| "Understanding The Ventura GeoScore" |
“Understanding The Ventura Geographical Score”, by Michael Ventura, copyright
2005.
The basic idea of the Ventura Geographical Score is to determine the
strength of a cross between two ancestors (for instance Danzig and Mr
Prospector), by comparing how many Stakes Winners the cross should have got
compared to how many it got in actuality.
Simply put, a better than average cross will produce more Stakes Winners
than it should and a worse than average cross will get less. And of course
the better than average crosses should be preferred by the breeder.
But how do we determine how many Stakes Winners a cross “Should” have got?
To answer, let us say that we searched the database of a software product
called G1Goldmine to find 412 Stakes Winners descended of Deputy Minister.
Then, not knowing a thing about these 412 Stakes Winners save that they were
Stakes Winners and that they were descended of Deputy Minister, how could we
best estimate what percentage of them have In Reality on the side of the
pedigree opposite to Deputy Minister?
The most logical answer is to look at what percentage of the Stakes Winning
Population as a whole has In Reality on the other side.
Fortunately, the G1Goldmine Program is sophisticated enough to let us do
this and we find that 872 of the 22,477 SWs in its database are descended of
In Reality in the damside for a percentage of 3.88%. Therefore the number
of Deputy Minister descended Stakes Winners that “Should” have In Reality on
the other side is 3.88% of Deputy Minister’s 412 SWs or a little less than
16 SWs (or more precisely 15.99 SWs).
The Regional Score
But wait! Undoubtedly the percentage of Stakes Winners with In Reality
varies from Country to Country in the same way the concentration of Deputy
Minister’s Stakes Winning Descendants undoubtedly also varies. This is a
serious problem for the calculations we made above because it creates a
situation where the relevant Percentage for Deputy Minister Stakes Winners
is NOT the percentage of Stakes Winners in the World with In Reality…
To use an example, In Reality might have 3.88% of the SWs in the World, but
if his SWs were concentrated entirely in Countries where the influence of
Deputy Minister is nowhere to be found, there’d be no chance of even 0.1% of
Deputy Minister’s SWs having In Reality on the other side, much less 3.88%!
But what would be the ideal solution to this problem of variant Geographical
Concentration?
It would be to first take things down to the level of individual Countries
and then add up the results of each Country to get the number of SWs the
cross should’ve got World-Wide. And thankfully, the G1Goldmine Program now
has the amazing capacity to sort and search Stakes Winners by their Country
of Birth, so we can in fact take the above calculations for determining what
percentage of In Reality’s Stakes Winners should have Deputy Minister on the
other side and perform them for each Individual Country and then add them
up.
For instance, Deputy Minister has 299 Stakes Winners in America on the
sireside and in that Country 7.35% of the SWs have In Reality on the
damside. Therefore, the Deputy Minister Sireside / In Reality Damside cross
“Should” have got 7.35% of 299 SWs, or 21.98 Stakes Winners, in the United
States of America.
Then looking at the reverse cross, we find that there are 178 Winners in
America with Deputy Minister on the Damside and that 7.99% of the G1 Winners
in America have In Reality on the Sireside. This means the In Reality
Sireside / Deputy Minister Damside Cross “Should” have got 7.99% of 178 SWs,
or 14.22 Stakes Winners, in the United States of America.
Now to determine how many G1 Winners Should have Deputy Minister on one side
and In Reality on the other we can merely add the 21.98 SWs predicted for
the Deputy Minister/In Reality Cross to the 14.22 SWs Predicted for the In
Reality/Deputy Minister Cross to get 36.2 Stakes Winners.
This means that 36.2 Stakes Winners is the “Control Score” of the Deputy
Minister / In Reality cross in America.
In actuality the cross got 46 Stakes Winners in America. Now to generate
the “Regional Score” we take the number of Stakes Winners the cross actually
got in a particular Country (in this case America) and Simply Divide it by
the number of Stakes Winners it was predicted to have got. In this case the
calculation would go like this: 46 actual SWs Divided By 36.2 predicted SWs
Equals a fair to middling Regional Score of 1.27, which is 27% better than
the Baseline Regional Score of 1.00.
Now, though the Regional Score of a cross is almost always based on less
data than the World-Wide Ventura GeoScore, I still consider it of great
potential use to breeders as the Racing Surfaces, Training Techniques, and
Race Distances Vary A LOT from Country to Country and thus it is well within
the bounds of possibility that a cross will do better in Australia because
it tends to produce horses that like hard tracks or such things; and as the
vast majority of breeders breed Stock with a particular Country in mind,
knowing that a cross does better or worse in a particular Country should be
a VERY useful piece of info.
