Start with how reality works. An animal gains energy by finding food and spends
energy searching for it and catching it. The animal that survives is not the one
that finds the richest patch. It is the one that knows when a patch has stopped
paying, and moves.
To maximize fitness, an animal adopts a foraging strategy that provides the most
benefit for the lowest cost, maximizing the net energy gained.
Optimal foraging theory
You have the same problem with a much worse feedback loop. A segment looks rich, the
team settles in, and three years later the cryptography is excellent, the
integrations are half finished, and nobody ever priced what entering that patch
would cost. That is not bad luck. That is what happens when the cost of the search
is never written down.
Charnov's marginal value theorem gives you the rule. Leave a patch
when its rate of return falls to the average rate of the whole environment,
including the cost of travelling to the next one. Read it in both directions,
because the second direction is where good teams go wrong: the harder the next patch
is to find, the longer it is rational to stay in this one. The chart is that rule,
and the two ways people miss it.
Then look at the incentives, which is where these relationships are actually
decided. An hourly agency is paid for residence time, so it has no reason to tell
you a patch is empty. We price per milestone, so our incentive points where yours
does: reach the decision point, take the decision, and move.