Dosage: Pedigree & Performance
Main Menu         Search

The Relationship Between the DI and the CD: DP Patterns

The core of contemporary Dosage methodology is the Dosage Profile, or DP.  It is the series of five numbers that summarize the aptitudinal contributions from chefs-de-race in the first four generations of a pedigree.  Quite often, much useful information can be derived from the DP alone.  For example, regardless of the DI or the CD, a horse with a DP having double-digit representation in the Brilliant group is often a candidate to win shorter races.  To illustrate the main point, among horses with a DI of exactly 3.00, those with at least 10 points in the Brilliant category of their DP have an average winning distance (AWD) of 7.88 furlongs.  Those with less than 10 points in the Brilliant category have an AWD of 8.01 furlongs.  Similarly, among horses with a DI of exactly 2.00, those with at least 10 points in the Brilliant category of their DP have an AWD of 8.25 furlongs.  Those with less than 10 points in the Brilliant category have an AWD of 8.43 furlongs.  Not surprisingly, a large component of inherited speed is an asset in shorter races.  Although the average DI is the same for each respective group, we do see real d    ifferences in the average CD.  The first group, those with DI 3.00 and at least ten Brilliant points in their DP, has an average CD of 0.89.  The second group, with less than ten points in their DP, has an average CD of 0.81.  Similarly, those with DI 2.00 and at least ten Brilliant points in their DP, have an average CD of 0.65 while those with less than ten points in their DP have an average CD of 0.56.  The principle here is that the relationship between DI and CD is the direct consequence of how the aptitudinal points are distributed within the DP.  Furthermore, it confirms that the CD is a more accurate measure of relative distance potential than is the DI.

There are numerous examples demonstrating the critical nature of point distribution within the DP.  The Table below presents several variations on how DP configuration affects average winning distance (AWD), the percentage of stakes wins in sprints (SPR), the percentage of stakes wins on the grass (TURF) and the percentage stakes of wins for juveniles (2YO) for the population falling within each DP distribution category.

The first column in the Table lists the general DP pattern.  For example, "Dominant Aptitude" refers to the situation in which the most points are found in one particular aptitudinal group; Brilliant (B), Intermediate (I), Classic (C), Solid (S) or Professional (P).  This category is subsequently divided according to which aptitudinal group dominates and shows how Dosage figures and performance characteristics can vary from pattern to pattern.  The next category focuses on the Brilliant group exclusively and shows the effects of total Brilliant points in the DP.  The following category does the same for the Classic group and so on.

The second column displays the symbols for the DP patterns themselves and shows the relationships among the various elements within the DP.   For example, the first pattern under Dominant Aptitude is designated B>I, C, S or P, indicating a DP in which there are more points found in the Brilliant category than in any other.  Similarly, under Wing Breeding, B, P>I, C or S means that more points are found in both the Brilliant and Professional categories than in Intermediate, Classic or Solid.  Next, SWs indicates what percentage of the general population of open stakes winners since 1983 have that particular DP pattern.  This is followed by columns showing the average winning distance (AWD), average points in each of the aptitudinal groups (or average DP), the average DI (ADI), the average CD (ACD), the average total points in the DP (PTS), SPR (the percentage of sprint stakes wins in the category), TURF (the percentage of turf stakes wins in the category) and 2YO (the percentage of juvenile stakes wins in the category).  The bottom row highlights the data for all of the stakes winners in the database and may be considered the average for the breed, suitable as a standard for comparing the various DP patterns.

