Three Types of Lies

Today I heard the following funny remark from one of my labmates. He said:

 

There are three types of lies: lies, damned lies, and simulation results!

 

The above saying is not totally original though. It rephrases the famous quote "There are three types of lies: lies, damned lies, and statistics," which was popularized by Mark Twain. But it’s so true about research, well, at least the research in EE and CS, that I like the wit from the words immdiately after hearing it, and can’t agree with it more. For a guy like me who conducts theoretical research all these years, research is nothing but the following cycle: 1) reading the related work in the literature, 2) finding a hole in these works to patch, 3) dervie a (not necessarily better) solution, 4) hurrily run some (selected) simulations to arguably show that the new solution is indeed better, and 5) eagerly looking for a venue to get the "better" results published. The cycle keeps going year in and year out.

 

I am not totally disputing this type of research approach. In fact, I am happily practising it right now and everybody else is doing it as well. But that doesn’t mean most of us have no idea how realistic our research assumptions are, and how true the stories in our papers tell. Unfortunately, that’s the way how one can rack up his publications and that’s the way life is. So, like what’s said in "The Shawshank Redemption", perhaps one day we will all get instituitionalized. Eventually, we will become the real Ph.D. — Pretty Huge Dumb, and having yet another Ph.D. — Permanent Health Damage.