File talk:Repetition Code On AWGN Channel Graph.jpg

I've got a couple of issues with this page: 1. I don't understand the second graph, where the 'practical bpsk' line shows better performance than the theoretical (?best case) line. 2. One of the graphs has no axis labels, the other has BER and SNR; surely BER and Eb/No (with an explanation of whether or not that allows for the FEC) would be appropriate. 3. I've added some stuff about sharing bit probabilities when decoding so that it's obvious that using that approach would give better results. But I didn't add graphs/formulae to back it up. (It's self-evidently true, but I don't have a reference and am too lazy to derive it right now).