You Can't Serve Two Masters

For about a year my skill list looked like a wish list: handstand, planche, front lever, back lever, muscle-up, pistol squat, L-sit, a bit of climbing on the side. I trained all of them a little, most weeks. I improved at none of them in a way I could feel.

My handstand only clicked when I parked almost everything else and became a handstand specialist for a few weeks. Short daily practice, one press-strength exercise, base strength kept ticking over. Within a block I’d moved further than in the previous six months.

Interference

Coaches call this the interference effect. Goals that draw on the same resources compete. Planche and handstand both hammer the wrists and shoulders. Heavy leg days sabotage the pull session after them. A long run on Saturday steals from a Sunday strength test. Each goal is fine on its own; together they split your recovery so thinly that none gets enough.

The rule I train by now: two push, two pull, two leg goals at a time — and only two or three of those are skills. Everything else is parked, written down, and waiting for the next block.

Teams have a recovery budget too

In engineering, the shared resource is attention. Every in-flight initiative costs context switches, review time, meetings and mental load. Five half-finished projects deliver nothing; two finished ones deliver two things.

This is what work-in-progress limits are for. Kanban boards cap the “doing” column not because people are lazy but because throughput drops when everyone juggles. The same logic scales up: a team with three goals a quarter outperforms a team with a twelve-line roadmap of parallel bets, because the three get finished.

Practical shapes it takes:

• WIP limits on the board, visible and respected. When the column is full, the next move is to help finish something, not start something.
• A short goal list per quarter. Two or three outcomes, not a backlog dump.
• An explicit “parked” list. Parking isn’t the same as dropping. In training I keep parked skills in my log so nothing is lost; teams can do the same with a visible “next block” column.
• A prioritisation tool for the incoming flood — something like the Eisenhower Matrix — so the parked list stays honest.

How to choose what to serve

Ask what the goals compete for. If two goals load the same joints — or the same reviewers, the same service, the same on-call rota — they probably shouldn’t run in the same block. Pair goals that use different resources: a handstand block sits happily alongside leg strength; a frontend performance push sits happily alongside backend capacity work owned by someone else.

Where the metaphor breaks

A body can’t be in two places, but a team can split. Sometimes the right answer isn’t “fewer goals” but “fewer goals per person”, with clear ownership. The principle survives; the unit changes.

And some work refuses to be parked. Incidents, security fixes and regulatory deadlines are the equivalent of a sudden injury: they take priority. Leave slack in the plan for them rather than pretending they won’t happen.

Takeaways

• List what each goal consumes; don’t run competing goals in the same block.
• Cap work in progress and finish before starting.
• Park ideas visibly so saying “not now” doesn’t feel like saying “never”.
• Specialise for a block, then rotate. Progress comes from depth, not breadth.

Related: Hypertrophy and Skill Phases: Build vs Harden Cycles (the same idea over a quarter) and Learning When Everything Wants Your Attention (the same idea applied to learning).