What is Proxygene?

Proxygene was conceived in the 1980's in response to various changes within the racing and breeding world. Stallion book sizes were growing, increasing globalisation was taking place and high value sales and 'closed' races had arrived. All three factors led to the blurring of statistical information used to judge a stallion's performance on the racecourse.

Historically aggregate prize money had been a relatively reliable and easy to compile expression of a stallions ability. Average earnings could be used to correct for quantity rather than quality but was easily distorted by even one large prize. Imports would also be ranked differently on tables based on quantity or quality, which in turn would be complicated by foreign earnings.

In the bloodstock world, level of ability is also used. Prizemoney is discarded for stakes or pattern race performance often expressed as a percentage of stakes winners to runners etc. This is useful information but inevitably has a 'cut off' point, whereby for example a non-winning classic placed animal may be rated inferior to a lesser horse who wins a minor stakes race.

Many other classifications are based upon the top horses only, the vast majority of horses being ignored. This is because top horses are easy to track but the whole of a crop is not. Are prize money and stakes races the most accurate assessment of ability? No, but they are easy to compile.

Progeny testing takes place in many other species, for 'advancement' of the breed. Bulls can be ranked by the performance of their offspring. A bull, which produces daughters who produce more milk, will be ranked above lesser bulls. Such progeny testing is directed towards the ultimate target, in such an example more milk. To be of statistical importance such testing should be controlled and directed, and based upon measurable attributes. It should also include a fixed population so that both good and bad examples of offspring are included.

In racing there is an ideal testing ground for progeny, the racecourse. Earnings, pattern race prestige and race times are all tools in analysis but are somewhat arbitrary. The most accurate individual appraisal is by rating for example those of Timeform. Such values are used in stallion promotion and are well understood. An exciting development was a table based upon the analysis of median Timeform ratings for stallion's offspring. This gave a good evaluation of whether a stallion was capable of elevating the standard of his produce within a fixed crop rather than of selected individuals alone. The increasing availability of computing power in the eighties meant that more of such information could be stored and processed to produce a more refined set of information.

British racing currently 'enjoys' the performances of over 10,000 racehorses each year. Compiling ratings for each of them is time consuming but ideally includes every one. It also needs to be as consistent from year to year as possible, to enable like for like comparison. It is also sensible to include all weather racing so no evidence is excluded. By restriction to British Racing a relatively complete population can be analysed. Performances abroad are ignored but British performances by foreign raiders are included.

Using this raw data Proxygene analyses stallion performances on a January to December basis. Each runner is analysed based upon age, 2 year old or older horse, and sex, filly vs colt or gelding. Sire and broodmare sire are recorded with an end of season rating. Where sufficient data exists the average rating for each sire and broodmare sire can be computed. This in itself produces valuable data but can be further refined. In the same way Average Earnings Indices can be adjusted to reflect the quality of the mates, so each runner is reanalysed with an allowance for any improvement a sire makes above the average of the corresponding broodmare sire (and vice versa). Further allowance is made for the quality of runner; to upgrade a rating of 110 to 120 is more worthy than upgrading from 40 to 50; and quantity of evidence, a high average rating from 100 runners will be more reliable than an average from just 10.

Over the last 30 years the leading Proxygene sires and Broodmare sires have been:-

SIRE BROODMARE SIRE
2017 Dubawi Galileo
2016 War Front Galileo
2015 Dubawi Danehill
2014 Galileo Danehill
2013 Galileo Mind Games
2012 Galileo Danehill
2011 Galileo Cozzene
2010 Oasis Dream Distant View
2009 Oasis Dream Distant View
2008 Pivotal Indian Ridge
2007 Galileo Dashing Blade
2006 Danehill Palace Music
2005 Danehill Efisio
2004 Pivotal Soviet Star
2003 Danzig Miswaki
2002 Danehill Alysheba
2001 Danzig Private Account
2000 Danzig Mr Prospector
1999 Nureyev Mr Prospector
1998 Sadlers Wells Round Table
1997 Sadlers Wells Frimley Park
1996 Sadlers Wells Darshaan
1995 Sadlers Wells Roberto
1994 Sadlers Wells Exclusive Native
1993 Sadlers Wells The Axe
1992 Sadlers Wells Alleged
1991 Green Desert Green Dancer
1990 Sadlers Wells Nijinsky
1989 Diesis Nijinsky
1988 Shirley Heights Blushing Groom
1987 Nureyev Northern Dancer
1986 Mill Reef Drone
1985 Young Generation Sea Bird

The intention of the project was to identify sire and broodmare sires particularly with a view to ability hidden in conventional statistics. Early evidence might allow early identification and opportunity for use before export or retirement arrived. With 30 years of data analysed (135,000+ annual performances) the leading sire over that time period is unsurprisingly Danzig. The table for broodmare sires is however much more unpredictable. For example until 2009 Danzig had been a negative influence as a broodmare sire, a table topped at the time by Mill Reef, a mantle taken on by Ahonoora now Mill Reef no longer appears close up in current pedigrees.

