Back in 1998 NM Todd Bardwick did an article in Colorado chess (the US Championship was being held n Denver) and he wrote the following:
"One of my curiosities was what is the relative strength of a grandmaster compared with a master, like myself. David Gliksman once told me that a 2200 master has a basic and underline basic understanding of what is going on. I can’t dispute that. On Saturday, November 16th, I got the honor of testing my abilities against the U.S. Champ (GM Joel Benjamin) in some blitz games. Basically, I was terminated. In all fairness, I psyched myself out after a game or two which didn't help at all. Joel sliced and diced me like I was a beginner. It was amazing to me the harmony in which his pieces danced around the board."
What I found interesting was GM Joel Benjamin's comments:
"After our games, we went to the bar for a pizza and I asked him about relative strengths of masters. He said that a 1700 rated player would probably have slightly better chances against a 2200 player than a 2200 player playing against a 2700. His reasoning was that a 2200 player can play erratically and his playing strength can fluctuate more than that of a GM. Then, I asked him how he would do against Kasparov. Joel responded that if he could get an opening edge against a GM of his strength, he would most likely win. However, in the same scenario, Kasparov would most likely draw or even beat him. Joel said that Kasparov would probably have to give him the white pieces, and draw odds for Joel to have a decent shot. For Joel to win a game against Kasparov, he said would be difficult."
Determining the Strength of Chess Players Based on Actual Play
The Relative Playing Strength of Chess Players: A Note