The World-Wide Ventura Geographical Score
Well, the Control Score of the World-Wide Ventura GeoScore is essentially
generated by adding up all the Control Scores of the Individual Countries.
Then we simply look at how many Stakes Winners the Cross actually did
produce World-Wide and divide that number by the World-Wide Control Score to
generate the World-Wide Ventura GeoScore, which always has the important
advantage of being based on more data than the Regional Scores save for
those rare instances where one ancestor in a cross is exclusively found in
the G1 Winners of a single Country.
Why The Ventura GeoScore?
As I’ve said before, the VGS is based on looking at how well a cross should
have done based on the number of SWs Descended of Ancestor A, the number of
Stakes Winners Descended of Ancestor B, and the level of Geographic overlap
between the two populations and then comparing that to how well the cross
actually did.
It is actually quite common for Scientists to use this technique. For
instance, a recent study reported in the book "Freakonomics" used
Californian Census Records to find if there’s a cause and effect
relationship between having a name much more common among Black People than
White People and having a lower yearly income. To determine this the
Scientists first looked at how much money the People with Especially Black
Sounding Names should have made based on the neighborhood they were born in,
the years of education and income earned by their Mothers, and whether or
not they had a Father in the house. Then they compared how much they should
have made based on these factors to how much they did make and found no
difference. They thus concluded that having a Black Sounding name is not
disadvantageous.
The great strength of applying this Technique to the question of crosses and
nicks is that it Isolates the effect of the cross on the number of Stakes
Winners by factoring in the quantity AND quality of foals descended from the
two ancestors involved in the cross. And it does this simply by taking into
account the total number of G1 Winners the ancestors have descended of
them.
You see, the number of G1 Winners descended from Danzig may not tell you how
many Descendants he has or how likely they were to win Stakes, but it DOES
tell the result of multiplying the total number of foals descended of Danzig
by the percentage of Stakes Winners in his descendant population. Thus
though quantity is not taken into account by itself and neither is quality
taken into account by itself, the COMBINATION of the two IS and therefore
both are taken into account.
And in another important advantage, this technique Isolates the effect of
the Cross and the Blood Compatibility of the Two Ancestors Involved from the
effect of the two ancestors in a so-called nick having their descendant
populations concentrated in the same Country or Countries, which is of
PROFOUND importance in properly evaluating nicks.
I say this because it is very common for a cross to have a high amount of
success primarily because the two aspects of the cross stood in the same
Country. (Note in the following that, if you so choose, the VGS can also be
used for just the winners of Grade/Group 1 Stakes.) For instance, with a
cool 26 G1 Winners Seattle Slew and Mr Prospector are often pointed to as a
very strong nick but the trouble is that 53% of Mr Prospector’s World-Wide
Influence as a progenitor of G1Ws is concentrated in America, which also
happens to be the best hot-bed of Seattle Slew blood there is.
And guess what? By taking this into account the carefully designed VGS
System gives to the Seattle Slew/ Mr Prospector cross, in spite of its
formidable reputation, a World-Wide G1 VGS of 1.57, which is another way of
saying it’s done only 58% better than average in terms of generating G1Ws.
However, with the use of the G1Goldmine Program we can note that a VERY high
proportion of the Seattle / Mr Prospector crossed G1Ws (including American
Eclipse Champions like Mineshaft, Declan’s Moon, and Tempera) get their
Seattle Slew through AP Indy and the VGS on that MORE SPECIFIC cross reveals
a MUCH STRONGER World-Wide Grade/Group 1 Ventura GeoScore of 3.72; which
means the cross of AP Indy and Mr Prospector is over three and a half times
better at producing G1Ws than it should have been even taking into
account the high quality of the two ancestors and the high degree of
Geographic Overlap.
Quite possibly this excellent result is driven partly by AP Indy’s need for
a more speed oriented influence on the damside. And significantly, even
though sons of AP Indy have only sired 6 G1 Winners so far, 2 of them
(Malibu Moon and Pulpit) have managed to sire G1 Winners (Declan’s Moon and
Tapit) from mares descended of Mr Prospector, which strongly indicates this
nick works through sons of AP Indy as well.
(Incidentally, the Seattle Slew/Mr Prospector cross’s G1 VGS of 1.57 would
go down to 1.28 if we left AP Indy’s G1 Winning Descendants out of the
calculations.)
