynasty! And I’m going to share the secret of how this happened.
Most does who are ranked high are part of the Promisedland/Sugar Moon Dynasty or the Rosasharn Dynasty or the ______________ Dynasty, etc. These are huge dynasties that involve a lot of data in many, many different herds over many generations and are building on each other’s high rankings. If your doe milks well and her dam milks well and great aunt milked well and her daughters milk well, the system ranks them all higher. If your doe milks well, but her granddam doesn’t and her sister doesn’t, then it ranks your doe lower. And it can be really hard to climb up to the top percentiles if your genetics start at the bottom.
For example, Rye is part of the Sugar Moon Dynasty (sire) and the Rosasharn Dynasty (dam), plus she and her dam milk well. Rye is currently ranked 95%.
However, right now, my does (born 2014-2020) do not have any relatives ranked higher than them that are not in my herd, Redstone Farm, or Bridgeport Farms. That is 3 herds. Now, I would like to take a minute to point out that I did not create these genetics. Many breeders who have been breeding much longer than I pulled these genetics together through careful breeding and line-breeding (Dragonfly Farm is one good example.) The genetics were already there when I bought them. I just put tons of data into the system (through year round milk tests.) And I repeatedly selected for high production. It took literally years of milk testing.
You can now look at my foundation does and accurately see why they are ranked the way they are. See my chart!
In the chart I made of my herd, it shows 4 generations. If you see the ranking go up in the next generation, that doe or her daughter milks better than her dam (in general.) If it drops, she milks worse than her dam.
My chart shows the percentile for each doe and then their 305 total lactations with butterfat and protein. P is for Predicted 305 lactations. Red highlighted lactations show how their predictions changed drastically from August to December. This shows why we don’t use predicted numbers for advertising. They change with each milk test. However, I wanted you to be able to easily compare each relative and some of them do not have completed 305s.
So I have been breeding/raising goats for almost 8 years and on milk test for the last 6 years. We have been milk testing YEAR ROUND since April 2017 (5 years.) I have studied hundreds of CDCB milk records, and followed the elite and top ten lists to understand how to breed for production. But a week ago, it finally clicked.
Let me preface by saying, I am sure I do not understand everything about production evaluations. I do not have a CDCB published source to back me up. I know I have a lot more to learn. But I wanted to share what I can see happening IN MY OWN HERD after 6 years of milk test. I also want to thank many breeders for teaching me all this and answering the same questions over and over again lol (Alexis Bachrach/Redstone Dairy Goats.)
A few basics:
Production evaluations take the milk test day raw data breeders put into the system and try to “rank” all the does on milk test in the nation. You will see a doe in the 17th percentile and another in the 97th percentile. 17th percentile means that out of 100 does, your doe should milk better than 16 does, but the other 83 does milk better than her. 97th percentile means that out of 100 does, your doe milks better than 96 of them.
To help us see (about) how many goats are on milk test with qualifying lactations, I counted how many does were on the elite list (that’s the top 5% of does.) There are 20 does per page and 24 pages and a last page of 3 does= 483 does on the elite list or top 5% of Nigerian dwarf does on test. Which means there should be about 9,660 Nigerian dwarf does with production evaluations. I also counted 82 does in the 99th percentile. So this is totally different from the “top ten” list where there are only 10 does on the list and if two tie, then there are only 9 places.
I would also like to note that when you look up the “Elite Doe Current” list on adgagenetics, it shows the does in an “order” which causes some people to mistakenly think that their doe (or buck) is #2 in the nation. As far as I understand, the 82 does in the 99th percentile are not ranked against each other and so that order changes all the time. Your doe may be #2 today and #40 tomorrow (someone correct me if I’m wrong here.)
This evaluation doesn’t just look at raw milk weights either. The system favors high butterfat and even more favors high protein. And this is not %, this is daily lbs of fat produced. A doe who milks 1 lb with 9% butterfat is actually not making as much fat per day as a doe doing 6 lbs with 6%. You can also look at total lbs of fat produced during a 305 day total: 65 lbs vs. 95 lbs of butterfat.
