A New Fitness Value System
hackerhmb.tumblr.com
hackerhmb.tumblr.com
Tangentially, the question of "who is the fittest?" comes up from time to time in various places. It's a fairly meaningless question of itself, but fully explaining why leads to roaming over stuff like multi-objective optimisation problems.
Which is neat if you like that sort of thing.
MOOPs generally don't have an optimal solution. There's a pareto front of good solutions, but you have to decide what you value.
In physiological terms, you can't have it all. You need to make tradeoffs. The adaptations for strength are competitive with adaptations for endurance, and so on. You can't be very strong and easily run a marathon.
Now, suppose we decide to develop an index of fitness. Any such index could only be tested according to specific tasks, which introduces the problem of task selection. Because you can quickly optimise those physiological tradeoffs to maximise index performance. If for example your fitness index has lots of running and swimming, endurance athletes will dominate. If there's lots of lifting and pushing, strength athletes will dominate. Average it over all possible tasks and I suspect the No Free Lunch Theorem will apply and everyone will suck equally.
There is no "fair" index because the entire concept of "general fitness" is nonsensical. All fitness is fitness for a task or group of tasks.
This is true of almost everything in life, because almost all the hairy problems we face are about trading off multiple objectives. The important lesson is learning that sometimes, you simply cannot have everything you want.