A Note On Sample Size
Crosses with high scores where the high score is based on more than one G1W
are to be preferred greatly to crosses where the high score is based on only
one SW. The reason for this is the Mathematical Principle that an unlikely
event is MUCH more likely to occur once than twice. For instance, the
chances of a specific 1000 to one shot happening to you Tomorrow is One out
of a Thousand (or 0.1%). In sharp contrast the chances of a specific 500 to
one shot happening to you twice tomorrow is calculated by Multiplying the
500 to one times 500 to one to equal a staggeringly unlikely probability of
250,000 to one! (or 0.000004%).
Another reason to be distrustful of high VGSs based on only one SW is that
when you’re talking about them you’re talking about a cross where only ONE
source of Strain B is proven with Strain A. For instance, the Lost
Soldier/Dr Carter cross has produced the Highly Talented G1W Lost In The Fog
from very little opportunity for a freakishly high VGS of 1,160.00, but
besides that being based on only one G1W (too small a sample size to be
valid), the only G1W with the cross gets his Dr Carter through Cloud Break
(the dam of Lost In The Fog).
Therefore, for all we know there is something unique to Cloud Break amongst
all daughters and descendants of Dr Carter that makes her especially
compatible with Lost Soldier (that something unique quite possibly being her
specific genotype and/or her damside ancestors). And in any event, we can’t
rightly speak of a general affinity between Dr Carter and Lost Soldier with
any certainty on account of this.
Still, you’ll most likely get better results breeding Dr Carter mares to
Lost Soldier than mares by any old random damsire; I’m just pointing out
that when a freakishly high score is based on a single SW, it isn’t nearly
as impressive as it seems.
Also, it should of course be noted that the larger the Sample Size, the
stronger the confidence you can put in a high VGS. For instance, Sadlers
Wells over Darshaan have combined in 63 SWs including the English and Irish
Derby Winner High Chapparral and the quadruple G1W Islington, and in spite
of this large sample size the cross still gets a Very Strong VGS of 2.5 (or
150 times better than the average score of 1.00).
Weaknesses Of The Ventura GeoScore
A problem with the VGS is that sometimes meaningful variables can slip in
besides the Compatibility of the Ancestors involved in the cross and affect
for good or ill the number of G1Ws produced by it.
As these variables should be taken into account by the VGS User when
relevant to the cross said user is checking, I here list them:
1. Inbreeding Avoidance. If you check the score for the cross of Mr
Prospector and Mr Prospector (or even a son of Mr Prospector like Gone West
crossed with Mr Prospector) the score will be pushed downward, not by the
weakness of the cross itself, but instead by the variable of breeders
avoiding inbreeding. And since the degree of inbreeding avoidance varies
according the perception of whether or not an ancestor should be inbred to,
the number of generations he is usually back in the pedigree, and the
question of whether or not his descendants are concentrated through a
handful or many different sources of him; there was simply no
way to fairly adjust the VGS Scores of inbreeding crosses upward and thus it
is the work of the user to understand that the VGSs for inbreeding crosses
are a “Law Unto Themselves” and cannot be fairly compared to non-inbreeding
VGS Scores.
2. Association. Sometimes one strain will be thoroughly associated
with another to a point where Inbreeding Avoidance will start depressing the
VGS. For instance, the cross of Seattle Slew and the mare Weekend Surprise
gets a Score of 2.54 and that Score is severely deflated because easily the
majority of Weekend Surprise’s G1 Winning Descendants are descended of her
through none other than AP Indy, the son she has with Seattle Slew
(obviously, many breeders have avoided crossing AP Indy and his sire Seattle
Slew
due to Inbreeding Avoidance and a secondary effect of this action on their
part is they also avoided crossing AP Indy’s dam Weekend Surprise with
Seattle Slew). Once again, it is the job of the VGS User to take this into
account on those fairly rare occasions when it
comes up.
3. The Same Farm Syndrome. When two stallions stood at the same farm
(at least recently enough that the fruits of their proximity might get
picked up by the G1Goldmine Program, which has the G1Ws for about the last
12 Years), that is a variable that strongly tends to increase the gross
amount of success a cross gets and thus it unnaturally drives up the VGS.
However, you’d be surprised how few big Ventura GeoScores seem to be driven
by this as it seems that besides some examples from South America (for
instance Southern Halo/Logical and Slew Gin Fizz/Cipayo) I haven’t been able
to find one example of a really high score based on Multiple G1 Winners
where both stood at the same farm and I strongly suspect that a lot of farms
nowadays don’t bother retaining mares by their stallions for breeding.
