# On average length of sums of unit vectors in R^n

Fix a number m and let us take a set, say A, of unit vectors {v_1,...,v_k} in R^n. Assume that k is large, say exponentially large in n (k=e^{cn}). Let X be the euclidean length of a random sum of m vectors in A (all sums are equally likely).

It is quite intuitive that the typical length of such sum should be sqrt(m) and it is not that hard to verify it.

Question: How does the variance of X behave? More precisely, is it true that VarX=o(sqrt(m)). If so, could one get a better rate? Maybe it is m^1/4?

I would be very grateful for any information or link.

-
This seems strange; your unit vectors, are these (1,0,...) and (0,1,0,0,...) and so on, or just vectors of length 1? If the latter, the expected mean should be the zero vector then, no? – Per Alexandersson Feb 17 '13 at 14:23
Is not the question asking for statistics for the total travel of a random walk, each of whose steps is of length $1$? – Joseph O'Rourke Feb 17 '13 at 14:58
It sounds like once you figure out what question you want to ask, the answer will be the Central Limit Theorem (for vector-valued summands, or for each coordinate if you don't mind losing a factor). If you mean something deeper than that, please clarify. – Douglas Zare Feb 17 '13 at 19:58
Per Alexandersson: there are n vectors with 1 as a coordinate and we have exponentially many, so I clearly mean any vector of legth 1. Joseph O'Rourke: it is - it is the about the average length of a random sum x_1+...+x_m, that is, it's euclidean length. Douglas Zare: I think I have really meant what I have written. Anyhow, how should I use the CLT in this case? I really need just a bound on the variance of the length of such a sum, nothing more. – TOM Feb 18 '13 at 3:06
TOM, your reply to Per adds to my confusion. You refer to a finite set of vectors (all 0 except for a single 1) then you say "any vector of length 1" (of which there are infinitely many). You put your vectors in $R^n$, not $Z^n$, remember. What is the complete definition of $A$? – Brendan McKay Feb 19 '13 at 1:38

I am assuming that the OP means that the initial $k$ vectors are themselves a sample of the uniform distribution on the sphere, so the question is both about the limiting distribution of the length (which is normal) and the speed of convergence to this distribution. This is a much studied (and highly nontrivial) question: for a nice analysis of the two dimensional case (and extensive references), see: