This is one of those things about experience and a little bit of training (from someone who knows - rather than thinks they know). I had the case where a fellow was going to evaluate a GC/MS system and one of his criteria was that he wanted to see match scores above 900 (out of 1000) to show that the instrument could do the job. He sent a sample of an essential oil for analysis. I had to laugh as I reviewed the peak table. One peak showed three different matching compounds, each with a match score of over 900 - I think even over 960. I guess that by this fellow's definition, the one chromatographic peak was correctly matched to each of three different compounds - and being that the essential oil was largely terpenes and terpene related compunds, I would not bet you more than lunch that the compound was any of the three listed.
One of my favorite examples for teaching is comparison of the spectra of peach lactone and coconut lactone. They each have one huge ion at m/z 85 and not much else in the spectrum. The trick question is how do you tell the differnce between these compounds. The experienced chemist answers: "Retention index of the peak!" So much for that fancy expensive mass spec and that neat and nifty software.
While inability to understand the value of mass spectral library search does not often result in something fatal, it does have an impact on careers. As I was teaching class one day, I guessed that I had some novices to GC/MS in the class. So, I took a bit of a detour to discuss the use of library hits and the importance of confirmation of ID's with actual compounds. My end to the story was to tell the class that when that manager comes down to the lab looking for something to say at a meeting, DO NOT give a tentative identification based on just a mass spectral library hit. Because two things will happen. 1) The word tetative will be forgotten and 2) you will be given full credit. There is nothing like having that VP who is never seen in the building walk past the next morning and greet you with "So, you're so and so...."
In that particular class there was a fellow standing at the back of the group where we were looking at the computer monitor, and I can still see him standing by the doorway into the next room. He turned absolutely white. For a moment I thought I was going to make a 911 call. I can only guess that he'd made a tentative identification - that became a career move and I brought up a painful memory.
And, rather than ramble on - I'll just say that the use of library searches to ID things works quite well. As long as you use your brain. It's not hard. I've not gotten into too much trouble over the years.
