Coding classes in East Lothian

The model is invented, not a survey of the Bass Rock. Fish grow on their own and are eaten when they meet gannets; gannets thrive when they catch fish and decline without them. In symbols, the fish change at a rate of x minus 0.02xy and the gannets at 0.01xy minus 0.8y. Starting from 60 units of fish and 40 of gannets, the learner steps the equations forward in time with a method called RK4, in hundredths of a year, for fifty years.

With RK4 the two populations rise and fall forever in a steady loop. Fish peak every 7.09 years, and gannets peak about 1.52 years after each fish peak: the predators need time to breed on the back of plenty, by which point the fish are already falling. Neither ever settles at the balance point of 80 fish and 50 gannets unless it starts there. The lag between prey and predator peaks is the signature of this kind of system, and ecologists look for it in real records.

Now the check. These equations have a hidden quantity that stays exactly the same along any true solution, a mixture of the two populations and their logarithms. RK4 keeps it constant. The simpler Euler method, which takes each step in a straight line, lets it creep upwards, and with larger steps the cycles spiral outwards: fish peaks grow from 120.7 to 175.7 over the fifty years. The graph still looks like a plausible ecosystem, yet it is an artefact of the method. A quantity that must not change is one of the strongest tests a simulation can have.

That is why coding still matters for an East Lothian teenager in 2026: it teaches you to question a simulation before trusting it. The longer argument is in why teenagers should still learn to code in 2026.