PATTERN SYMBOL SWs    AWD     B     I     C     S       P ADI ACD PTS SPR TURF 2YO
DOMINANT APTITUDE B>I,C,S,P 27.6% 7.71    10.66    4.26 5.99    1.20 0.53    5.21 1.06    22.64    42.7%    22.2%    15.6%
I>B,C,S,P 5.1% 7.94 5.21 9.36 6.08 1.04 0.55 4.73 0.81 22.23 34.6% 23.6% 11.7%
C>B,I,S,P 55.7% 8.29 6.22 4.07    12.57 1.37 0.83 2.17 0.54 25.07 25.3% 37.0% 12.0%
S>B,I,C,P 1.2% 8.82 4.42 3.01 4.54 8.73 0.66 0.83 0.03 21.37 14.4% 47.9% 9.6%
P>B,I,S,C 0.6% 8.98 4.12 2.48 5.11 1.12    11.16 0.65    -0.56 23.99 14.5% 55.2% 7.6%
BRILLIANT (B) PTS B=0 1.2% 8.27 0.00 2.61 6.66 1.59 1.41 1.56 -0.04 12.27 28.1% 32.1% 10.7%
B<=5 35.6% 8.22 3.40 3.61 8.69 1.34 0.88 2.38 0.48 17.93 28.1% 35.8% 12.7%
B>5,<=10 43.9% 8.09 7.77 4.38 9.90 1.33 0.75 3.33 0.77 24.14 31.0% 30.6% 13.1%
B>10,<=15 16.8% 7.93 12.43 5.63 11.11 1.49 0.68 4.23 0.94 31.35 36.4% 26.4% 12.9%
B>15,<=20 2.8% 7.72 17.39 6.81 11.11 1.45 0.63 5.42 1.08 37.39 45.5% 19.3% 16.1%
B>20 0.9% 7.83 23.40 6.92 12.79 3.47 1.04 4.55 1.03 47.62 37.2% 19.7% 13.4%
CLASSIC (C) PTS
 C=0 1.1% 7.69 6.13 3.05 0.00 0.70 0.57 4.39 1.32 10.45 44.0% 31.6% 15.6%
C<=5 23.2% 7.79 6.67 3.68 3.35 1.25 0.56 5.29 0.92 15.52 41.0% 25.1% 15.1%
C>5,<=10 38.2% 8.00 7.09 4.19 7.94 1.25 0.71 2.97 0.72 21.18 33.6% 27.4% 13.8%
C>10,<=15 24.4% 8.25 7.86 4.78 12.74 1.40 0.88 2.39 0.61 27.65 26.4% 35.7% 12.3%
C>15,<=20 9.8% 8.49 8.19 5.30 17.69 1.86 1.12 2.02 0.49 34.16 20.1% 41.8% 9.3%
C>20 4.4% 8.81 9.69 6.10 24.50 2.12 1.34 1.91 0.46 43.75 13.5% 50.6% 8.2%
HIGH B/C PTS B>10,C>10 10.0% 8.10 13.63 6.03 16.25 1.49 0.86 2.91 0.80 38.27 31.0% 29.0% 13.3%
"DOWN-THE-LADDER"  B>I>C>S>P 2.2% 7.71 12.22 7.46 4.80 2.05 0.12 5.32 1.11 26.65 45.7% 18.5% 15.1%
"DOUBLE ZERO"  S=0,P=0 37.7% 7.81 7.26 4.16 8.77 0.00 0.00 4.65 0.94 20.20 39.0% 24.7% 15.4%
"TRIPLE ZERO"   C=0,S=0,P=0 0.7% 7.67 6.36 2.95 0.00 0.00 0.00 1.68 9.31 45.1% 28.2% 16.4%
I=0,S=0,P=0 2.5% 7.94 5.15 0.00 6.60 0.00 0.00 2.92 0.93 11.75 35.5% 32.1% 16.6%
PTS IN ALL 5 GROUPS B,I,C,S,P>0 14.6% 8.50 7.47 4.78 11.69 3.11 2.46 1.76 0.38 29.51 20.9% 39.2% 9.1%
DOMINANT CLASSICITY S=0,P=0, C>(B+I) 9.2% 8.21 5.31 3.05 13.21 0.00 0.00 2.27 0.62 21.57 25.9% 38.7% 13.7%
C>(B+I+S+P) 21.4% 8.39 5.21 3.03 14.88 1.01 0.60 1.86 0.46 24.74 21.7% 42.9% 11.0%
DOMINANT C PTS
W/O DOMINANT CLASSICITY   		C>B, I, S, or P
but not
C>(B+I+S+P)		 34.3% 8.23 6.85 4.72 11.13 1.60 0.98 2.37 0.59 25.28 27.6% 33.3% 12.6%
TOTAL PTS PTS<=10 7.8% 7.87 2.89 1.56 3.26 0.38 0.27 3.64 0.77 8.36 37.4% 27.9% 15.4%
PTS>10,<=20 34.4% 7.92 5.60 3.10 6.56 0.82 0.48 3.68 0.76 16.55 36.3% 27.2% 15.0%
PTS>20,<=30 37.9% 8.15 8.01 4.79 10.53 1.53 0.83 3.02 0.69 25.69 29.2% 33.2% 12.2%
PTS>30,<=40 14.6% 8.37 10.21 6.33 14.79 2.23 1.32 2.63 0.63 34.88 24.7% 35.1% 10.3%
PTS>40,<=50 3.9% 8.51 12.88 8.71 18.77 3.06 1.54 2.61 0.63 44.96 22.4% 40.2% 9.8%
PTS>50 1.4% 8.37 15.68 9.71 23.86 3.13 2.18 2.47 0.61 54.56 25.5% 37.5% 12.8%
"WING" BREEDING B,P>I,C,S 0.4% 8.46 7.60 2.45 2.88 0.36 7.44 1.51 0.18 20.73 26.7% 41.7% 7.5%
ALL PATTERNS   100.0% 8.10 7.40 4.41 9.74 1.38 0.79 3.21 0.71 23.72 31.3% 31.3% 13.0%

We can examine in greater detail how DP distributions can affect overall type by examining several DP configurations for a given DI.  The DPs in the following Table are all equivalent to a DI of 3.00.  Note, however, the wide range of possible CD values depending on how the DP is configured.