It also ranks them against their herd-mates to try to account for management. Maybe you feed grass hay and milk once a day with average temps during milking season being in the 100s, but a herd on the other side of the country feeds alfalfa hay, milks twice a day, and has a moderate climate. Obviously, the raw data for the 1st herd would be below the raw data in the 2nd herd. But in each herd, what if one doe milks way better than her herd-mates? And what if those does are actually close relatives? Genetics!
So the evaluations are trying to see past weather, feed, management, etc to see which genetics actually produce the most milk and components in the breed.
However, you can have a high producing doe with an amazing milk record that is ranked 70th percentile. And a doe with a crummy milk record who is 94%. What?! I saw this all the time and was so confused. Does it mean that the 94% doe is really better? I should use a buck out of her? Or that I should sell my doe ranked 70th?
This happens because the evaluations are not ranking each individual doe. They are ranking families. The system doesn’t just want a high producing doe. It wants a whole huge family of does who all milk well and better than their herd-mates in multiple herds all over the country.
The exciting part:
Basically, production evaluations are trying to identify “dynasties.” The definition of dynasty is “a succession of people from the same family who play a prominent role in politics, or their field.” Another definition said at least 4 generations.
Yes! The evaluations want 4+ generations of high producing does that are all related and consistently producing offspring who are producing well.
Okay, so the really exciting part of this is that it happened in my herd! I made my own baby dynasty! And I’m going to share the secret of how this happened.
Most does who are ranked high are part of the Promisedland/Sugar Moon Dynasty or the Rosasharn Dynasty or the ______________ Dynasty, etc. These are huge dynasties that involve a lot of data in many, many different herds over many generations and are building on each other’s high rankings. If your doe milks well and her dam milks well and great aunt milked well and her daughters milk well, the system ranks them all higher. If your doe milks well, but her granddam doesn’t and her sister doesn’t, then it ranks your doe lower. And it can be really hard to climb up to the top percentiles if your genetics start at the bottom.
For example, Rye is part of the Sugar Moon Dynasty (sire) and the Rosasharn Dynasty (dam), plus she and her dam milk well. Rye is currently ranked 95%.
However, right now, my does (born 2014-2020) do not have any relatives ranked higher than them that are not in my herd, Redstone Farm, or Bridgeport Farms. That is 3 herds. Now, I would like to take a minute to point out that I did not create these genetics. Many breeders who have been breeding much longer than I pulled these genetics together through careful breeding and line-breeding (Dragonfly Farm is one good example.) The genetics were already there when I bought them. I just put tons of data into the system (through year round milk tests.) And I repeatedly selected for high production. It took literally years of milk testing.
You can now look at my foundation does and accurately see why they are ranked the way they are. See my chart!
In the chart I made of my herd, it shows 4 generations. If you see the ranking go up in the next generation, that doe or her daughter milks better than her dam (in general.) If it drops, she milks worse than her dam.
My chart shows the percentile for each doe and then their 305 total lactations with butterfat and protein. P is for Predicted 305 lactations. Red highlighted lactations show how their predictions changed drastically from August to December. This shows why we don’t use predicted numbers for advertising. They change with each milk test. However, I wanted you to be able to easily compare each relative and some of them do not have completed 305s.
SGCH Redstone C Mimolette is currently 96% and SG Dragonfly CP Tumacacori “Cori” is 96%. Mimolette’s daughter and Cori’s granddaughter Caprice is 99% (with an amazing personal milk record.) Caprice’s full sister Chocolat is ranked 97% with a commendable steady 4+ lb/day extended lactation. Caprice and Chocolat’s sire (D’Ambert) is ranked 99th percentile. Rouelle is a Mimolette granddaughter in the 97th percentile. Filetta is a Rouelle daughter sired by Cori’s other son also ranked 98th percentile. Both does came up to 4+ lbs at the end of their lactations. Rouelle peaked at 6.1 lbs. Gourmandise is another Mimolette granddaughter through a different daughter, ranked 99%. There are several other notable relatives, including Mimolette and Cori’s dams and sisters and all of D’Ambert’s daughters.
But the point is, there are your 4+ generations of high producing, closely related does. And they are not part of any other dynasty or riding on any other milk records besides their own family.
