Ever wonder why EPDs can change? Before reading too far, check out this link provided by AAA describing and clearly illustrating this feature: https://www.angus.org/tools-resources/national-cattle-evaluation/accuracy-and-possible-change
Before we begin, I must confess this is not an easy topic to clearly articulate. Moreover, it requires getting a bit more technical than most of our newsletters get into. I'd encourage you to read this through, reread it, and then give me a call if you have any clarifying questions. I believe this is a critical topic and can help both registered and commercial cattlemen elevate their breeding toolkit to the next level. Let's begin.
"Possible potential" change describes how much an EPD can shift based on its accuracy. For example, if a given animal starts with WW EPD value of +80 (top 15%) but a very low accuracy of 0.05, then the possible potential change for this animal’s WW EPD is massive. Its WW EPD could change dramatically as progeny are reported, as a matter of fact, the EPD could shift +-15 pounds based on the Angus Potential Change chart (link attached). This is because the EPD is very very low accuracy, so until progeny are reported, the EPD is unstable and not reliable. That +-15 could put the actual WW EPD value to +95 (top 1%) or as low as +65 (50th percentile or ~breed average).
Now here’s the real kicker, as this animal’s progeny are continually reported, the accuracy will begin to increase but the possible potential change will begin to decrease because the sire’s passed on traits are becoming more predictable, regardless of whether they are better or worse. The predictably increases as the accuracy increases and the potential change decreases as the EPD prediction tightens or “settles” around what it actually is.
This “possible change” should actually give producers some comfort if they understand it and embrace this reality. This is a feature of EPDs, not a bug. This is how statistics work. Anything can be modeled and subsequent predictions can be made, but knowing how predictable the model is far more important. But in ALL cases, quality models increase in accuracy and predictability as more data is added (barring dishonest data or errors).
In practice, this can be extremely useful knowledge. For example, knowing this information can be helpful when evaluating the use of a calving-ease sire to AI heifers to. A high-accuracy (ie., 0.90 accuracy) high calving-ease sire (ie., +10 CED) has little room for additional "potential change" because the accuracy is so high. Meaning, you shouldn't expect too many surprises and you should have high confidence that the sire will indeed be "calving-ease."
Another less discussed example, though potentially just as valuable, is leveraging EPD accuracies and the concept of possible change for experimenting with complete outcross sires. This can inform you on how much “risk” you are taking when testing a new sire. As with most things though, with great risk, there can also be great reward. Let me explain. Often, lesser known pedigrees or international sires have “poor” or sporadic EPDs simply due to very few data points to build on, so the resulting EPDs are extremely low accuracy with a high likelihood for potential change. If you do your research on an outcross sire though, and you feel confident that he will transmit the traits you want but you are worried because his EPDs "suck" then don't lose all hope - yet. If his accuracies are 0.05 then there could be a massive positive EPD shift as those progeny are reported. If the progeny perform like you hoped in whatever trait you expected, then its EPDs would both improve and increase in accuracy simultaneously (high reward). BUT the opposite could also occur, because they are low accuracy, and if the sire doesn't produce the traits you expected, the EPDs values could decline while the accuracy's increase (high risk). Meaning, as their EPDs are getting worse they are also increasing in accuracy. For example, just as there are known sires that you can depend on for calving-ease based on their high accuracy (ie., 0.90) and high CED EPD values (ie., +10), there are also sires you can count on, with confidence, to absolutely wreck havoc on your heifers during calving season based on their high accuracy (ie., 0.90) and low CED EPD values (ie., -2). You KNOW you will likely have several "hard pulls" and likely lost calves and heifers because of it.
All of this also helps explain how some sire’s EPDs shift so dramatically the first year after their progeny birthweights, weaning weights, and yearling weights are reported. But, as the EPDs shift, they will begin to settle and tighten in parallel with the accuracy increasing. Connealy Craftsman, the sire with the most reported progeny in 2024 is a prime example showcasing the phenomenon of possible change. When he sold in March of 2022, he was advertised as a calving-ease heifer bull with a CED of +9 with an accuracy of somewhere between 0.25-0.35, meaning that it was possible his CED EPD could increase or decrease by as much as 7 to 8 points (based on the chart above). As of today, after 774 calving reports were submitted from his progeny, his CED now sits at a +1 with an accuracy of 0.83. A staggering 8 point shift that verifies he is certainly NOT a calving ease sire. Now, the reality is this doesn't happen often. Between data reporting and genomics, we have as much confidence in our predictions now as we ever have, but the fact is, these shifts CAN happen .Though, we shouldn't necessarily be shocked. If anything, we should accept this as a possible reality when using new unproven sires. Aftercall, the data told us this could happen when he first sold. New sires have greater risk, but again, they can also have greater reward.
High accuracy means low possible change. Low accuracy means high possible change. It doesn’t mean that they will shift, just that there’s a reasonable likelihood they can.
But remember, if you don’t report the data, then the EPDs won’t change accordingly! I hope this is helpful to the folks interested.
For the full article and others like it, feel free to check it out on our VICCI newsletter page: https://www.viccicattleco.com/newsletter
The attached video, curtesy of Chris Earl, is a resident VICCI herd sire, ZA Cowboy King (20954620) as a 4 month old calf. He was recently acquired as the top selling and record setting bull at Zumbrunnen Angus’s Annual Production sale. He’s an example of a young prospect that we feel confident in his ability transmit the traits we want into his progeny. Even still, with unproven sires there is both greater risk, but also potentially greater reward. You never know until the progeny prove it!
Happy Breeding
Josh Bedell,
President & Owner
VICCI Cattle Co.
Comments