Breeding Forum - Inbreeding, Linebreeding, Outcrossing

Almost all discussion about duplications stops at the sixth generation (2+4+8+16+32+64 = 126 ancestors). Often discussion ends at the fifth generation.

Because it is too difficult to write out 7 or 8 or 9 or 10 or more generations on paper or display them on a computer screen.

I am getting a bit obscure here but when you add an extra generation e.g. go from a 5 gen pedigree to a 6 gen pedigree the duplication groups can change, and the same going to a 7 gen pedigree from a 6 gen pedigree, and so on.

The correct way (my opinion) to identify duplications if to take the ancestor with year of birth closest to the subject horse and look for any duplication of that horse. Then move to the next closest ancestor by year of birth and repeat.

Of course analysing pedigrees by identifying duplications is nothing more than a short-hand way of reducing a pedigree to a handful of ancestors.

A while back I wrote a program to identify the duplications in 10 gen pedigrees (2+4+8+16+32+64+128+256+512+1024 = 2046 ancestors).

Often people treat duplications (or inbreeding) as an indivisible unit.

The duplicated ancestor group can be males (sires) or females (dams), and there are three possibilities for each group.

A simple example is a duplication group of a sire twice in a pedigree. He is a male. The offspring of that male twice can be two sons, two daughters, or a son and a daughter.

I will post below the 10 generation duplication groups of 15 horses: 9 high rated; 6 low rated.

When I did the work the idea was to see if there were differences between the duplication groups of the high rated and the low rated horses.

If there is a difference it is that the low rated horses have more groups of duplicated sires producing sons only (highlighted in grey) close up in their pedigrees i.e. in the first 10 groups, often in the first five groups.

It would be great if a 10 generation prdigree could be reduced to one number to identify quality.

When identifying a duplicated ancestor I treated full siblings as one.

The first Nearco duplication group is St Simon (m); St Simon (m); Angelica (f) X St Simon (m); St Simon (m).

The program took the sex of the duplicated group ancestor as "f" from Angelica (the one female in the 4 male plus 1 female group). Angelica is a full sibling of St Simon.

Treve (131) - 4 groups of male ancestors producing all male offspring

Ribot (143) - 4 groups

Bella Canta (37) - 4 groups

Winx (130) - 4 groups

Be Excellent (28) - 6 groups

Beauchamp Kite (22) - 7 groups

Brigadier Gerard (143) - 4 groups

Beaulie (29) - 7 groups

Frankel (147) - 8 groups

Galileo (132) - 2 groups

Goldikova (131) - 5 groups

Becquarius (70) - 2 groups

Nearco (138) - 1 group

Northern Dancer (141) - 3 groups

I was in error with the sex of the duplicated ancestor of Nearco's first group (80% male; 20% female) so the groups immediately above could be in error.

Posted by Diomed

I have a couple of programs developed for my own use, one analyses six-generation pedogrees, the other ten-generation pedigrees.

If you can analyse one pedigree it is a short step to have a load file with thousands of pedigrees and a results file to accept the analysis of those pedigrees.

A few people ask for suggestions and I give my picks.

In November I was asked by a person in California to suggest a CA sire for a mare. I analysed the mare with all USA sires and listed a handful of what I thought would suit the mare. One CA sire that was under consideration by the mare owner was in my suggestions so it gave them more confidence.

On 31st December an Australian sent me a list of nine Karaka yearlings and asked my opinion. I like to work from the biggest possible list and discard. I analysed all the 1,030 (?) Karaka yearlings and was interested in 17. After looking at the pedigrees in more detail I sent a short list of six with six second choices. The info was passed to a trainer who bought one that was on both our lists.

I find pedigrees of horses offered at the sales are not well put together.

If I want to breed a foal I start without a mare and without a sire.

I live in Ireland. I gather Excel files from the three main breeding stock sales in Europe (IRE; GB; FR) and complete the mare pedigrees to five generations (and that automatically gives me ten or more generations.)

Then I prepare pedigrees of all the sires at stud in Europe, about 500. This is more difficult as new sires go to stud, and other sires leave.

What I am looking for is the biggest possible sample of sire and dam combinations.

I December 2023 I compared 2,308 European mares on offer with 497 sires at stud, or 1,147,076 "foals" / test-matings, 2,308 x 497).

My program counts colt factors and filly factors.

See Clive Harper's 1997 book, The Thoroughbred Breeders' Handbook, or read foalmare.com for a list of these factors.

I want as many factors as possible in the foal (the maximum is 8).

We do not know the sex of the foal in advance so try to plan a pedigree that will produce a useful horse of either sex i.e with as many colt factors and as many filly factors as possible.

I discard many horses (pedigrees) with this ready reckoner method, and obviously throw out many that could be good.

When I have what I consider the better pedigrees I look at each on screen in TesioPower in great detail. These might be less than one tenth of one percent of the original sample.

At this stage I look at the 7th, 8th, 9th, 10th 11th 12th generations, comparing the sire side with the dam side of the "foal".

There are weaknesses with this method.

Many good horses do not have good matching in the first six generations e.g. Via Sistina, but have matching elsewhere, often the sire and dam dam lines.

Over the last 50 years with fewer stallions a stud serving bigger books good pedigrees are becoming rare. The result imo is more slow horses.

This trend might appear to be a negative but I look on it as a positive if you can analyse pedigrees.

You can breed a better horse (although it is very difficult), and you can avoid the good looking, good conformation slow horses offered at the sales.

In future I will try to add in a comparison of the dams on the sire line with the dams on the dam line as I believe good horses can have links here, often but not always the 4th dam of the sire and the 4th dam of the dam.

My analysis is an inbreeding analysis (I prefer the term duplications.)

Very little dam racing ability is inherited in her foal. My work says about 25% of the figure above the industry average of about 70 i.e. a dam rated 110 tends to have foals rated 80 [70+25% of (110-70)]. Professor Patrick Cunningham estimated it at 35%. Either way foals do not inherit black type.

The best predictor of success in a foal (my opinion) is full siblings in its sire and dam in the 4th, 5th or 6th generation (4th and 5th best). Better still if they are full sibling sister and brother. Better again if the sex of the offspring of the full siblings is the opposite sex to the full sibling i.e. a female FS producing a son, a male FS producing a daughter. Better again if there are more than two full siblings in the six generations.

Commentators say the great breeder Federici Tesio never revealed the secrets of his success, yet in his book Breeding The Racehorse, Chapter 6, Breeding For Speed, he gives three requirements: inbreeding; nicks (which is inbreeding); the best quality parents. Then Tesio spends a few pages talking about full sibling brothers and sisters - and people say he revealed nothing.

I'm a lucky one, who has Clive Harpers "The Thoughbred Breeder's Handbook" in my privat colletion of Horsebreeding books.

At page 22 is an explanation of so called 3/4 relatives, a term I often use in my short pedigree analysis:

...Back to horses; the term '3/4 relative', or as Englishwriters prefer; is not as simple as it appears at first. Let us consider some of the different patterns at first which typically come under this broad rubric.

(a) Horses by same sire out of mares, e.g. Aloe and Winalot - both by Son-in-Law out of mares by Gallinule

(b) Horses out of the same mare but by sons of the same sire, e.g. New Moon and Moon Priestress - both out of Selene

(c) Horses by the same sire out of daughters of the same mare, e.g. Searching and Striking - both by War Admiral out of Big Hurry and Baby League respectivly who are both daughters of La Troienne

Mr. Harper explains even more complex pedigree patterns, but for "beginners", I will make

it easier to follow. From personal expericence I know, how difficult is it to impart my knowledge.