Working with other herds:
Now, let’s talk about the importance of working together with other breeders. The system will not give a buck an evaluation without at least one daughter on milk test in another herd, so 5 daughters in 2 herds. I literally could not have done this on my own. There have to be at least 2 herds involved. And the more herds, the better. Karen Coutu with Bridgeport Farms put Caprice on milk test for the last 2 years. And milked her twice a day through the winter. She did the same with Caprice’s daughter Tiffany. Alexis Bachrach with Redstone Dairy Goats has been breeding, retaining and doing milk test for many years with many of Mimolette’s close relatives, including her full brother Redstone Starship Trooper, her dam SG Redstone SF Dark Star, her maternal half sister SG Redstone Pony Bellatrix and many other relatives. Joanne Karohl with Dragonfly Farm has production evaluations on several of Tumacacori's close relatives, including her dam, full sister, much of her herd, etc.
I do not have a production evaluation for any of my other bucks: Jude, Barbeillon, and Gorgonzola do not have enough daughters on test or enough daughters on test in other herds to get an evaluation. Giulianni you will notice does not yet have any daughters in other herds on milk test but will next year. We are all excited to see his evaluation next year!
The rankings are constantly changing (new rankings are released twice a year: Aug & Dec) depending on how everyone is CURRENTLY milking. That means that most of my does were not ranked super high in previous years. D’Ambert was 88th percentile, then 90th, 97th in August, and now 99th percentile as his daughters got more mature lactations on record.
And our top ten doe Zoella was ranked in the 99th percentile right after she made the top ten list in 2016. Zoella is also part of the Promisedland/Sugar Moon Dynasty and has many relatives at the top. But in recent years, her production hasn’t been fabulous (especially as she gets older.) But more importantly, her offspring have not been milking very well. She has 3 daughters (Zelu, Zufi, Zola) in my herd who are being way out-milked by Mimolette x Cori offspring. And her sons don’t have a lot of daughters or sisters, but certainly none who are doing well (except Gourmandise.) Zoella is currently ranked 80th percentile.
Now, where does Rouelle’s production come from or Chateau’s or Tamie’s? Their dams consistently milk worse than they do. In Tamie’s case, her dam and her granddam milk worse. Genetics are more like dice rolls. If you start with high producing genetics that are consistently ranked high, you are more likely to see high producing offspring. However, sometimes the genes a doe carries (genotype) are not seen in herself (phenotype is the genetics we can see.) So she can have a daughter who milks better than she does. So in general, you may actually have offspring milk better out of a doe on the elite list with a crummy milk record. However, if the lower ranking doe is just ranked low because she doesn’t have enough relatives on test, then you may actually want her son and plan to prove out lots of relatives to see how predictable is it.
But the point is, there are your 4+ generations of high producing, closely related does. And they are not part of any other dynasty or riding on any other milk records besides their own family.
Working with other herds:
Now, let’s talk about the importance of working together with other breeders. The system will not give a buck an evaluation without at least one daughter on milk test in another herd, so 5 daughters in 2 herds. I literally could not have done this on my own. There have to be at least 2 herds involved. And the more herds, the better. Karen Coutu with Bridgeport Farms put Caprice on milk test for the last 2 years. And milked her twice a day through the winter. She did the same with Caprice’s daughter Tiffany. Alexis Bachrach with Redstone Dairy Goats has been breeding, retaining and doing milk test for many years with many of Mimolette’s close relatives, including her full brother Redstone Starship Trooper, her dam SG Redstone SF Dark Star, her maternal half sister SG Redstone Pony Bellatrix and many other relatives. Joanne Karohl with Dragonfly Farm has production evaluations on several of Tumacacori's close relatives, including her dam, full sister, much of her herd, etc.
I do not have a production evaluation for any of my other bucks: Jude, Barbeillon, and Gorgonzola do not have enough daughters on test or enough daughters on test in other herds to get an evaluation. Giulianni you will notice does not yet have any daughters in other herds on milk test but will next year. We are all excited to see his evaluation next year!
The rankings are constantly changing (new rankings are released twice a year: Aug & Dec) depending on how everyone is CURRENTLY milking. That means that most of my does were not ranked super high in previous years. D’Ambert was 88th percentile, then 90th, 97th in August, and now 99th percentile as his daughters got more mature lactations on record.
And our top ten doe Zoella was ranked in the 99th percentile right after she made the top ten list in 2016. Zoella is also part of the Promisedland/Sugar Moon Dynasty and has many relatives at the top. But in recent years, her production hasn’t been fabulous (especially as she gets older.) But more importantly, her offspring have not been milking very well. She has 3 daughters (Zelu, Zufi, Zola) in my herd who are being way out-milked by Mimolette x Cori offspring. And her sons don’t have a lot of daughters or sisters, but certainly none who are doing well (except Gourmandise.) Zoella is currently ranked 80th percentile.
Now, where does Rouelle’s production come from or Chateau’s or Tamie’s? Their dams consistently milk worse than they do. In Tamie’s case, her dam and her granddam milk worse. Genetics are more like dice rolls. If you start with high producing genetics that are consistently ranked high, you are more likely to see high producing offspring. However, sometimes the genes a doe carries (genotype) are not seen in herself (phenotype is the genetics we can see.) So she can have a daughter who milks better than she does. So in general, you may actually have offspring milk better out of a doe on the elite list with a crummy milk record. However, if the lower ranking doe is just ranked low because she doesn’t have enough relatives on test, then you may actually want her son and plan to prove out lots of relatives to see how predictable is it.
Breeding/culling choices:
I should add that until this year, I made breeding/culling choices based on raw data from their milk tests. There are things a computer cannot see like a single kid lactation, an illness, show circuit or injury, grain/mineral changes. Something like mastitis can drastically affect her production but the computer doesn’t know what happened. It can only work with the data you put in. Those are things only a herd owner would know.
Looking at milk tests, I asked:
I should add that until this year, I made breeding/culling choices based on raw data from their milk tests. There are things a computer cannot see like a single kid lactation, an illness, show circuit or injury, grain/mineral changes. Something like mastitis can drastically affect her production but the computer doesn’t know what happened. It can only work with the data you put in. Those are things only a herd owner would know.
Looking at milk tests, I asked:
- Who peaked the highest?
- Who held 90% of their peak (4-6 lbs) past 100 DIM?
- Who came up in the spring and finished their lactation at 90% peak?
- How low did they drop? (3 lbs vs. 1 lb)
- Who bounced back after shows or illness?
- How many kids did each doe have?
- Who had the highest totals? Best total fat? Best total protein? (here you might use predicted totals if they don’t get a full 305.)
- Who held 90% of their peak (4-6 lbs) past 100 DIM?
- Who came up in the spring and finished their lactation at 90% peak?
- How low did they drop? (3 lbs vs. 1 lb)
- Who bounced back after shows or illness?
- How many kids did each doe have?
- Who had the highest totals? Best total fat? Best total protein? (here you might use predicted totals if they don’t get a full 305.)
Another point: I was well aware when I brought in Giulianni that his parents and full sister had high production rankings. Honestly, I was looking closer at their raw data. Gianna (his dam) peaked at 6.8 lbs. She milked 5.3 lbs at 198 DIM. At 300 DIM, she milked 3.6 lbs! Giulianni’s sire also had a dam and daughters with good production (raw data and percentile rankings.)
If you look at the August chart vs. the December chart, you will notice the impact Giulianni’s line had on his daughters who are now mixed with my herd. However, they still accurately indicate the current best milkers in my herd. Chateau has a beautiful ff lactation in progress and her ranking pulled up her dam and granddam’s rankings. As more milk tests and lactations are added, these rankings will continue to change.
If you look at the August chart vs. the December chart, you will notice the impact Giulianni’s line had on his daughters who are now mixed with my herd. However, they still accurately indicate the current best milkers in my herd. Chateau has a beautiful ff lactation in progress and her ranking pulled up her dam and granddam’s rankings. As more milk tests and lactations are added, these rankings will continue to change.
Feel free to browse the raw data milk records of our highest ranked does:
Noall's Ark Caprice Des Dieux- 99th percentile:
Noall's Ark N Filetta- 98th percentile:
SG Noall's Ark J Rouelle - 97th percentile:
SG Noall's Ark D'A Chocolat- 97th percentile:
(look at how much she is milking at over 305 days in milk!)
Noall's Ark G Gourmandise- 99th percentile: