>just had a calculus exam
>final two questions were obviously meant to be solved with specific methods (taylor, and elipson-delta)
>tfw couldnt remember either
>tfw thought i got the correct answer anyways
>tfw just wrote a short python script that shows my answer was wrong by a factor of 4000
should i just an hero?
Just had a calculus exam
>elipson-delta
nobody needs that anyway
>taylor
how can you not remember that?
>professor said there was no point learning elipson-delta. But one of the problems was about proving a limit, how tf do you do that without elipson
idk if taylor was meant to be used for the second question but it about approximating a series with an error margin less than 1/100. i know the normal taylor shit
>fucking cucked by my professors desu
"Elipson-delta" is not a mathematical concept, so your professor was right about that.
Many ways. Just computing the limit normally is a proof, you can also use stuff like finding two other functions with the same limit, one that is strictly smaller or equal and other geq and so on who knows
results are up online. turns out i got it right on both questions
>feelsgoodman
Good job, user!
i stopped counting how many times i was close to killing myself after an exam that i actually did well in. i hate this shit
This. I usually did about 15% better than i anticipated, well aside from fucking language which is riddled with autistic exceptions. I got Cs in foreign language and it dropped my GPA from 3.8 to 3.5 i was we so god damn mad
>do well on all the exams i think i did shit on
>bomb the ones i think i did good on
Grats brainletbro
This still makes you a brainlet because your script said you were wrong by a factor of 4000 lmao.
always this
But i still dont know wtf is wrong with my script Veeky Forumsbros help..
N>= 40000 is the correct answer which is pretty much what i got. But my correct(?) script says n=10
Brainlet here. How do you show that this series converges to 2?
en.wikipedia.org
Set the upper limit to R and then the integral for lim R->infinity = 2
>while true
>if() break
*endless screeching*
>fucking cucked by my professors desu
No, you're just a brainlet, you couldn't even remember how to use Taylor expansion.
>No, you're just a brainlet, you couldn't even remember how to use Taylor expansion.
For a power series*
I know normal taylor series (one of the questions on the exam was about it)
And i got the question right anyway.
ok nvm maybe im a brainlet. But can someone tell me whats wrong with my script?
the script isn't wrong you dunce
desmos.com
Then where is the messup? This is the examiners motivation
what the fuck
The total sum is not two, it is around 2.61.
See wolframalpha.com
These results aren't incompatible. If the absolute error is less than .01 for N > 10, then it is also less than .01 for N > 4000.
It's important to recognize that the integral on the second line bounds the sum above, it isn't equal to the sum.
True, but the integral approximation actually yields an error that is pretty close to the real error, since shifting the domain of integration by 1 also gives a lower bound on error.
Shit.But why isnt it 2? I thought you could use the Maclaurin–Cauchy test en.wikipedia.org
The sum is continuous for all n in [1,infinity) and decreasing. So the generalized integral should give the sum. But why doesnt it?
fuck math desu this shit sucks
Think about why the integral test works.
>the sum is equal to the integral
ah yes...
Under those conditions, the integral test gives a lower bound for the sum.
Man im stupid. hopefully i get some points at least even though my math was fucked
Math is fucking gay
no u