The argument that was being made was that nowhere in the article does it say that the record that was broken was the average-attendence record. If you've ever taken a statistics course or a public relations course, you'd understand that you can take any set of data and spin it however you like based on the wording of the article it is presented in.
In this case, it is said that the NHL set a new all-time attendence record for the month of October. Then the writer of the article says that the average attendence of the league is 16,820. The reader is automatically going to associate the two pieces of information and assume that the league set a new all-time average-attendence rating. Sure the article could have been poorly written. But this article was written by the NHL. You'd think they'd specify. What I believe happened was that the league set a new total-attendence record, aided by unprecedented number of games played in October by the league, and then fed us information about how the league's average-attendence rating was above that of the last four years.
So you're saying that the league never had a better average-attendence rating than 2002's 16,226 before this year? That I'm not so sure about. It had to have beat that mark in the 90's. But of course that wouldn't result in a record total-attendence rating because there wasn't 30 teams 10 years ago.
Or the article was just written poorly.