Here's one variable that, if included in the algorithm, would certainly cause a different result -- number of port days vs. at-sea days. For example, given that port days generate shorex revenue but at-sea days generate onboard revenue, what's likely the number of $ that a passenger will spend in cruise line bucks on a port intensive 14 day cruise vs. a 14 day transAtlantic? While those would be at the extremes, certainly there would, on average, be a difference?
I'm sure there are many more, and more subtle ones that have been teased out of the data over the years by all of the lines.