http://cache.gawker.com/assets/images/9/2009/09/500x_dflag.jpg
Interesting article on the inherent delays for games processing inputs, in both 30fps and 60fps.
They used a display with 4 60fps frames (67ms) lag. Games showing “67ms” on that chart (including SF4 and guitar hero, no surprise there) are lagless.
Anything else is inexcusable.
Its a good thing to inform the masses about input lag problems, but this test seems quite funny to me. They went all the distance to get specific mod by Benheck to have the most stylish input gear, but then tested on a laggy LCD. They could have easily used a stick or padhack with a single LED-modded button and could use a CRT screen to eliminate display lag.
Furthermore, testing for visual clues of a characters doing something just isn’t right. How do they know when exactly is Master Chief supposed to jump after a button press is detected? 
They could just stick to a fighting game, training mode and “show input” option to do the job. Oh and a CRT.