Talk:Nekhoroshev estimates

Estimate formula
The formula presented just can't be right! Obviously, the closer our trajectory should stay to the invariant torus, the shorter the possible time. Thus upper bound should decrease with $$\varepsilon\to 0$$, while it increases! I changed the sign so that makes now more sense (exponentially fast drift away), but the whole formula should be revised, 'cause there's the same mistake in the cited source!

By the way, when $$\varepsilon\to\infty$$ T should go to infinity (farther distance requires more time, on compact manifold there's even an upper distance bound!) It's *very* strange that in the current formula upper time bound tends to a constant. --Anton (talk) 23:38, 2 December 2010 (UTC)

Sorry for the buzz, I didn't get the first time that there's the same epsilon in Hamiltonian and estimates. Well, maybe the statement should be clarified a bit? --Anton (talk) 00:20, 3 December 2010 (UTC)