A memoir   –   Karel Miedema

Early Years – Staples & Flip-flops

My fascination with pedigrees began in about 1985, from discussions with American geologist Elliot ‘Tim’ Miller, who’d made his fortune mining rubies in Kenya. He had settled in Cape Town and bred from a few mares, also buying the odd yearling. Among the latter Prince Florimund, who got Horse of the Year honors in 1983 after which Tim sent him to be trained by Charlie Whittingham.

Prince Florimund won a G3 and was runner-up in G2, before being returned to South Africa for stud. He proved rather unremarkable, but did produce a filly bred by Tim which became dual Broodmare of the Year, mainly through the exploits of multiple Horse of the Year Pocket Power and his Gr1 full sister. Both were sired by stallion Jet Master, whose was bred from a mating I had designed for a small breeder. Multiple Horse of the Year, Jet Master went on to become six times Champion Sire. Interesting how Tim’s and mine paths crossed that way so many years later!

Tim had done extensive research on pedigrees of major winners in Europe. At the time pedigree work was tedious, all paper files and folders, but when computer programs became more accessible in the early Eighties Tim & I (with the help of a friend) designed a computerized database showing six-generation pedigrees with duplications. That was something!

At the time we concentrated on sire influence (as was the general custom in pedigree-land). Using an English publication titled ‘Pedigrees of Leading Winners 1912-1959’, Tim noted a high incidence of specific patterns, involving duplication of sires within four generations.

These patterns were sires doubled along top and bottom lines of a pedigree, and along the central axis. Tim dubbed these ‘staples’ and ‘flip-flops’, according to the way they visibly showed up in standard pedigree boxes.

A pedigree can be given letters and numbers to indicate specific positions, with subsequent generations named A, B, C, D, etc., and the columns of top-to-bottom sires and dams 1,2,3,4, etc.

  • The first generation will be sire and dam (A1-A2), then the four grandparents (B1 to B4), then the next generations C1-C8, D1-D16, E1-E32, F1-F64 etc.
  • Notable staples were C1-C7, C1-D15, D1-D15, D1-C7.
  • Notable flip-flops C3-C5,C3-D9, D7-C5, D7-D9.

They were duplications in the same or adjacent generations, in other words.

As this thinking evolved, it turns out that staples and flip-flops in essence are the same thing.

Staple indicating Top-of-sire to Bottom-of-mare, and Flip-flop being Bottom-of-sire to Top-of-mare.

This was at a time when conventional wisdom dictated that the only sex considered in duplications was male-to-male.



In time, Tim (who was a decade older than me) had a succession of girlfriends who showed no interest in thoroughbreds and he gradually lost interest & touch.
I went my own way, searching for extensions of Tim’s original thought.

That led to the notion of duplicating ancestors who were related through combinations of their own ancestors.
- Full siblings, half siblings, and the like.
- Such related ancestors could be male as well as female.
- That was a whole new way of looking at things. I termed it Kin-Breeding.

Much research went into it, including regular write-ups with predictions of what could work for stallions at the time of them going to stud.

The outcome of such predictions reinforced the belief in kin-breeding theory, although no specific, deeper research was done at the time.

More extensive research came in about 2009, as a result from meeting Charles ‘Chuck’ Fipke, who requested an evaluation of his extensive mare-band, with the stated objective of ‘breeding a Kentucky Derby winner and another Storm Cat’. Clearer it could not be.

That put the focus firmly on researching G1 winners and the way they were bred, ignoring all else.

As a result of the research the way of thinking shifted to kin-breeding through female-line ancestors, notably through successive generations, originating from a specific significant female ancestor, identified as and named SigAnc.

In other words, the focus became female rather than male.


Performance Ratings

In January 2016 I presented a paper at the Asian Racing Conference in Mumbai on the subject of the origins of the calculation of Performance Ratings, sharing a platform with professionals from Hong Kong and UK who covered different aspects of the same subject.

