Which bootstrap?

There is one small hitch in my plan to use the bootstrap algorithm in my comparo: which one to use? The term is pretty broad, basically encompassing any technique that subsamples a sample to deduce the quality of an estimator from that population. The name comes from the phrase, "pulling yourself up by your bootstraps." You don't know anything about the population distribution, so you repeatedly subsample your sample to simulate the distribution.

The Bag of Little Bootstraps (BLB) algorithm that I'm officially comparing seems to favor the percentile bootstrap, though they admit that other techniques may be more appropriate in various situations. Because my data is highly correlated within physical blocks, I'm leaning towards using the Block Bootstrap. This also plays into my desire to keep all my subsamples within a physical block of data.

It also makes for an interesting matchup between the BLB and the Bootstrap Metropolis-Hastings (BMH) algorithm. If I use blocks as my subsamples for both, it will be a straight-up fight to see which one can do a better job of estimating the true blocksum variance.