I once (a year or two ago) tried to determine the distribution of all backs relative to a 520 card set. (I was going to try to make my 520 mirror that distribution)
I used Bill Browns superset spreadsheet, Scot Readers data, the back multipliers from SCD and t206.org., and the list of toughest backs that were the subject of more than one thread.
I'm sure my method had faults, but regardless the target number I arrived at was 8/520 SC150-25's
Take that for what it's worth