I’ve been hands-on involved in rating (for private use, at first) racehorse performances in Europe since the late ‘Sixties, the years in the UK of Nijinsky, Brigadier Gerard, Mill Reef – then Roberto, High Top et al. The method applied is the same one as used by English commercial rating operations Timeform (since 1948) and Raceform (part of Racing Post).

The essence of a performance rating is how a horse compares in ability to the average horse of the same age and sex.

  • Performance ratings are traditionally expressed as a number representing weight, more specifically as lbs (pounds).
  • Ratings are on a scale with 140lbs (equivalent to 10st) as a traditional top (seldomly attained) going down to zero.
  • Ratings are calculated using weight carried by horses in a race, and differences in performance using lengths.

The distribution of ratings follows a bell curve (or possibly better described as log-normal curve), with a long tail at the high end and fatter bottom at the other end.

The average rating has to be kept the same from year to year, so as to safeguard the integrity of rating successive generations. There is a difference of about five pounds between the sexes, males (as might be expected) superior to females.

Rating individual horses leads to being able to create a profile for offspring of stallions, and see how the offspring compares to that of the ‘average’ stallion. When split by sex, a further refinement can be made.

  • Presently under research is a database of ratings from Racing Post (UK), containing over 1.5m individual performances from 150.000 horses in 150.000 races.
  • The database covers the period from 1998 to date, with all races that took place in UK/Ireland and most of the major races worldwide.
  • The database includes over 5000 individual sires and over 5000 broodmare sires.

Performance ratings for sires show the best sires with progeny averages of 10lbs and more above the population average.

For broodmare sires there is a distinct regression to the mean, with the ‘best’ broodmare sires just 2 or 3lbs above average (and most converging around the average).

In simplified language, this means that there’s no such thing as broodmare sire influence (in the way the rest of the world regards broodmare influence)!

This is in tune with the listing of G1 winners grouped by sire, and the names which crop up as broodmare sires of these G1 winners – they are not what might have been expected.

The thought process following from this is:

  • First-generation Sires matter, but other males in the make-up of a pedigree individually don’t matter when it comes to influencing racing ability.
  • Studying sire x broodmare sire combinations (or extensions of this) is nothing more than an exercise in futility.
  • This fits in with the genetic lottery, where sire and dam each pass on half of their genetic worth, with the other half lost.
  • Full siblings on average will be 50% the same. ON AVERAGE! In reality this similarity will vary from 0% to 100%.
  • On top of that you have to figure out what’s individually dominant and what’s not.
  • Like coat color, where bay is dominant over chestnut, resulting in (on average!) 25% of a population being chestnut and 75% bay – still a chestnut gene can be passed on, going unnoticed, for many generations – then suddenly appearing as if out of nowhere when it meets another hidden chestnut.
  • Now facture in Racing Ability and try to define that . . . 


Sire Lines

So: First-generation Sires matter.

  • But how long can a sire-line last when it comes to excellence?
  • How do we define a sire-line anyway?

Sires currently at stud descend, for over 90% of them, from Phalaris, through two sons, Pharos (full brother to Fairway) and Sickle (full brother to Pharamond & half brother to Hyperion).

We assume that a sire-line-sire is a stallion who has at least one grandson who is a major G1 producing sire.

It is amazing how thin these lines are.

  • Sickle has one son who got a lasting legacy, Unbreakable.
  • Unbreakable has one son, Polynesian.
  • Polynesian has one son, Native Dancer.
  • Native Dancer has one son, Raise A Native.
  • Raise A Native has one son, Mr Prospector.
  • Now we are looking at the sons of Mr P to continue the line. They have to be sires of note to begin with. So: Fappiano, Forty Niner, Gone West, Seeking The Gold, Machiavellian, Kingmambo, Smart Strike …  and none of these have (so far) a son who’s able to continue the line.
  • Pharos has one son who got a lasting legacy, Nearco.
  • Nearco got Nasrullah and 3-part brother Royal Charger, and border-line Nearctic.
  • Nasrullah has one son, Bold Ruler, who got What A Pleasure, Sir Gaylord, Secretariat, Boldnesian. Of these, Boldnesian got Seattle Slew, sire of AP Indy, who has . . . who knows.
  • Royal Charger got Turn To, who has Hail To Reason, sire of Halo and Roberto. Halo sired Sunday Silence who has Deep Impact. But after that?
  • Nearctic got Northern Dancer.

