Reorganizing again

The outline of the paper is changing again, but not the content. We're just going to present the topics in a slightly more cohesive manner given that the true subject of the paper has become the block variance, not the sum itself.

1. Introduction
2. Sampling framework and notation
3. Exact distribution of sample sum
1. Distribution as a function of the block variance
2. Block variance in the uniform case (my adivser would like to use a different word than "uniform" since it sort of implies a Uniform distribution - I'll think of something).
3. Block variance in the correlated relevance case
4. Block variance in the general correlated case
4. Methods for estimating the block sum
1. Simple sample variance
2. Bag of Little Bootstraps
3. Bootstrap Metropolis Hastings
4. Kernel estimation of relevance
5. General kernel estimation
6. Results and discussion
5. Implementation details
6. Future research