Didn't he already show you a wc -l? Did you test it yourself and get different results?

No. He only added the wc after I asked how many lines were in the file. And I asked because the only way I could reproduce anything like his results was if I ran tac on a large file that contained few or no newlines. It is faster, but not by the margins his test showed.

Then again, I'm using a win32 native port of tac and I do not know how close it is to the *nix version, or if it uses the same algorithm.