Consider a continuum of random variables such that each takes the value $1$ with probability $p$ and $0$ with probability $1-p$. The random variables should be essentially pairwise independent.

Sun and Zhang (2009, Journal of Economic Theory, 144, p. 432-443, doi:10.1016/j.jet.2008.05.001, Theorem 1 and Corollary 2) show that there exists a $\Omega$, a suitable probability space on $\Omega$, an extension $\bar \lambda$ of the Lebesgue measure on $[0,1]$, a Fubini extension on $[0,1]\times\Omega$ and a corresponding process $g\colon [0,1]\times \Omega \mapsto \mathbf R$, such that the random variables $g(t, \cdot)$ are essentially pairwise independent with the required distribution and a law of large numbers is valid. I hope this correctly summarizes their findings, but I am grateful for any corrections. Now my question is how to formulate a law of large numbers here. I don't want to use their formulation as it is too general for me and it requires even more notation.

Is the statement "$\int_{[0,1]} g(t,\omega) \mathrm{d}\bar \lambda=\bar \lambda(\{t\in [0,1]|g(t,\omega)=1\})=p\ \ $ almost surely" correct?


Your Answer

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

Browse other questions tagged or ask your own question.