4 ms·
I feel like that example assumes the conclusion. A better analogy would be "spinning objects theories". It's a category so broad you could always find something
by ravi-delia 4y ago
I feel like that example assumes the conclusion. A better analogy would be "spinning objects theories". It's a category so broad you could always find something which works, but since it's not really thought of as a particular solution, it isn't so vulnerable to overfitting. Epicycles were bad largely because they gave a false impression of closing in on reality; every new round of additions consisted of smaller and smaller tweaks. Spinning object theories are even more broad and adaptable, but since they aren't basically one solution you don't have the same issue. When Kepler found a spinning object theory which worked, it wasn't overfitting.
- lisper 4y agoWhat on earth is a "spinning object theory"???
- ravi-delia 4y agoPlanets and sun in epicycles around the earth, planets and moons in cycles around the sun, planets in cycles around the sun and moons in cycles around planets, that sort of thing. I used it as an example because "supersymmetry" is more a descriptor, than a theory unto itself. If you make your "theory" big enough it's going to technically be unfalsifiable, but that's more because you drew too large a bound than any mark against it.
- lisper 4y agoAh. OK, in that case... > When Kepler found a spinning object theory which worked, it wasn't overfitting. Well, yeah, that's true, but the important thing about Kepler was not the fact that it was a "spinning object theory", it was the switch from a "circular theory" to an "elliptical theory". (Also the switch from geocentrism to heliocentrism, but that was Copernicus, not Kepler). This is the sort of thing that the supersymmetrists are hoping to achieve, but have not yet achieved, most likely because nature is not supersymmetric in point of actual fact.