Also, it is quite clear that outside of South America they by and large fail
to pick out stallions to stand at their farms based on the idea that they’ll
be compatible with the daughters of stallions they already stand (long gone
indeed is the era of Bull Hancock!), so even if these same farm stallions
get increased opportunities with each other they are quite likely to fail
with each other anyway as there is much truth to the proverb that doing
something stupid a whole bunch of times won’t make you any
more successful. Anyway, if you find a VGS based on more than 1 G1 Winner
which was evidently inflated by both stallions involved in the cross
standing at the same farm, please drop me an e-mail.
4. The Same State Syndrome. In America there will sometimes be
concentrations of a particular ancestor in the mare population of a certain
State. For instance, Valid Appeal stood his entire Stud Career in Florida
and my studies indicate that even today there’s still more Valid Appeal in
Florida than Kentucky, New York, or California. This fact can drive up the
VGSs of crosses involving Valid Appeal and a Stallion that stood in Florida
such as End Sweep, who sired Two G1 Winners (Trippi and Nany’s Sweep) from
Valid Appeal mares to get a huge VGS of 44.62 with him. But though the VGS
of this cross was undoubtedly inflated by End Sweep siring a
lot of his foals in Florida, a State with more Valid Appeal than America as
a whole, the cross is still clearly far better than average even if you take
that into account. And as evidence of this fact we can note that though End
Sweep’s fellow Forty Niner son Distorted Humor most assuredly did not have
the advantage of standing in Florida, he has still managed to combine with
Valid Appeal in the Australian Oaks (G1) Winner Rinky Dink in spite of VERY
limited opportunity with Valid Appeal. Also, another Forty Niner son named
Gold Case(who also did not stand in Florida), managed to sire the 3-Time G2W
Randaroo from a Valid Appeal mare.
How To Use The Ventura GeoScore
Let’s say you’re a mare owner and you wanted to use the VGS Value in making
breeding decisions. How would you best go about it?
Well, the first thing I recommend doing is looking up the sire of your mare
in the G1Goldmine Program on both the sireside and the damside so that you
may search all the Stakes Winners descended of him. Now see if you can
find any ancestor who seems to be showing up a lot on the side of the
pedigree opposite to your mare’s sire (and especially look for an ancestor
who seems to be showing up with your mare’s sire more often than you’d have
expected).
Also if so choose you can use the Affinity matrix feature of Goldmine to
look up the Percentages of the Stakes Winners with your mare's sire in the
damside with certain ancestors in the sireside; keeping your eye out for
sireside ancestors with a higher percentage than you'd think they'd be based
on opportunity.
If your search is successful I highly recommend searching the Ventura
GeoScore on the cross to see if it gets a good Score and if it does you can
mark that down as an ancestor to look for in potential Stallions for your
mare.
For instance, if you had a mare by Medaglia D'Oro and used the G1Goldmine
Program to look through his Stakes Winning Descendants you’d almost
certainly find that a remarkable proportion of them (3 out of 10) have Cure
the Blues on the side opposite to Medaglia D'Oro.
And then when you checked the VGS on the cross of Medaglia D'Oro (which is
interestingly enough found in the Champion Rachel Alexandra), a Remarkable
Ventura GeoScore of 59.7, or more than 59 times better than the average
cross.
And finally the lesson the owner of the Medaglia D'Oro mare should glean
from this would be to give special consideration to Good Stallions descended
of Cure the Blues for his mare.
Of course, ancestors of your mare besides her sire are also very relevant to
breeding her and if you have to go back further in her pedigree to find a
high VGS it’s not all that big a deal.
Also, if you could find more than one high VGS that you can effect with one
mating it certainly stands to reason that would be even better than getting
one high one.
For instance, the tremendous racehorse and Breeder’s Cup Classic Record
Holder Ghostzapper undoubtedly takes the cake in regards to packing a
plethora of High VGSs. And I say that as he possesses no less than 18 G1
VGS Scores over 2.00 based on Multiple G1Ws, no less than 11 over 4.00 based
on multiple G1Ws, and no less than 5 over 6.00 based on multiple G1Ws (if
you ever find another horse with anywhere near this many high VGS, please
send me an e-mail me on the subject as I sure haven’t found any).
Coda
I realize that the Ventura GeoScore is a pretty complicated beast; and if
you have any as yet unanswered questions about it you may contact me with
them at: VenturaPedigreeAnalysis@comcast.net, through my home e-mail of
rockysocreates8@yahoo.com, or through the "Contact Michael Ventura" page of
this website.
Note: The G1Goldmine Program (www.g1goldmine.com) Possesses The Capacity to
Generate The Ventura Geographical Score Discussed Above.
Which is just as well given that calculating it by hand is very difficult
(believe me I've tried it).