Now we’re talking, you might think.

  • Northern Dancer has Storm Bird, Nijinsky, Danzig, Vice Regent, Sadler’s Wells.
  • Storm Bird got Storm Cat, whose legacy is through Hennessy, Johannesburg, Scat Daddy.. and then?
  • Nijinsky has no presence, Vice Regent got Deputy Minister.
  • Danzig got Danehill, and War Front maybe. Danehill has Redoute’s Choice and maybe there’s still something else not yet on the radar. Same for War Front.
  • That leaves, for Northern Dancer, Sadler’s Wells, who has Galileo (anything is possible still with him) and El Prado (Medaglia d’Oro, Kitten’s Joy are options).

That’s it. That’s all.

Now if sire-lines are so thin in number, and made of such relatively short-lived sequences, and if broodmare sires are not influential, what’s left?


Female Lines

The presence of a SigAnc, a sequence of kin-bred female line ancestors, a closely kinbred dam, and close kin-breeding – these are all factors which have shown to have a positive influence on superior racing ability. It is where the quest for Kentucky Derby winners and Storm Cats starts.

There are, however, major winners who do not have kinbred dams, and are themselves not kin-bred.

Where would they fit in?

  • If first-generation sires matter (the top-line in a pedigree box), and if bottom-female line mares matter, then perhaps there might be a kin-relationship between the two.
  • Kin-breeding through a kin-bred top-line sire-ancestor to a kin-bred bottom-line female ancestor. The Staple.

Returning to the sire-line sequence discussed earlier, the one of Polymelus, Phalaris, Sickle, Unbreakable, Polynesian, Native Dancer, Raise a Native, Mr P.

  • Start with Polymelus, who is kinbred 92% (meaning the 1×1 sire x dam combo exceeds the kin-breeding threshold – our required level of 89/90%), with 2×1 *Arcadia x *Maid Marian 91%. If further extension is needed, his dam *Maid Marian is 3×2 *Haricot x *dau of Young Melbourne 95%, and there’s a top-to-bottom sire-dam kinlink (Staple) 4×3 Doncaster x *dau of Young Melbourne 93%.
  • Phalaris is kinbred 93%, with 2×1 *Maid Marian x *Bromus 93%; the Staple is 1×1 Polymelus x Bromus 92%
  • Sickle is kinbred 95%, with 3×4 *Cheery x *Dongola 90%; Staple is 2×2 Polymelus x *Serenissima 95%
  • Unbreakable is 93%, with 4×2 *Gondolette x *Hour Glass 90%; Staple 2×1 Phalaris x *Blue Grass 91%, 2×2 Phalaris x *Hour Glass 92%, 3×2 Polymelus x *Hour Glass 91%
  • Polynesian is kinbred 89%, with 3x2x4 Selene x Blue Grass x Bird Loose 91-93%; Staple 4×4 Polymelus x *Bird Loose 91% (Polynesian’s dam *Black Polly is by a son of Polymelus, and thus also Staple Polymelus x *Bird Loose 91%)
  • Native Dancer is 85% ncu; Staple is 5×5 Polymelus x *Miss Fiora 92%
  • Raise A Native is 84% with 3×2 *Miyako x *Lady Glory 90%; 3×3 *Miyako x *Beloved 94%; 5×5 *la Grisette x *Padula 90%; Staple 6×5 Polymelus x *Padula 92%
  • Mr Prospector is 82% with 2×3 *Raise You x *Miss Dogwood 91%; Staple 7×5 Polymelus x *Frizeur 89%; 7×6 Polymelus x *Frizette 91%; 7×7 Polymelus x *Ondulee 96%

The impression is that a top-line sire-ancestor can be regarded as a SigAncsire, who (amazingly) in most cases links to exactly the bottom-line *SigAncdam.

That’s the Staple.

