The error probability for a class of binary recursive feedback strategies is evaluated. An exact analysis is given for both a nonsequential and a sequential decision strategy. We obtain the interesting result that even for the nonsequential receiver the error probability vanishes exponentially fast at channel capacity. A similar result had been previously obtained by Horstein for the sequential receiver, but was believed to be a consequence of the sequential nature of his decision strategy. For rates below capacity our feedback strategies have two error exponents, i.e., a lower error exponent <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^- (R)</tex> and an upper error exponent <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^+ (R)</tex> . The lower error exponent <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^- (R)</tex> exhibits an anomalous behavior in that <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^- (R)</tex> increases monotonically from <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^- (0) = 0</tex> to <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">E^- (C) = E(C)</tex> as the rate <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">R</tex> increases from 0 to capacity.