Mann's original Hockey Stick graph was flawed. When this was discovered through the peer review process, Mann and his group redrew the graph using better (more) data. The new graph was still shaped like a hockey stick and the revised study (now quite old) has never been "de-bunked" by the critics of the original study.
The warming skeptic crowd always points out the problems with the first Hockey Stick and ignores the fact that the errors were corrected and accepted long ago.