/sci/ - Science & Math

>>15123121
>which seems the exact same as what I said but with a missing step (is it?)
When [math]x[/math] goes to minus infinity [math]e^x[/math] goes to zero, so I'm not sure why you're discarding the [math]\exp (-1/(2 x^2))[/math] factor when looking at the asymptotics, you just worsen the estimate.