Below is a sample of the top 50 sires and broodmare sires of 2007 and the ratings achieved as an example.

9275 75 Galileo 11826 9 Dashing Blade
9068 127 Pivotal 11277 3 Dodge
8115 75 Danehill 10665 122 Danehill
7830 2 Ghazi 9257 23 Rousillon
7478 2 Dubai Millennium 8366 2 Dance in Time
7321 129 Danehill Dancer 8106 15 Kaldoun
7168 5 Van Nistelrooy 7847 17 Rudimentary
7079 15 Mr Greeley 7493 5 Clever Trick
6720 96 Dansili 7349 3 Flying Spur
6596 1 Nureyev 7149 3 Milford
6440 17 Danzig 7077 45 Generous
6428 8 Monsun 7038 15 Bahri
6222 65 Rock of Gibraltar 6832 53 Thatching
6159 4 Carson City 6821 17 Clantime
6095 3 Royal Abjar 6746 91 Warning
6005 95 Sadlers Wells 6605 28 Reprimand
5793 101 Green Desert 6598 75 Last Tycoon
5567 16 Lemon Drop Kid 6446 23 Arazi
5398 42 Acclamation 6396 73 Mr Prospector
5375 1 Prized 6357 5 Cosmonaut
5349 4 Belong to Me 6330 155 Darshaan
5298 61 Indian Ridge 6088 7 Glow
5128 48 Oasis Dream 6074 10 Damister
5112 80 Danetime 6052 4 Octagonal
5079 67 In the Wings 5989 198 Sadlers Wells
4919 118 Cape Cross 5829 38 Soviet Star
4622 16 Dynaformer 5716 7 Definite Article
4591 15 Storm Cat 5712 13 Indian King
4544 12 Chester House 5683 85 Selkirk
4350 95 Montjeu 5577 9 Ajdal
4346 64 Invincible Spirit 5572 4 Celestial Storm
4196 6 Arch 5502 1 Kala Shikari
4168 3 Kings Signet 5467 7 Kendor
4094 58 Kingmambo 5382 7 Alnasr Alwasheek
3853 136 Royal Applause 5357 5 Sunshine Forever
3827 78 Singspiel 5355 29 Ahonoora
3792 5 Smart Strike 5266 4 Future Storm
3789 2 AP Indy 5173 9 Manila
3697 97 Cadeaux Genereux 5113 30 Storm Cat
3682 34 Rahy 5110 3 Grey Dawn
3665 67 Machiavellian 5074 16 Song
3611 1 Capote 5071 9 Timeless Times
3605 31 Elusive Quality 5010 102 Indian Ridge
3576 3 Gilded Time 5000 1 Atraf
3491 1 Weldnaas 4904 2 Sunday Silence
3403 7 Dixieland Band 4723 5 Electric
3392 10 Dalakhani 4669 43 Danzig
3382 16 Magic Ring 4606 8 Anshan
3380 10 Silver Hawk 4552 16 Tinas Pet
3204 1 Shadeed 4546 21 Silver Hawk

 

Key

First column sire rating
Second column number of runners
Third column sire
Fourth column broodmare sire rating
Fifth column examples analysed
Sixth column broodmare sire

The next stage of Proxygene's evolution turned towards how this data could be applied, particularly in the breeding of a superior racehorse. The concept of nicks and 'blood affinities' has been around for some time, and has been somewhat difficult to prove mainly due to patchy evidence. In a small sire sample size of for example 5 runners, are 2 stakes winners truly of significance? However, if all British runners of a sire over a 25 year period are analysed by rating comparison this should provide better evidence of compatability.

For each runner, if we know the average rating of products by the same sire and combine this information with the average of products of its broodmare sire, we should have an idea of expected level of ability. If that runner races to a rating beyond expectation this would be a positive result on an individual basis. The greater the rating is beyond expectation, the greater the positive value. Anything below expectation generates a negative result. In the early stages of development this could be used to prove or disprove certain theories, for example do Darshaan mares nick with Sadler's Wells? (They do!)

