You have seen it happen, maybe to you.
A young bull comes up for sale with a set of numbers that stop you cold. His growth figures sit near the top of the breed. His carcass numbers are stacked. The dollar index is high enough that you circle the lot. You buy him, or a few straws of his semen, and you breed a good set of cows.
Then the calves hit the ground. And they are fine. Just fine. Not the calves those numbers promised. A little lighter than you hoped, a little plainer.
So what happened? Did the numbers lie? Did the seller pull one over on you?
Usually, no. The numbers did exactly what they were built to do. You just read them for something they never claimed to be. This piece is about how to read them right, so a high number never fools you the same way twice.
Before the statistics, hold onto something you already own. You can already tell a better animal from a poorer one. When you have two to choose from, you size them up by weighing what you know about each. Nobody taught you a formula for it. An EPD is that same everyday judgment, done bigger. It lets you compare cattle you will never stand next to, with the environment stripped out so the comparison is fair. For the definition of each trait and each dollar index, see our guide to the traits. Here we go after the meaning behind them.
Read the numbers like a weather forecast
Here is the frame that makes the rest click. When the weatherman says seventy percent chance of rain, you don’t hear a promise. You hear three things at once. His best guess. How sure he is. And in the back of your mind, that it could still go the other way. You plan the day around all three.
An EPD works the same way. It is a forecast about an animal’s future calves. Like any honest forecast, it comes in three parts: a best guess, a measure of how sure that guess is, and a spread of what could actually happen. Most confusion about these numbers comes from grabbing the first part and ignoring the other two. Hold all three at once, and the whole system makes plain sense.
The best guess is a difference, not a measurement
An EPD is an Expected Progeny Difference. Expected means predicted, not measured. Progeny means calves. Difference means it is measured against another animal, not on its own.
So an EPD does not say how much your calves will weigh. It says how much heavier or lighter one bull’s calves should run compared to another’s, bred to the same kind of cows. If Bull A has a weaning weight EPD of +80 and Bull B has +60, you expect Bull A’s calves to wean about twenty pounds heavier on average. Not eighty pounds. Twenty. That difference is the whole point.
That word difference also tells you what “zero” means. Zero is not average, and it is not nothing. It is a flag planted in the ground years ago, called the genetic base. Every EPD is measured as a distance from that flag. The breed has kept moving since, so a +50 today does not mean what a +50 meant when your father was buying bulls. And another breed set its flag in a different spot, which is why you can never lay an Angus EPD next to a Hereford EPD and call it a fair fight. An Angus number only means something next to another Angus number.
One more piece hides in the name. An EPD predicts the progeny difference, which is only half the animal’s own genetics. A parent hands each calf a random half of its genes, like dealing half a deck. The EPD is built around that half, which is why a calf never comes out exactly like the paper says.
How sure we are: accuracy
The best guess is only half the story; the other half is how much to trust it. That is accuracy. It runs from 0 to 1. The higher the number, the more real information stands behind the guess.
Think about two used trucks with the same price on the windshield. One comes with a fat folder of service records, the whole history. The other comes with the seller’s word and nothing else. Same price, very different purchase. The folder is accuracy. It doesn’t change the price on the glass; it changes how much you can trust the price is right.
A young bull’s number leans mostly on his pedigree. As he sires calves and those calves get weighed, real data pours in and his accuracy climbs. A proven bull with hundreds of calves carries accuracy near the top; a yearling with only his parents behind him sits low.
So a weaning weight EPD of +90 at 0.30 accuracy and a +90 at 0.90 accuracy are not the same purchase, even though the page shows the same number. The first can still move a long way as data comes in, up or down. The second is close to settled. You are buying two different levels of certainty for the same headline number.
Genomic testing changed things for young animals. A DNA test reads tens of thousands of markers across an animal’s genes. Run through the Association’s single-step evaluation, it raises a young animal’s accuracy before that animal ever breeds, buying a yearling about a first calf crop’s worth of information ahead of time. It does not change what the EPD means. It just lets you trust a young animal’s number sooner, so you guess less on cattle you haven’t seen produce.