Download The VGS System Here For Free...
2005.
The basic idea of the Ventura Geographical Score is to determine the
strength of a cross between two ancestors (for instance Danzig and Mr
Prospector), by comparing how many Stakes Winners the cross should have got
compared to how many it got in actuality.
Simply put, a better than average cross will produce more Stakes Winners
than it should and a worse than average cross will get less. And of course
the better than average crosses should be preferred by the breeder.
But how do we determine how many Stakes Winners a cross “Should” have got?
To answer, let us say that we searched the database of a software product
called G1Goldmine to find 412 Stakes Winners descended of Deputy Minister.
Then, not knowing a thing about these 412 Stakes Winners save that they were
Stakes Winners and that they were descended of Deputy Minister, how could we
best estimate what percentage of them have In Reality on the side of the
pedigree opposite to Deputy Minister?
The most logical answer is to look at what percentage of the Stakes Winning
Population as a whole has In Reality on the other side.
Fortunately, the G1Goldmine Program is sophisticated enough to let us do
this and we find that 872 of the 22,477 SWs in its database are descended of
In Reality in the damside for a percentage of 3.88%. Therefore the number
of Deputy Minister descended Stakes Winners that “Should” have In Reality on
the other side is 3.88% of Deputy Minister’s 412 SWs or a little less than
16 SWs (or more precisely 15.99 SWs).
The Regional Score
But wait! Undoubtedly the percentage of Stakes Winners with In Reality
varies from Country to Country in the same way the concentration of Deputy
Minister’s Stakes Winning Descendants undoubtedly also varies. This is a
serious problem for the calculations we made above because it creates a
situation where the relevant Percentage for Deputy Minister Stakes Winners
is NOT the percentage of Stakes Winners in the World with In Reality…
To use an example, In Reality might have 3.88% of the SWs in the World, but
if his SWs were concentrated entirely in Countries where the influence of
Deputy Minister is nowhere to be found, there’d be no chance of even 0.1% of
Deputy Minister’s SWs having In Reality on the other side, much less 3.88%!
But what would be the ideal solution to this problem of variant Geographical
Concentration?
It would be to first take things down to the level of individual Countries
and then add up the results of each Country to get the number of SWs the
cross should’ve got World-Wide. And thankfully, the G1Goldmine Program now
has the amazing capacity to sort and search Stakes Winners by their Country
of Birth, so we can in fact take the above calculations for determining what
percentage of In Reality’s Stakes Winners should have Deputy Minister on the
other side and perform them for each Individual Country and then add them
up.
For instance, Deputy Minister has 299 Stakes Winners in America on the
sireside and in that Country 7.35% of the SWs have In Reality on the
damside. Therefore, the Deputy Minister Sireside / In Reality Damside cross
“Should” have got 7.35% of 299 SWs, or 21.98 Stakes Winners, in the United
States of America.
Then looking at the reverse cross, we find that there are 178 Winners in
America with Deputy Minister on the Damside and that 7.99% of the G1 Winners
in America have In Reality on the Sireside. This means the In Reality
Sireside / Deputy Minister Damside Cross “Should” have got 7.99% of 178 SWs,
or 14.22 Stakes Winners, in the United States of America.
Now to determine how many G1 Winners Should have Deputy Minister on one side
and In Reality on the other we can merely add the 21.98 SWs predicted for
the Deputy Minister/In Reality Cross to the 14.22 SWs Predicted for the In
Reality/Deputy Minister Cross to get 36.2 Stakes Winners.
This means that 36.2 Stakes Winners is the “Control Score” of the Deputy
Minister / In Reality cross in America.
In actuality the cross got 46 Stakes Winners in America. Now to generate
the “Regional Score” we take the number of Stakes Winners the cross actually
got in a particular Country (in this case America) and Simply Divide it by
the number of Stakes Winners it was predicted to have got. In this case the
calculation would go like this: 46 actual SWs Divided By 36.2 predicted SWs
Equals a fair to middling Regional Score of 1.27, which is 27% better than
the Baseline Regional Score of 1.00.
Now, though the Regional Score of a cross is almost always based on less
data than the World-Wide Ventura GeoScore, I still consider it of great
potential use to breeders as the Racing Surfaces, Training Techniques, and
Race Distances Vary A LOT from Country to Country and thus it is well within
the bounds of possibility that a cross will do better in Australia because
it tends to produce horses that like hard tracks or such things; and as the
vast majority of breeders breed Stock with a particular Country in mind,
knowing that a cross does better or worse in a particular Country should be
a VERY useful piece of info.
