[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math


View post   

File: 18 KB, 460x276, einstein.jpg [View same] [iqdb] [saucenao] [google]
2983471 No.2983471 [Reply] [Original]

hey /sci/

Physifag here.
We're starting to study the Brownian motion and for the "historical background" we're supposed to do a small assignments, which is... statistics... which i loathe and suck at.

Now, the assignment is nothing more than a random walk, however, with inequal chances of going left or right.

We're supposed to calculate the "supposed position" after n steps in terms of n and p; and the variance. Since the choices are "left" or "right", i think this is some kind of binom...

I'm pretty sure the position is just the mean, which in this case would be:
n*p - n*q (seeing going left as positive, going right as negative).
= n(p-q) = n(p-(1-p)) = n(2*p-1)

However, the same trick obviously doesn't work for the variance:
n*p*q - n*q*p = 0

So, erm, what's the variance?

>> No.2983520
File: 1.93 MB, 1920x1040, cutey_Emma_av.png [View same] [iqdb] [saucenao] [google]
2983520

how about looking online first? (Asshole)

like like

http://mathworld.wolfram.com/RandomWalk1-Dimensional.html

http://en.wikipedia.org/wiki/Random_walk#One-dimensional_random_walk

>> No.2983580

>>2983520
Well, my friendly /sci/entist, i already did.

Unfortunately, if you had done so yourself you'd have noticed wikipedia only talks about random walks with p = q = 50%.

Wolfram thing is constantly only talking about the mean amounts of steps left or right, not about taking both into account, so i'm unsure if i can just use the variance n*p*q. If i can, then please tell me so, if not, it's not on wolfram.

>> No.2983597

Fucking binomial distribution, how does it work?

>> No.2983614

fucking quantum physics
i didn't go into physics to do probabilities

>> No.2983622

>>2983597
EX = n*p
Var X = n*p*(1-p)

But if you hadn't noticed, i need the location, not either the steps left or the steps right. The EX i need (and which i calculated) is
n*p - n*q = n*(2p-1)

And now i'm not sure if i also have to touch the variance of the binom...

>> No.2983632

>>2983614
my thoughts exactly

>> No.2983636

>>2983614
What's the connection between quantum physics and statistics? You obviously are thinking about statistical physics.

>> No.2983650
File: 81 KB, 480x390, hypnotoadmb8[1].gif [View same] [iqdb] [saucenao] [google]
2983650

>>2983622
Yes, you haALL GLORY TO THE HYPNOTOAD

>> No.2983657

>>2983650
Well, erm, thanks for trying, i guess. lol.

>> No.2983668

>>2983657
Seriously though, variation is npq.

>> No.2983882

>>2983668
>>2983668
Actually, wouldn't it be 2npq?

You could say that the calculation is like
D(np, npq) - D(nq, npq) = D(np-nq, (npq)^2 + (-npq)^2) = D(n(p-q), 2npq) ?

>> No.2983939

>>2983882
What exactly does D stand for?

>> No.2983989

>>2983668
>>2983882
Upon further consideration I agree I fucked up and didn't take into account that when you build analogies with binomial distribution, total step is actually 2, not 1. So, variance is 4pq.

>> No.2983993

>>2983989
Ugh, 4npq