The spread: why the calf is never exactly the paper
Now the third part, the one most people skip. Every honest forecast has a spread, and the best guess sits in the middle of it, not at the edge.
When you plan a mating, the best guess for the calf is the average of the two parents’ EPDs. Sire at +80, dam at +60, the expected calf sits at +70. That is the parent average. But no calf inherits the exact average. Each calf gets its own random half from each parent, its own hand off the deck. One full brother draws a strong hand; the next draws a weaker one. Cattlemen call this Mendelian sampling. You have seen it your whole life without a name for it: the two full sisters out of your best cow who turned out nothing alike.
So the parent average is the center of a scatter, not the outcome. It tells you where the calves will cluster on average, not what any single calf will be. That is the honest reason a mating is never a guarantee. You are buying a spread, and the paper prints the middle of it.
That scatter of calves is one kind of spread, and it is there no matter how well you know the bull. Even a bull with hundreds of calves behind him throws a strong one and a plain one out of the same cow. Accuracy governs a second kind of spread. It is not about how the calves fall around the average; it is about how well we know where that average sits to begin with. A well-proven bull’s true figure sits in a tight range and barely moves as new calves come in. A barely-proven bull’s true figure could still land a good ways off in either direction. Low accuracy, wide range; high accuracy, tight range.
Why the hot young bull sometimes breaks your heart
Put accuracy and spread together and you finally understand the young bull everybody chased and nobody could explain later.
That flashy young bull’s EPD is already a cautious estimate; the evaluation does not throw out a wild number just because an animal is young. So the number is honest. The catch is the spread around it. Low accuracy means a wide range of where it could settle, higher or lower. Stack Mendelian sampling on top, because each calf draws its own random half, and you have a wide spread sitting on another.
Here is the part that catches good cattlemen. Every year a whole crowd of young bulls carries big, unproven numbers. Chase the single most extreme one, and odds are you grabbed a bull riding a lucky short record, sitting on the high edge of its own spread. As his calves come in, that pick tends to come back toward the pack. That is not the number lying to you. It is what happens when you single out the most extreme bet from many soft ones. Same reason a rookie’s hot three weeks tells you far less than a veteran’s twelve-year career average: a short record can swing either way, while a long one has already settled toward the truth.
None of this makes young bulls bad bets. It means you size the bet to the certainty: spread the risk across several young sires instead of the whole herd on one, and lean on genomic-tested bulls. The point isn’t fear. It’s knowing which part of the curve you’re standing on.
What you can and can’t push on
Not every trait passes to the next generation with the same strength. Heritability is the share of what you see in an animal that actually gets handed down, rather than coming from feed, weather, or luck. Some traits are strongly heritable, so what you select for shows up quickly in the calves. Growth and carcass traits run moderate to strong: carcass and yearling weight sit toward the high end, while weaning weight is more moderate. Others are weakly heritable, so selection moves them slowly no matter how hard you push.
Fertility traits, like whether a heifer breeds and calves on time, are low and slow. That is not a flaw in the numbers, it is biology. Most of whether a heifer breeds comes down to how you developed her and what the grass did that year, not her genes alone. So the fertility EPDs are real and worth using, but they move like a heavy gate on a muddy hinge, steadily over years.
And here is the twist worth sitting with. The low-and-slow traits are exactly the ones where a good number, backed by DNA, is worth the most, because your own eye can barely see them. You cannot look at a heifer and know how her daughters will breed. The number earns its keep hardest where you are most blind.
The Association publishes its heritabilities and genetic correlations, and re-estimates them as the database grows. Either way, you can’t just “select harder” and get everything at once. Low-heritability traits move at their own pace, which is one reason you build a program over years instead of chasing it all in one bull.
Garbage in, garbage out
An EPD is only as good as the honest records and fair comparisons feeding it. The fairness comes from a contemporary group: calves raised together, same sex, same herd, same season, fed and weighed the same. Inside that group, the differences you measure are close to genetic, because everybody shared the same environment. A heavy calf in an easy year does not get credited for genetics he does not have.