The World-Wide Ventura Geographical Score
Well, the Control Score of the World-Wide Ventura GeoScore is essentially
generated by adding up all the Control Scores of the Individual Countries.
Then we simply look at how many Stakes Winners the Cross actually did
produce World-Wide and divide that number by the World-Wide Control Score to
generate the World-Wide Ventura GeoScore, which always has the important
advantage of being based on more data than the Regional Scores save for
those rare instances where one ancestor in a cross is exclusively found in
the G1 Winners of a single Country.
Why The Ventura GeoScore?
As I’ve said before, the VGS is based on looking at how well a cross should
have done based on the number of SWs Descended of Ancestor A, the number of
Stakes Winners Descended of Ancestor B, and the level of Geographic overlap
between the two populations and then comparing that to how well the cross
actually did.
It is actually quite common for Scientists to use this technique. For
instance, a recent study reported in the book "Freakonomics" used
Californian Census Records to find if there’s a cause and effect
relationship between having a name much more common among Black People than
White People and having a lower yearly income. To determine this the
Scientists first looked at how much money the People with Especially Black
Sounding Names should have made based on the neighborhood they were born in,
the years of education and income earned by their Mothers, and whether or
not they had a Father in the house. Then they compared how much they should
have made based on these factors to how much they did make and found no
difference. They thus concluded that having a Black Sounding name is not
disadvantageous.
The great strength of applying this Technique to the question of crosses and
nicks is that it Isolates the effect of the cross on the number of Stakes
Winners by factoring in the quantity AND quality of foals descended from the
two ancestors involved in the cross. And it does this simply by taking into
account the total number of G1 Winners the ancestors have descended of
them.
You see, the number of G1 Winners descended from Danzig may not tell you how
many Descendants he has or how likely they were to win Stakes, but it DOES
tell the result of multiplying the total number of foals descended of Danzig
by the percentage of Stakes Winners in his descendant population. Thus
though quantity is not taken into account by itself and neither is quality
taken into account by itself, the COMBINATION of the two IS and therefore
both are taken into account.
And in another important advantage, this technique Isolates the effect of
the Cross and the Blood Compatibility of the Two Ancestors Involved from the
effect of the two ancestors in a so-called nick having their descendant
populations concentrated in the same Country or Countries, which is of
PROFOUND importance in properly evaluating nicks.
I say this because it is very common for a cross to have a high amount of
success primarily because the two aspects of the cross stood in the same
Country. (Note in the following that, if you so choose, the VGS can also be
used for just the winners of Grade/Group 1 Stakes.) For instance, with a
cool 26 G1 Winners Seattle Slew and Mr Prospector are often pointed to as a
very strong nick but the trouble is that 53% of Mr Prospector’s World-Wide
Influence as a progenitor of G1Ws is concentrated in America, which also
happens to be the best hot-bed of Seattle Slew blood there is.
And guess what? By taking this into account the carefully designed VGS
System gives to the Seattle Slew/ Mr Prospector cross, in spite of its
formidable reputation, a World-Wide G1 VGS of 1.57, which is another way of
saying it’s done only 58% better than average in terms of generating G1Ws.
However, with the use of the G1Goldmine Program we can note that a VERY high
proportion of the Seattle / Mr Prospector crossed G1Ws (including American
Eclipse Champions like Mineshaft, Declan’s Moon, and Tempera) get their
Seattle Slew through AP Indy and the VGS on that MORE SPECIFIC cross reveals
a MUCH STRONGER World-Wide Grade/Group 1 Ventura GeoScore of 3.72; which
means the cross of AP Indy and Mr Prospector is over three and a half times
better at producing G1Ws than it should have been even taking into
account the high quality of the two ancestors and the high degree of
Geographic Overlap.
Quite possibly this excellent result is driven partly by AP Indy’s need for
a more speed oriented influence on the damside. And significantly, even
though sons of AP Indy have only sired 6 G1 Winners so far, 2 of them
(Malibu Moon and Pulpit) have managed to sire G1 Winners (Declan’s Moon and
Tapit) from mares descended of Mr Prospector, which strongly indicates this
nick works through sons of AP Indy as well.
(Incidentally, the Seattle Slew/Mr Prospector cross’s G1 VGS of 1.57 would
go down to 1.28 if we left AP Indy’s G1 Winning Descendants out of the
calculations.)
