http://biwiki.editme.com/bench is the article being discussed.

I agree with Rip.  The purpose of benchmarking is to compare alternatives.  There is nothing that says the alternatives must be "similar".

Certainly it makes sense to compare the alternatives of investing in an S&P 500 index fund vs. buying large-cap stocks individually to see if your skill (or luck!) at actively managing your own portfolio is producing an improvement over passive (index) investing.  My spreadsheet is specifically designed to help with that sort of comparison (portfolio vs. index fund).

It can also make sense to compare alternatives such as a stock/bond mix vs. 100% stocks.  However, (if 100% stocks is represented by an index) then you're not only comparing active vs. passive you're also comparing different strategies.  That's going to make it hard(er) to figure out whether it's the active part or the strategy part that's making the difference.

You could also compare alternatives such as actively investing in small stocks vs. actively investing in large stocks or commodities.  Again, you're mixing both skill/luck and strategy in such a comparison which complicates the interpretation.  Also, my spreadsheet will be no help at all with doing that (comparing two actively managed portfolios).

And, of course, benchmarking can only tell you what would have happened in the past.  It says nothing about whether or not something similar will happen again in the future.