Proposition
In , the maximum number of vectors such that for all is .
\begin{proof}
Suppose vectors in satisfy the condition. We show any vectors are linearly independent.
Otherwise, assume that for . Separate coefficients into positive () and negative () indices:
Then . Since and , each term is . Thus .
Notice that . Since all , this implies all .
Thus, any vectors are linearly independent, so , which means .
\end{proof}
Lemma
In , the maximum number of vectors such that and for all and is .
\begin{proof}
Since there does not exist anti-parallel pairs, for such , we can get such that .
Then by ^gwjb1o, .
\end{proof}
Proposition
In , the maximum number of vectors such that for all is .
\begin{proof}
By Induction on .
In , if , we can have at most one positive and one negative number (e.g., ), which deduces that .
Now assume the bound holds for .
Let be a set in satisfying the condition.
If no two vectors are anti-parallel , the bound is smaller by ^70d15f.
If there exists a pair of opposite vectors and , then for any other vector , we must have .
This forces all remaining vectors to lie in the -dimensional subspace orthogonal to . By the inductive hypothesis, there are at most such vectors. Then total .
\end{proof}