This is why a single calf reported by itself, with no herdmates to stand against, tells the evaluation almost nothing. Feed the system careful records and honest weights, and the numbers earn their trust. Feed it sloppy or lonely data, and the number is only wearing a costume of precision.
You can see this in interim EPDs, the ones shown with an “I” in front. When a young calf has no contemporaries and both parents have real EPDs, its interim number is literally just the average of its two parents, carried at a very low accuracy. That is a placeholder with a fancy label, the parent average and nothing more, until the animal’s own records come in. Knowing that keeps you from paying proven-bull money for an unproven guess.
Why the numbers are hard to fake
The contemporary group does one more thing, and it is the quiet reason the numbers can be trusted at all. It makes them hard to fake.
Picture two herds. One is fed hard and weans three bull calves at 700, 750, and 800 pounds. The other is on tough country and weans three at 500, 550, and 600. Very different cattle, you would think. But the evaluation does not keep the weights. It keeps how each calf did against its own herdmates. Subtract each herd’s own average and both herds tell the same story: one calf fifty pounds under, one right on the average, one fifty pounds over. The extra 200 pounds of feed in the first herd told the evaluation nothing. It lifted every calf together and washed out.
That is the heart of it. You cannot buy a good EPD with a feed truck. Anything you do to the whole group, better grass, more feed, a growth implant, helps every calf the same and cancels. Only the gaps between animals raised side by side survive, and gaps are the one thing feed and bragging cannot manufacture.
Inside a group the math pins you down further. Those gaps have to add up to zero. Every group has a top, a middle, and a bottom, so you cannot have all your calves above their own average any more than every calf can stand taller than the herd. “They are all my best ones” cannot be true of a whole herd. The numbers spread out because they are measured against each other.
And you do not own the ruler. Your cattle are placed on one breed-wide scale, tied to every other herd through the same AI sires everyone breeds to. If your young bull lights it up in your pen but his half brothers come in ordinary across twenty other herds, the evaluation sees all of it and settles him in the middle. Hiding your poor calves does not help either: drop the bottom of a group and you only raise the bar for the ones left, so your good calf is now just average among good calves.
None of this makes cheating impossible. A man can still falsify a weight or fudge a birth date to slip a calf into a softer group. What the structure does is make the quiet, everyday kind of number-padding pointless, and make the outright lies collide, sooner or later, with how those genetics perform in everybody else’s pens. A herd truly can run above breed average, when the genetics are honestly there. You just cannot feed or talk your way to it.
Why chasing one number backfires
Now the economics, because the dollar indexes grow straight out of everything above. Say you decide growth is king. You buy the highest yearling-weight bulls you can find, year after year, and ignore the rest of the page. Your calves do get bigger. But your cows get bigger too, because growth and mature size travel together. Bigger cows eat more, every day of their lives. Your feed bill climbs, your stocking rate drops, and the cows that used to hold flesh on grass now need help. You chased one number off a cliff and dragged three others with it.
That is the trap. Traits are correlated. Pull on one and others come along, some the way you want and some the way you don’t. Milk is the sharpest example. More milk sounds like more pounds weaned, and up to a point it is. But milk has an economic optimum, not an “up is always better.” Past a point, a heavier-milking cow costs more to keep than the extra weaning weight returns, and in a dry year she falls apart first. The best amount of milk is the amount your country can feed. You cannot fix this by staring harder at one column. You fix it by weighing many traits together, by their economic value, which is what the dollar indexes are for.
Dollar indexes: many forecasts, priced out in dollars
A $Value index bundles a whole set of EPDs into a single figure: expected dollars of profit per head, under a stated breeding goal. Each trait is weighted by how much it moves money toward that goal. It does the trade-off math for you, in dollars.
But keep the forecast frame in mind, because a $Value is two things stacked together. It is a genetic prediction, same as an EPD. And it is a set of price assumptions: what fat cattle bring, what feed costs, what a bred heifer is worth. Those assumptions get refreshed once a year against a rolling market average. So a $Value has a vintage, like a truck has a model year. When the market shifts away from the assumptions baked in, the dollars can be off even when the genetics are dead on.