A Note On Sample Size
Crosses with high scores where the high score is based on more than one G1W
are to be preferred greatly to crosses where the high score is based on only
one SW. The reason for this is the Mathematical Principle that an unlikely
event is MUCH more likely to occur once than twice. For instance, the
chances of a specific 1000 to one shot happening to you Tomorrow is One out
of a Thousand (or 0.1%). In sharp contrast the chances of a specific 500 to
one shot happening to you twice tomorrow is calculated by Multiplying the
500 to one times 500 to one to equal a staggeringly unlikely probability of
250,000 to one! (or 0.000004%).
Another reason to be distrustful of high VGSs based on only one SW is that
when you’re talking about them you’re talking about a cross where only ONE
source of Strain B is proven with Strain A. For instance, the Lost
Soldier/Dr Carter cross has produced the Highly Talented G1W Lost In The Fog
from very little opportunity for a freakishly high VGS of 1,160.00, but
besides that being based on only one G1W (too small a sample size to be
valid), the only G1W with the cross gets his Dr Carter through Cloud Break
(the dam of Lost In The Fog).
Therefore, for all we know there is something unique to Cloud Break amongst
all daughters and descendants of Dr Carter that makes her especially
compatible with Lost Soldier (that something unique quite possibly being her
specific genotype and/or her damside ancestors). And in any event, we can’t
rightly speak of a general affinity between Dr Carter and Lost Soldier with
any certainty on account of this.
Still, you’ll most likely get better results breeding Dr Carter mares to
Lost Soldier than mares by any old random damsire; I’m just pointing out
that when a freakishly high score is based on a single SW, it isn’t nearly
as impressive as it seems.
Also, it should of course be noted that the larger the Sample Size, the
stronger the confidence you can put in a high VGS. For instance, Sadlers
Wells over Darshaan have combined in 63 SWs including the English and Irish
Derby Winner High Chapparral and the quadruple G1W Islington, and in spite
of this large sample size the cross still gets a Very Strong VGS of 2.5 (or
150 times better than the average score of 1.00).
Weaknesses Of The Ventura GeoScore
A problem with the VGS is that sometimes meaningful variables can slip in
besides the Compatibility of the Ancestors involved in the cross and affect
for good or ill the number of G1Ws produced by it.
As these variables should be taken into account by the VGS User when
relevant to the cross said user is checking, I here list them:
1. Inbreeding Avoidance. If you check the score for the cross of Mr
Prospector and Mr Prospector (or even a son of Mr Prospector like Gone West
crossed with Mr Prospector) the score will be pushed downward, not by the
weakness of the cross itself, but instead by the variable of breeders
avoiding inbreeding. And since the degree of inbreeding avoidance varies
according the perception of whether or not an ancestor should be inbred to,
the number of generations he is usually back in the pedigree, and the
question of whether or not his descendants are concentrated through a
handful or many different sources of him; there was simply no
way to fairly adjust the VGS Scores of inbreeding crosses upward and thus it
is the work of the user to understand that the VGSs for inbreeding crosses
are a “Law Unto Themselves” and cannot be fairly compared to non-inbreeding
VGS Scores.
2. Association. Sometimes one strain will be thoroughly associated
with another to a point where Inbreeding Avoidance will start depressing the
VGS. For instance, the cross of Seattle Slew and the mare Weekend Surprise
gets a Score of 2.54 and that Score is severely deflated because easily the
majority of Weekend Surprise’s G1 Winning Descendants are descended of her
through none other than AP Indy, the son she has with Seattle Slew
(obviously, many breeders have avoided crossing AP Indy and his sire Seattle
Slew
due to Inbreeding Avoidance and a secondary effect of this action on their
part is they also avoided crossing AP Indy’s dam Weekend Surprise with
Seattle Slew). Once again, it is the job of the VGS User to take this into
account on those fairly rare occasions when it
comes up.
3. The Same Farm Syndrome. When two stallions stood at the same farm
(at least recently enough that the fruits of their proximity might get
picked up by the G1Goldmine Program, which has the G1Ws for about the last
12 Years), that is a variable that strongly tends to increase the gross
amount of success a cross gets and thus it unnaturally drives up the VGS.
However, you’d be surprised how few big Ventura GeoScores seem to be driven
by this as it seems that besides some examples from South America (for
instance Southern Halo/Logical and Slew Gin Fizz/Cipayo) I haven’t been able
to find one example of a really high score based on Multiple G1 Winners
where both stood at the same farm and I strongly suspect that a lot of farms
nowadays don’t bother retaining mares by their stallions for breeding.
