Gilbert-Elliot DFR is poor when compared to the repetition redundancy baseline

[vatch123]

The DFR for Gilbert Elliot also doesn't follow the results in the paper.

[vatch123]

The most recent bug hunting. See changes in commit 6fb346e fixes this. The performance is now comparable to the repetition redundancy

[vatch123]

Can't figure what is actually wrong that can cause this issue?

[vatch123]

On a side note, the bernouli results are also not at par with the paper but can't figure them out too