The bowler who refuses to get worse ends up deciding how everyone else bowls.
You learn to bowl without putting your fingers in the ball. Your average climbs fast, because the early gains are easy, and then it stalls at 160. Someone who bowls seriously tells you that to improve you need a different grip and a different style, and that you will get much worse before you get better. You look at your 160, which wins most casual games, and decide it is good enough. Ten years later you have ten years of experience, or the same year of experience ten times.
Nothing in that story is unusual. The puzzling part comes later, when the bowler who stopped improving ends up running the league. Erik Dietrich, writing about software teams, called this person the Expert Beginner. The question is how a plateau, which is ordinary, becomes the standard a whole group is held to.
The stage where you can’t see the gap
Dietrich starts from the Dreyfus model of skill acquisition, which runs from Novice through Advanced Beginner, Competent and Proficient to Expert. The crucial step is from Advanced Beginner to Competent, because Advanced Beginner is the last stage where you have no understanding of the big picture. A Competent practitioner knows what they do not know. An Advanced Beginner does not, so an Advanced Beginner can confuse themselves with an Expert in a way that a Competent practitioner never could.
Advanced Beginners break one of two ways: they move to Competent and start to grasp the big picture, or they “graduate” to Expert Beginner by assuming they have already graduated to Expert.
The plateau itself isn’t the danger. The danger is the decision to redefine the plateau as the summit. Once the bowler has made it, anyone bowling differently is not better, only wrong. The decision is easy to make because the plateau is Locally Optimal: 160 does win casual games, the code does ship, the team does deliver. Every path away from the current method begins by getting worse, and the Expert Beginner has built an identity around never getting worse. It helps that Expert Beginners usually reach their station through default rather than merit. They stuck around, so they have never had to persuade anyone of the value of their ideas, and they sit at the unskilled end of the Dunning-Kruger effect, where “if I don’t understand it, it must be easy.”
How one plateau becomes the team’s
The individual Expert Beginner is a nuisance. The group dynamic he creates is the real damage, and it starts once he controls hiring and mentoring. The group learns two lessons:
- If you wait your turn, you can wield unquestioned authority regardless of merit
- It is okay and even preferred to be mediocre here
He doesn’t consciously seek worse candidates. He rationalizes that ambitious people “aren’t a good fit.” The talented leave because they won’t tolerate the ceiling, and the less talented stay because they are grateful for the job. Bruce Webster called this the Dead Sea Effect: the organization evaporates its best people and concentrates the residue, and the worse the group gets, the harder it is to attract anyone who could fix it.
Challenges from inside meet a pivot pattern that cannot be falsified. Is the newcomer’s solution simple? Too simplistic. Complex? Over-engineered. Is the newcomer younger? Rookie mistakes. Older? Out of touch. Faster? Careless. Slower? Inefficient. The goal is not to evaluate the solution but to defend the Expert Beginner’s position as the standard of measurement. Picture the bowler, now captain, explaining to every recruit with a proper grip why their scores don’t count yet.
Why better metrics won’t fix it
The natural response is to blame a few bad promotions, or to call for better performance metrics. But the Expert Beginner is not a hiring failure but an emergent property of any system that rewards tenure over demonstrated growth. Better metrics won’t help, because he controls the metrics. What works is exposure to external standards: conferences, open source, cross-team rotations, anything that forces the local standard to compete with the global one. The Expert Beginner survives by keeping his environment closed, so the fix is to open it.
Opening the group helps, but the cause sits in every learner, because the Expert Beginner isn’t a different kind of person. The valley between Advanced Beginner and Competent is where every learner has to accept being worse than they thought. Grokking gives that valley a shape: the jump from memorized pattern to compressed rule tends to arrive long after the learner seems to have stalled, and every felt signal during the stall (boredom, frustration, the sense that more practice is wasted) says to stop just before the jump would have landed. The Expert Beginner obeyed those signals and then built an identity around the decision. Most of us have a 160 somewhere. The open question is whether we have also put ourselves in charge of the league.
References:
- Erik Dietrich, How Developers Stop Learning: Rise of the Expert Beginner (series)
- Bruce F. Webster, The Dead Sea Effect