Let $X_1, X_2, \ldots $ be a sequence of independent identically distributed random variables with values in the space $C([0,1])$ of real continuous functions on $[0,1$]. Assume for simplicity that the $X_i$'s are centered, in the sense that $E( X_i(t)) = 0$ for $t \in [0,1]$). Are there any necessary and sufficient conditions (on the common distribution of the $X_i$'s) in order that $$ \limsup_{n \to \infty} \frac {1}{\sqrt {2n \ln \ln (n)}} \, \| X_1 + \cdots +X_n \|_\infty < \infty $$ almost surely (where $\| \cdot \|_\infty$ is the sup norm)? What would then be the value of the limsup?


What you want is the so-called bounded law of the iterated logarithm (BLIL) in $C[0,1]$.

A necessary and sufficient condition for the BLIL in an arbitrary real separable Banach space $B$ was given by Ledoux and Talagrand '88 -- see e.g. Theorem A in the subsequent paper by Einmahl '93. This necessary and sufficient condition involves the boundedness of $(S_n/\sqrt{n\ln\ln n})_{n\ge3}$ in probability, which may be hard to verify in $C[0,1]$.

A sufficient condition (involving a certain, rather weak condition on the modulus of continuity of $X_1(\cdot)$) for the BLIL in $C[0,1]$ is implied by Theorem 2.1 in the earlier paper by Kuelbs '76.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.