Re: Probability in an infinite sample space

Robert Low wrote:

Han de Bruijn wrote:

But e = 1/n and that depends on n .

No, e is an infinitesimal. That's just a very small number,
remember?

Isn't e = 1/n just a very small number iff n is very large?

I thought that was my point. You're letting n tend to infinity
*after* you use your purported infinitesimal: but the infinitesimal
isn't defined until after you take the limit.

The infinitesimal is defined _before_ you take the limit. It disappears
_after_ you have taken the limit. It becomes zero then, remember? Don't
take limits with infinitesimals that have already vanished. Dangerous !

Hey, *you* are the one trying to define a new framework (infinitesimal
probabilities). The rest of us are quite happy to observe that standard
probability doesn't work here, and use a viable alternative.

Han de Bruijn

.

Relevant Pages

  • Re: Probability in an infinite sample space
    ... of picking 39 is also zero. ... specified multiple of 3. ... zero, but infinitesimals. ... You *cannot* assign uniform infinitesimal probabilities ...
    (sci.math)
  • Re: crank crank crank
    ... > Robinson's infinitesimals), as opposed to the ... But non-standard analysis is standard mathematics;-). ... of what they are meaning. ...
    (sci.math)
  • Re: Calculus XOR Probability
    ... I will not claim that such infinitesimals are measurable quantities. ... physicists will use infinitesimals in a naive way, ... Dirac delta-functions lying about. ... Poe and Robert Low, mathematical formulas do no longer express what they ...
    (sci.math)
  • Re: crank crank crank
    ...  > Robinson's infinitesimals), as opposed to the ... This was pointed out by several posters. ... NSA is still "standard mathematics." ...
    (sci.math)
  • definitions ha
    ... Dik T Winter wrote: ... >>> Robinson's infinitesimals), as opposed to the ... And when they use standard terminology ... of what they are meaning. ...
    (sci.math)