Consider the sex of six babies born in a hospital. Each birth is a random event—whether the first baby is a boy or a girl has no effect on the sex of any of the next five babies. With that in mind, let's take a look at three possible birth sequences:
BBBGGG
GGGGGG
BGBBGB
Are they all equally likely?
Our intuitions tells us that they are not. Six girls in a row is so improbable! But each sequence does indeed carry the same probability. Because births are independent events and B is as likely as G, there is no reason for one sequence of six to be any more or less likely than another.*
Even if we can accept this as true it remains counterintuitive. This is because our brains are programed to find, or if necessary generate, causation. This process is reflexive, but it can get us in trouble by inhibiting our ability to spot randomness. It leads us to jump to a conclusion—"non-random!"—as soon as we spot any sort of pattern.
But random systems produce things that look like patterns all the time. Famously, the bombing of London during WWII leveled much of the city but left certain portions untouched. This led some to believe that German spies were located in these areas. But statistical analysis later showed that the gaps in bombing were the expected result of random strikes.
In summary: when we detect patterns, we assume they can't be the result of randomness and look for other causes. And unfortunately, we often think we find them.
The Draft is Random
The NFL draft is largely a random event. We are fairly good at sorting players into general categories (first round talent, second round talent, late round talent, etc.). But outcomes within each category are essentially random. Even using a relatively low benchmark for success, the hit rate for first round picks is approximately 53%. However, it is easy to see some patterns in the chaos. And when we see patterns, we imagine causes. And once we imagine causes, we reject the idea that the process is random.
Of course, it's easy to find causal explanations for draft outcomes. The player was hardworking. The player was immature. The team that drafted him had a scheme that maximized his talents. He played in a pro-system that made the transition from college to the NFL smoother. His injury history put him at high risk of missing time. And on, and on.
Explanations for both successes and failures are incredibly easy to generate. And that's the problem. The ease with which we can explain draft results makes it exceedingly difficult for us to acknowledge the fundamental randomness of the event.
Tiers Not Rankings
Nobody is better than anyone else at figuring out which college players will succeed in the NFL. As 538 puts it:
None of [the NFL's] individual actors have the ability to “beat the market” in the long run. Some do see short-term deviations from the mean, but those prove unsustainable over larger samples. The implication is that much of what each team gets from its draft picks — the very entryway to the league for almost every NFL player — is determined by pure chance.
It pains me to write this as someone who spends hours scouting college prospects each year. But we must approach the process with humility. The most we can do is group players into broad categories, or tiers. This is a valuable exercise, so it's not to say that time spent grinding film is worthless. But creating tiers is the most we can do—the data shows us that attempting to rank prospects within these broader categories is futile.
In the face of this randomness, the very best thing you can do is trade back within a tier. Take more swings at players with similar career prospects. This is true whether you're the GM of an NFL franchise or the manager of a fantasy football team.
We are degenerate, irrepressible see-ers of causation. Know that, resist it, and acknowledge randomness, as uncomfortable as it is.
*Example from Thinking Fast and Slow, by Daniel Kahneman