The Lieb–Thirring inequality revisited