What I think people who aren't in the software development field may not realize is just how much can go into such an algorithm. At most what's "known" publicly are the general parameters, such as, say, average age of accounts, average utilitization, maximum utiltization, etc. But there's a lot of detail necessary to implement each of these, and the results can vary tremendously based on these details. Just some of these details could include:

• Are all three reporting services equally weighed, or is there a different weight for TransUnion vs Experian vs Equifax?
  
• If information is found in multiple reports and differs, how do they resolve that?
  
• If information is found in multiple reports and is the same, do they count it twice or once?
  
• If different accounts have reported at different times in a particular report (when I get a report, it's timely for some accounts, but out of date for other accounts), how/do they compensate for that?
  
• Do they use just one report from each agency, or do they get multiple reports over some period of time?
  
• Do they have special knowledge about some accounts (such as the fact that Amex reports "hi balance" in the same field as a CC reports "hi limit") or not?
  
• Exactly how far back do they go in history? Does it vary for different types of accounts?
  
• Do they ignore certain account issuers because of a history of unreliability of the data from such issuers?
  
• Do they weigh certain account issuers higher than others?
  
• Do they weigh certain types of accounts higher than others?

Since it's a computer program that's implementing the algorithm (and likely a full-time staff maintaining that algorithm), there's no reason for the algorithm to be simple/short enough to explain in a newspaper/magazine story. In fact, there's not even a reason for the algorithm to be simple enough to explain in "plain English". I wouldn't be surprised if it requires a complex many-page flowchart to explain it.