For interest some sire-Staples (*SigAncdam in bold) of ‘Chuck Fipke’ stallions:

  • Perfect Soul: 8×7 Polymelus x Miss Albany 90%; 8×8 Polymelus x *Miss Finlay 92% (8×9 Polymelus x *Castelline 93%)
  • Jersey Town: 8×9 Unbreakable x *Mirawala 93%; 9×9 Sickle x *Mirawala 96% (NB *Mirawala = 2×1 Polymelus x *Miranda 91%)
  • Tale Of Ekati : 1×2 Tale Of The Cat x *Maplejinsky 89%;  10×13 Polymelus x *Maid Of Erin 91% (also: *Silence Beauty is 2×4 Halo x Hail to Beauty 91%)
  • Not Bourbon: 8×7 Polymelus x *Kerry Blue 90%
  • Bee Jersey: 3×2 Gone West x *Run For Lassie 92%; 11×9 Polymelus x *Lady Lawless 93%
  • Java’s War: 5x4x5 Red God x Red God x *Miss Uppity 95%; 10×7 Polymelus x *Kindergarten 91%
  • Perfect Timber: 3×3 Northern Dancer x *Coup de Folie 92%; 8×11 Polymelus x *Dazzling 90%
  • Danish Dynaformer: 9×8 Polymelus x Gumdrop 91%; 9×11 Polymelus x *Sainotta 92%

Example of kin-breeding with Staple and Flip-Flop in a major racehorse & sire:

  • Nasrullah – 85%; kinbred 4×3 *Sibola x *Lady Josephine 89%;
  • dam *Mumtaz Begum kinbred 85%; 3×2 *Wild Arum x *Lady Josephine 91%
  • Staple 3×3 Phalaris x *Lady Josephine 89%; 4×5 Polymelus x *Palotta 91%;
  • Flip-flop 3×4 *Catnip x Swynford 91%

Example of kin-breeding with Staple and Flip-Flop in a champion racemare:

  • Forever Unbridled – 74%; kinbred 6×6 *Subterranean x *Closer 90%; 11×10 *Saratoga Belle x *Meetme 92%
  • dam *Lemons Forever 78%; kinbred 6x8x7x5 Miss Dogwood x Dalmary x *Lady Lawless x *Closer 88-93%; 10×9 *Barmaid x *Meetme 92%
  • Staple 8×10 Unbreakable x *Meetme 89%; 9×10 Sickle x *Meetme 90%; 10×10 Phalaris x *Meetme 89%; 11×10 Polymelus x *Meetme 89%
  • Flip-flop 7×9 *U-Boat x Unbreakable 90%


The Right Way?

The right way may well be

a proper G1 producing sire,

on a properly kin-bred dam,

to give a properly kinbred offspring,

on the background of a Staple & Flip-flop

Internal duplications are meaningless, as will close conventional inbreeding be – unless it has meaning.” Karel Miedema



The omni-presence of Polymelus as SigAncsire is baffling. He’s got it all, though.

  • Polymelus is kinbred 92%; 2×1 *Arcadia x *Maid Marian 90%
  • dam *Maid Marian 90%; 3×2 *Haricot x **X Young Melbourne 95% (2nd dam *Quiver 92%; 3rd dam *dau of Young Melbourne 95%; 4th dam *Brown Bess 94%)
  • Staple 1×1 Cyllene x *Maid Marian; 4×3 Doncaster x *dau of Young Melbourne 93%
  • Flip-Flop 2×2 *Arcadia x Hampton 90%; 2×3 *Arcadia x Lord Clifden 90%

A multiple SW in UK between 1904 to 1907, Polymelus was five times Champion Sire.

Jockey Danny Maher, who frequently rode him said “Polymelus has the longest stride of any horse I have known. Because of that he easily becomes unbalanced, which may explain why he either won his races in a canter, or ran disappointingly”.

Karel Miedema
(last updated July 2020)


More from Insight

How Good Is Galileo


How Inbred Is Enable


Sire Lines Rule, OK?


Thoughts to ponder on the making of racing ability


Mitochondrial Inheritance


Ready To Win

A simple approach to mark your card