I saw a demonstration of this sum recently. It went like this:
S=1-1+1-1+...
=1-(1-1+1-...)
the periods indicating a continuation to infinity, it is assumed that:
S=1-S
2S=1
S=1/2

Then again, the alternating harmonic series converges to ln2 but by rearranging the terms in the equation, it also converges to (ln2)/2 and (ln2)/4 and so forth. The problem in your reasoning is that you take a partial sum as reference. I know that it's counter-intuitive just like 0.00000...1= 0 but the reason it gives that result is only because the sum is infinite. The moment it becomes finite, these results all go wrong.

CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.