The bigger decision isn’t chasing the largest index number. It is picking the index that matches your operation. Do you keep your own replacement heifers, feed calves out and sell on a carcass grid, or hand everything off at weaning? A maternal index rewards the cow-calf man keeping daughters. A terminal index rewards the man selling pounds and carcass. A bull who tops a terminal index can be the wrong bull for a man keeping heifers. Match the index to how you make money, and the biggest single number on the page stops mattering as much as the right number for you.
Accuracy pays here too, in one specific coin: less risk. A genomic-tested young animal is a lower-risk bet than an untested one of the same age, because more of what you’re buying has already been observed. You are not paying for a better animal, necessarily. You are paying for a narrower spread. On a big-money purchase, that is worth real dollars.
What the number will never tell you
Here is the honest part, the part that earns your trust in the rest. A number that pretends to do everything is lying to you.
The numbers are blind to a lot of what makes an animal worth owning. An EPD can’t see feet and legs, or whether he holds up for a fifth breeding season. The foot EPDs help with structure, but they do not replace standing at the fence and watching him move. An EPD can’t fully see fertility, the trait that drives profit harder than any single production number. It captures disposition only in part. It knows nothing about whether an animal fits your country and your grass, or what your buyers pay for. A bull can post an elite index and stand on bad feet, and a bull on bad feet does not stay in service long enough to sire that index into your herd.
And remember the two hard fences. A $Value models a market, and its price assumptions can shift under your boots between the day you buy and the day you sell. And every EPD lives inside its own breed, measured from a different flag than any other breed’s. You cannot line them up and read the gap.
What they genuinely give you
So why trust them at all? Because they do something your eye cannot.
They take decades of records and boil them into a number that lets you compare two animals you will never stand side by side. A bull in Montana and a bull in Georgia, born in different years, weighed against their own herdmates on the same honest scale. No cow sense eyeballs that across a thousand miles and three years. The numbers do it every week, when the national evaluation runs fresh.
They also take the risk out of what you can’t see until it’s too late. Carcass merit hides inside the animal until harvest. Maternal milk doesn’t show until a bull’s daughters raise calves. Longevity takes years to prove. The numbers, especially with genomics behind them, give you an honest early read on those hidden traits before you’ve committed a decade of breeding. That is the real gift: they de-risk the traits that used to cost a lifetime to learn.
The true place of the numbers
So where do they belong? In your hand, not in the driver’s seat.
An EPD is a decision tool, not the decision. The numbers narrow the field: they take a catalog of a hundred bulls you could never sort by eye and hand you the fifteen worth walking out to look at, and they tell you which carry traits you could never judge from the fence. Then your job begins. You walk the pen. You look at feet and legs and how he travels. You think about your grass, your winters, your cows, and the cow family behind him. The numbers cannot do that part. You can, and it decides whether the animal earns his keep. Hold the numbers and the animal in the same view.
Now put the disappointing bull to rest. He did not lie to you. You bought a low-accuracy number as if it were a high-accuracy one, and the lottery of genes plus the wide range around an unproven number did the rest. Read next time for accuracy, not just size. A young bull’s headline number is the rookie’s hot month, not the veteran’s career line.
One last thing, because it keeps the whole picture honest. No breeder computes these EPDs, not us and not anyone. They come out of the American Angus Association’s national evaluation, run across the whole breed every week, on more data than any ranch could hold. What a good breeder does is read them well and combine them with what a number was never built to see. At Deluisio Angus, that is the job: let the evaluation do what only it can do, and bring the cow sense that no evaluation ever will.
You already size up cattle in plain, everyday ways. The numbers extend that, they do not replace it. Now you can read them the same honest way and put the two together.
For the source records and methods behind this piece, see our methodology notes. Trait-by-trait definitions live in our guide to the traits. The common objections to all this, answered one at a time, are in the coffee-counter complaints about EPDs.
Leave a Reply