Like many 'seek gold where gold has already been found' systems the volume of data may still be small. As a consequence some systems will draw upon evidence from more remote generations usually within the sire line. On average a grandparent contributes only 25% of the genetic makeup of an individual, although in theory this can range from 0 to 50%. Proxygene's next step therefore, was to analyse each nick not just by sire, but by all male ancestors of the sire up to great great grandsire, to see if further evidence of gold could be sieved out. As a further refinement by combining results of a dams 3 closest broodmare sires, potentially negative nicks could be filtered out, with the hope of increasing the chances of a successful product.

A typical Proxygene result would look like this. (just the top 50 printed of 900 sires analysed)

TRULY A GIFT is an Arazi mare out of a Darshaan mare out of a Caerleon mare.

Arazi
Darshaan
Caerleon
Rating
Sire
5116 91 5981 215 4793 283 139798 Danehill
5890 17 4495 2 4741 0 138730 Nureyev
4547 29 4133 12 5554 4 137315 Danzig
6810 1 14275 80 2896 0 134805 Gold Away
4470 167 5448 1490 5151 130 129806 Sadlers Wells
4373 22 4040 69 5802 104 103121 Warning
4464 165 2870 550 5713 92 97753 Green Desert
4771 64 4236 89 5170 74 95046 Danehill Dancer
4432 15 5655 2 4128 0 92211 Nijinsky
3844 12 8023 12 3676 32 90179 Mozart
5148 61 3258 0 3371 14 89379 Storm Cat
4007 9 6380 170 5153 49 88368 Rock of Gibraltar
3265 0 6715 22 6053 32 85944 Lomond
5016 4 4874 0 3934 4 85545 Stravinsky
4651 4 6034 21 4834 8 85441 Flying Spur
3384 21 3600 325 7135 53 85036 Galileo
6737 3 3326 0 3074 0 83416 Van Nistelrooy
3938 9 5728 12 4164 0 83253 Exceed And Excel
5344 37 4576 116 4173 76 81619 Fasliyev
4272 23 3515 2 5380 30 81592 Dayjur
5735 1 5008 21 3444 0 80328 Oratorio
2164 71 6032 264 5444 0 79966 Caerleon
6229 30 3657 72 4021 31 79905 Zilzal
2890 128 4558 233 3767 294 79078 Machiavellian
4009 31 5283 196 2871 102 78800 Fairy King
3777 0 6057 1 3787 0 78457 Lammtara
4555 5 9448 35 1688 84 78449 Trempolino
3127 77 3727 0 3034 49 78359 Alleged
3999 0 3858 0 4380 0 76900 Topsider
4050 3 5060 10 3735 0 76565 Zieten
3629 0 3368 0 4436 0 76075 Polish Numbers
3943 1 5667 64 4497 20 75692 Act One
2657 47 2663 1 5267 63 75500 Mr Prospector
4130 30 5846 8 3817 0 75095 Niniski
3946 6 4139 84 4647 60 75059 Soviet Star
3990 54 5088 202 4731 51 73874 Hernando
4724 7 2992 21 4261 114 73733 Giant's Causeway
3261 0 7833 12 3563 0 73684 Seattle Dancer
5813 19 5136 18 3870 5 73606 Eagle Eyed
3352 0 5399 0 5844 0 73532 Tate Gallery
2549 27 6237 33 2589 0 73218 El Gran Senor
6407 51 2333 10 4244 75 73172 Spinning World
4766 3 3325 2 2517 0 72736 Storm Bird
4501 26 2557 4 3142 29 72440 Roberto
3232 0 4224 0 5791 0 72141 Ajdal
3644 13 2960 0 5428 101 71775 Miswaki
2913 2 5059 7 4230 4 70832 Known Fact
4989 51 4231 391 4795 53 70477 Polish Precedent
3808 28 4094 112 4457 38 70469 Peintre Celebre
3109 66 4965 38 2310 253 70124 Diesis

Key

First column is proxygene nick for Arazi alone with guide to volume of evidence in column 2. Each further column relates to nicks for Darshaan and Caerleon with a combined rating in front of each potential or deceased sire. The value of the rating is not so important other than to rank each sire. Deceased stallions are included so more speculative use can be made of them in assessing their sons as sires without runners yet.

Based on this result Truly A Gift was booked to Soviet Star.

If you have any questions or suggestions please contact us by email at proxygene.(enquiries@antispamproxygene.co.uk- just remove the antispam), or preferably via facebook for a quicker reply.

return to home page