It’s bull-sale season. As you comb through your options, whether you are selecting your next herd sire from amongst your go-to seedstock supplier’s offering, attending a sale or using Angus Media’s new global sale book search (see here), there’s a table I encourage you to use.

It’s called “Accuracy and Associated Possible Change,” and it can be found online at https://bit.ly/03AccAPC. You can navigate to it from www.angus.org  by choosing “Tools & Resources” in the upper navigation and selecting “Accuracy and Possible Change” from the submenu.

It’s a big table, but it is oh so important in putting differences in expected progeny difference (EPD) values into perspective when you are pricing your next herd sire. Bull A may excel in calving ease direct (CED) by 3, but is that enough to matter? Bull B may have 10 pounds (lb.) more weaning weight (WW) EPD, but is that enough to matter?

The answer is it depends — on your breeding objectives, but also on the accuracy and the associated possible change in those numbers.

Basic concepts

Accuracy (Acc) is easy. Think of it as the confidence we have that the current EPD will be the animal’s true EPD. Sires used heavily in multiple herds through artificial insemination (AI) may reach accuracy values of 0.95, or even 0.99, meaning we are 99% confident that bull’s EPD is his true value.

While the best genetics in the breed may be being born right now, they will start with EPDs that are pedigree averages of their sire’s and dam’s EPDs and an accuracy of 0.05. As registered Angus breeders submit data (individual performance, genomic profile and progeny records) to Angus Herd Improvement Records (AHIR^®), the accuracies increase. While most yearlings won’t have progeny, genomic testing can provide prediction insight equivalent to having progeny (see “Progeny Equivalents” table at https://bit.ly/03stakesPE) and get us to accuracies on yearlings of 0.35-0.45.

Possible change is really pretty simple, too. For each 0.05 accuracy for each EPD, there is a possible change value assigned. That value is 1 standard deviation (SD).

Animal performance within a population will fall within a normal distribution — that bell-shaped curve — with most animals (68%) falling within 1 SD of the average, 95% falling within 2 SD, and 99% falling within 3 SD.

Tied to an accuracy value of a bull’s EPD, it also gives us a window within which we can expect a chosen bull’s true EPD value to fall. For example, at an accuracy value of 0.35, the possible change value is 10.2 for WW. If our bull has a WW EPD of 67, we have a 68% confidence his true EPD will fall between 56.8 and 77.2 (67 ± 10.2). We would have 95% confidence his true EPD would fall within 46.6 and 87.4 (67 ± 20.4) lb.

Seem like a large window? At 0.80 accuracy, the possible change value for WW is much less — 3.1. Consider a bull with a 67 WW EPD at 0.80 accuracy. With 68% confidence, we would expect that bull’s true EPD to fall within 63.9 and 70.1; with 95% confidence, we would expect his true EPD to be between 60.8 and 73.2.

Another point, to be “statistically different,” values must be at least 1 SD apart. With a 0.35 accuracy, a WW EPD of 70 would not be significantly different from a WW EPD of 65. However, at a 0.80 accuracy there would be a significant difference.

Putting it into practice

So, knowing what we know, how do we compare the two sets of EPDs in Fig. 1? Let’s start with CED. Bull A has a 7.0 CED; Bull B has a 13 CED. At an accuracy of 0.35, the possible change is 6.7, putting Bull A’s 1 SD window at 0.3 to 13.7 and Bull B’s window at 6.3 to 19.7. With only a 6-lb. difference, the CED EPDs of these two bulls are not significantly different, but it’s close. How much problem calving ease is for your herd will determine if the difference is significant enough for you to pay more for.