Intuitively we are supposed to achieve a coding gain with repetitive code for n>2 with a decrease in overall rate.
We are considering AWGN channel.
Surprisingly, repetitive code performs worse in AWGN channel (that is, no coding gain). Even a convolutional code with lesser minimum Hamming distance (dmin) achieves some coding gain. But repetitive code (n,1) does not achieve any coding gain.
Why is it so?
P.S: Performance measure: BER (bit-error-rate) vs SNR (signal-to-noise ratio) per bit (Eb/No).