Random commentary about C++, STL, Boost, Perl, Python, Algorithms, Problem Solving and Web Search
Everytime after you toss the coin you will interchange the meanings of H and T (heads becomes tails and viceversa), thus the bias will not be only on T. If you have another solution, please do tell :)
Von Neumann solution fits well? Toss the coin in pairs and reject the similar outcomes. There by we can rule out the biased outcomes as the probability of HT and TH is same.