This short note intends to go through the numbers that characterize the precision of our models and compare them with those in the industry.
If you have been exploring our site, you will find proven models that have been thoroughly tested (see https://championbred.com.au/why-championbred/ for more details). That means that if you use our tool, your chances of finding a great horse will increase significantly. So how much of an edge are we talking about, and how often does it show up? Let’s start with the second one. When we test how likely it is that someone could beat the model’s selection by chance, the probability is less than 0.5%. To put that in context, in many scientific settings, it is a common convention to use p<0.05 (less than 5%) as a threshold for statistical significance.
What about how strong the effect is? That depends on how the reports are used. Top results tend to deliver 200% to 300% better than random selection within a given auction catalog or sample. For example, at an auction with 7.5% SWs, top reports reach 25% to 30%. Of course, these values oscillate from auction to auction, but we got this by testing with Australian auctions the models had never seen. We have made recommendations to one of our clients for auctions during 2025, so we will soon start having numbers to compare to these previous tests. For more about this, see our next article about previous results at auctions (reference 2).
We can compare this with the stallion fees and their SWs percentages. Top stallions present numbers close to 15%, with most of them not even reaching 10%. You can end up spending a lot of money for what is still, in the end, a fairly modest improvement in your chances. Even among the very top end of the market, horses costing more than 1 million AUD (see reference 1) average about a 13% stakes winner rate, while the average for horses below 100000 AUD is around 4% stakes winner rate.
The line added to the graph to explore the correlation between stallion fee and SWs% has an R of 0.62, indicating a moderate relationship. Assuming auction prices follow stallion fees, for every 100000 AUD spent, the buyer will get an average 2% extra chance. This gets much worse if we consider the 13% SWs that millionaire horses tend to average.
Let’s remove the most expensive stallions and zoom in. After doing that, the correlation almost disappears, dropping to 0.08, suggesting a disconnection between the variables.
What recommendations emerge from this simple analysis?
- Searching for affinities and using tools for patterns should pay off better than selecting according to stallion fees;
- Factors that determine market prices seem to increase the chances of success, but only slightly. For example, an obvious mathematical conclusion if you have 1 million AUD available should be: do not buy one expensive horse, but several cheaper ones. For example, if you buy 6 horses below 100000 AUD (assuming 4% for each), your overall chance will go up to 18.5%.
- Cheap or new stallions with high-quality mares might present excellent chances.
- Intermediate stallions, often forgotten and out of fashion, can create interesting opportunities.
Write us with your opinion to guide us towards new short articles and get extra tokens to explore Championbred’s capabilities better.
References: