The truly bad sometimes find ways of covering their tracks. I've seen bad newbies, and bad oldies.

Even the "average" bad hire costs a lot more than what the article suggests. $15K for a $50K job seems very low. What's the value of the "average" bad hire turning a client reference into an angry client? Or having them hire someone very bad for one of their positions? Or undermining company morale?

What is the probability of an "average" bad hire doing those things? That was at the core of my original point. This seems to come up on HN every time a hiring discussion moves into talking about "bad" hires. People bring up realistic negative possibilities without attaching a probability to them, making it impossible to determine a ballpark expected value, or in this case cost, of the action. Now, for various reasons calculating an accurate probability is virtually impossible as well, but one should be able to come up with a rough idea of what it is. Additionally, every company should have some sort of controls in place to minimize the damage a new hire could do, even if the action is a genuine mistake or intentionally malicious.

Good question. Some of it is quibbling on words. My general take on people's ability is it's a fat tailed distribution on both sides.

There's a middle that is fairly tight in ability, and they perform good or bad largely dependent on the circumstances system (including management) that they're put in. The upside and downside surprises can go very far on both sides of the mean. Which means slightly bad hires tend to be products of a bad system, which is what should be fixed. But very bad hires can do an inordinate amount of damage even relative to their small probability. P(x) * Value(x) comes out to a very large negative #. (Similarly P(x) * Value(x) comes out to a very high absolute # for the great hires)

Now what's the probability of a truly awful hire? Depends on how well you hire. I suspect it's somewhere between 1% and 10% based on the organization. Same on the upside.