I really think that you need to use the length of the longest citation cycles to get a good idea of where the crackpot-clusters are.

The assumption being that the good journals outnumber the bad journals, so even if crackpots can take over one or two, the longest cycle they can get is 2, whereas over time all good journals will tend to link to most other good journals.

And it's there that the El Naschie papers fail. Most of the sections quoted by that article fail to adhere to basic standards like "explain the appearance of all magic numbers":

He then suggests quitting at the second stage of this iteration and getting
"2×2×17=68"
of something -- but it's not clear what, nor why the number 2×2×17 should show up.
But never mind! He then notes that 68 is
"1/2(1371), where again 137 is a rough approximation to the reciprocal of the fine structure constant. "
Of course, can always find some formula linking any two numbers, and the possible meaning of this formula linking the numbers 137 and 68 is not discussed.