DP CD
 0-3-0-0-1   0.25 
 0-3-0-1-0   0.50
 0-2-2-0-0   0.50
 2-0-2-0-0   1.00
 3-0-0-0-1   1.00
 3-0-0-1-0   1.25

In all cases, the CD range for any DI spans one full CD unit.  For DI 3.00, the median CD is 0.75, and the range is from 0.25 to 1.25.  It is reasonable to suspect that a horse with DI 3.00 and CD 0.25 may be quite different in type from a horse with DI 3.00 and CD 1.25.  In fact, we find real differences between those horses with DI 3.00 and CDs greater than and less than 0.75.  In the former case, 36% of the races won are sprints and the average distance of their wins is 7.88 furlongs.  In the latter case, the percentage falls to 30% sprint wins and the average distance rises to 8.24 furlongs.  Thus, we confirm that horses with the same DI are not necessarily the same type in terms of performance attributes. 

For reference, the general equation to find the median CD for any given DI is:

Median CD = [3 x (DI - 1)]/[2 x (DI + 1)]

The median, maximum and minimum CDs for some common DI values are displayed in the next Table.  

DI   MEDIAN CD   MAXIMUM CD   MINIMUM CD
0.00 -1.50 -1.00 -2.00
0.50 -0.50 0.00 -1.00
1.00 0.00 0.50 -0.50
2.00 0.50 1.00 0.00
3.00 0.75 1.25 0.25
4.00 0.90 1.40 0.40
5.00 1.00 1.50 0.50
6.00 1.07 1.57 0.57
7.00 1.13 1.63 0.63
8.00 1.17 1.67 0.67
9.00 1.20 1.70 0.70
10.00 1.23 1.73 0.73
20.00 1.36 1.86 0.86
30.00 1.40 1.90 0.90

Some important observations emerge from this analysis.  First, differences in DI at the lower end of the DI range are more significant than they are at the higher end.  For example, the difference in the median CD for DIs 1.00 and 2.00 is 0.50 CD units.  In contrast, the difference in the median CDs for DIs 10.00 and 20.00 is only 0.14 CD units.  Doubling the DI at the lower end has a far greater effect on the median CD than it does at the higher end.  Second, and most important, neither the DI alone nor the CD alone is sufficient for an adequate aptitudinal evaluation of a pedigree.  The critical component remains the DP, and its configuration is responsible for the interplay between the DI and the CD.  Consequently, all of the Dosage figure components - the DP, the DI and the CD - are complementary and all are necessary for the best and most accurate interpretation.

To emphasize this point, we can evaluate the results obtained from over 50 SWs on dirt with a DI of exactly 3.00 and at least five stakes wins.  These were divided into groups by CD range with Group 1 having a CD of 0.90 or more, Group 2 having a CD range of 0.80 to 0.89 and Group 3 having a CD range of less than 0.80.  Group 1 had 82 stakes wins among them.  Group 2 had 93 and Group 3 had 95.  A statistical t-Test assuming unequal variances (in order to determine if the Groups are statistically different in their average winning distance (AWD)) afforded the following for Groups 1 and 3:

     DI 3, CD 0.8 to 0.89    DI 3, CD< 0.8
Mean AWD 7.41 8.45
Variance 1.838301716 1.321668533
Observations 82 95
Hypothesized Mean Difference 0  
df 160  
t Stat -5.418164946  
P(T<=t) one-tail 1.08787E-07  
t Critical one-tail 1.654432253  
P(T<=t) two-tail 2.17574E-07  
t Critical two-tail 1.97490408  

The analysis clearly show that the Groups at the ends of the CD spectrum are different from each other in a statistically significant way as expressed by the P-values in each Table.  A P-value of 0.05 or less indicates statistical significance.

We can see the differences visually by charting the average CD of each Group (ACD) versus the average winning distance (AWD).  The correlation is outstanding, with a correlation coefficient of 0.97.