Before we built Championbred, we sat down with a simple question: does a pedigree really have the statistical power people believe it has?
Everyone assumes it matters. We assumed it too and that is why we created this platform. But assumptions were not enough, so, we needed proof. Here we make a short summary of the full article you can find below.
Does sibling performance is more alike than for unrelated horses?
We tested this using race records from a sample of more than 60,000 Thoroughbreds. The answer is clear: yes, they do and the statistics leave no room for doubt.
How did we do it? We compared two groups:
Then we compared them against random pairs of horses from the same years and racing environments. If the pedigree had no real effect, these groups would look the same in their race performance. They didn’t.
How did we test it?
To make sure the results were not luck, we used two well-established statistical tools:
-Mann-Whitney U test: checks if one group ranks consistently closer together than another one.
-Kolmogorov-Smirnov test: compares full performance curves, not just averages.
Both tests are non-parametric, meaning they work even when the data isn’t perfectly clean or normally distributed, which is ideal for real racing results.
What did we find?
Across all performance levels:
-Siblings were significantly more similar to each other than random horses.
-This effect was strongest at higher levels of performance (top 20%, top 5%).
-Even 3/4 siblings showed clear, measurable similarity. Sometimes nearly as strong as full siblings.
-The similarity remained visible as horses aged, meaning the genetic signal didn’t fade.
The statistics were not weak at all. P-values fell far below the 0.05 threshold, confirming that the chances of these patterns appearing by accident are extremely low.
What does it mean?
Heritability is moderate, but it is measurable and statistically proven. For breeders, this means shared bloodlines don’t just matter in theory — they show up in real results on the track. Full and even partial siblings offer a genuine, measurable edge. In fact, ¾ siblings show a statistically strong effect, especially when they share the same mare. One plausible explanation is that mares vary far more in their performance than stallions, so having the right mare tends to matter more for overall probabilities.
Many of these points are broadly understood in the industry, but for us it was essential not to assume anything. We wanted to take careful, evidence-based steps to build a strong foundation for all our products. Our goal is to ensure that by using Championbred, you gain a significant edge and proving this was our first step.