Also, it is quite clear that outside of South America they by and large fail
to pick out stallions to stand at their farms based on the idea that they’ll
be compatible with the daughters of stallions they already stand (long gone
indeed is the era of Bull Hancock!), so even if these same farm stallions
get increased opportunities with each other they are quite likely to fail
with each other anyway as there is much truth to the proverb that doing
something stupid a whole bunch of times won’t make you any
more successful. Anyway, if you find a VGS based on more than 1 G1 Winner
which was evidently inflated by both stallions involved in the cross
standing at the same farm, please drop me an e-mail.
4. The Same State Syndrome. In America there will sometimes be
concentrations of a particular ancestor in the mare population of a certain
State. For instance, Valid Appeal stood his entire Stud Career in Florida
and my studies indicate that even today there’s still more Valid Appeal in
Florida than Kentucky, New York, or California. This fact can drive up the
VGSs of crosses involving Valid Appeal and a Stallion that stood in Florida
such as End Sweep, who sired Two G1 Winners (Trippi and Nany’s Sweep) from
Valid Appeal mares to get a huge VGS of 44.62 with him. But though the VGS
of this cross was undoubtedly inflated by End Sweep siring a
lot of his foals in Florida, a State with more Valid Appeal than America as
a whole, the cross is still clearly far better than average even if you take
that into account. And as evidence of this fact we can note that though End
Sweep’s fellow Forty Niner son Distorted Humor most assuredly did not have
the advantage of standing in Florida, he has still managed to combine with
Valid Appeal in the Australian Oaks (G1) Winner Rinky Dink in spite of VERY
limited opportunity with Valid Appeal. Also, another Forty Niner son named
Gold Case(who also did not stand in Florida), managed to sire the 3-Time G2W
Randaroo from a Valid Appeal mare.
How To Use The Ventura GeoScore
Let’s say you’re a mare owner and you wanted to use the VGS Value in making
breeding decisions. How would you best go about it?
Well, the first thing I recommend doing is looking up the sire of your mare
in the G1Goldmine Program on both the sireside and the damside so that you
may search all the Stakes Winners descended of him. Now see if you can
find any ancestor who seems to be showing up a lot on the side of the
pedigree opposite to your mare’s sire (and especially look for an ancestor
who seems to be showing up with your mare’s sire more often than you’d have
expected).
Also if so choose you can use the Affinity matrix feature of Goldmine to
look up the Percentages of the Stakes Winners with your mare's sire in the
damside with certain ancestors in the sireside; keeping your eye out for
sireside ancestors with a higher percentage than you'd think they'd be based
on opportunity.
If your search is successful I highly recommend searching the Ventura
GeoScore on the cross to see if it gets a good Score and if it does you can
mark that down as an ancestor to look for in potential Stallions for your
mare.
For instance, if you had a mare by Medaglia D'Oro and used the G1Goldmine
Program to look through his Stakes Winning Descendants you’d almost
certainly find that a remarkable proportion of them (3 out of 10) have Cure
the Blues on the side opposite to Medaglia D'Oro.
And then when you checked the VGS on the cross of Medaglia D'Oro (which is
interestingly enough found in the Champion Rachel Alexandra), a Remarkable
Ventura GeoScore of 59.7, or more than 59 times better than the average
cross.
And finally the lesson the owner of the Medaglia D'Oro mare should glean
from this would be to give special consideration to Good Stallions descended
of Cure the Blues for his mare.
Of course, ancestors of your mare besides her sire are also very relevant to
breeding her and if you have to go back further in her pedigree to find a
high VGS it’s not all that big a deal.
Also, if you could find more than one high VGS that you can effect with one
mating it certainly stands to reason that would be even better than getting
one high one.
For instance, the tremendous racehorse and Breeder’s Cup Classic Record
Holder Ghostzapper undoubtedly takes the cake in regards to packing a
plethora of High VGSs. And I say that as he possesses no less than 18 G1
VGS Scores over 2.00 based on Multiple G1Ws, no less than 11 over 4.00 based
on multiple G1Ws, and no less than 5 over 6.00 based on multiple G1Ws (if
you ever find another horse with anywhere near this many high VGS, please
send me an e-mail me on the subject as I sure haven’t found any).
Coda
I realize that the Ventura GeoScore is a pretty complicated beast; and if
you have any as yet unanswered questions about it you may contact me with
them at: VenturaPedigreeAnalysis@comcast.net, through my home e-mail of
rockysocreates8@yahoo.com, or through the "Contact Michael Ventura" page of
this website.
Note: The G1Goldmine Program (www.g1goldmine.com) Possesses The Capacity to
Generate The Ventura Geographical Score Discussed Above.
Which is just as well given that calculating it by hand is very difficult
(believe me I've tried it).
Download The VGS System Here For Free...
 |
| |
 |
| |
 |
| |
 |
| |
 |