How Many Rookies is Normal in the NHL?

Several days ago, Ender and I were discussing the merits of the 2007-08 Edmonton Oilers.  The conversation started around whether or not the offer sheet to Dustin Penner in the summer of 2007 was a good idea but we ended up touching on a wide variety of things throughout the conversation.  One of those things was my idea that the Oilers employed too many rookies that season.  They had seven different players who played 60 games that season who hadn't played 60 career NHL games when the year began.  That struck me as being a very large number and it was one of the reasons that I thought the Oilers were a bad team that year.  In this post I'm going to look at how many of these kinds of players have been employed by NHL teams since the lockout.  Is seven a large number, or are there actually lots of teams who play that many rookies all year long?  What's normal?  This is an attempt to begin to answer that question using some pretty specific criteria, namely, how many players did teams have in the lineup for 60 regular season games over one season who hadn't yet had 60 regular season games of NHL experience.  To make the post easier to write, I'll be referring to those players as rookies from here on in.

The results are in and I was correct that the Oilers of 2007-08 were anomalous.  No other team in the four years since the lockout employed seven or more rookies.  As you'd expect, there were a lot more rookies in the year immediately after the lockout and yet none of those teams employed as many rookies as the Oilers did in 2007-08.  In fact, only four teams (two in 2005-06, two in the three subsequent years) employed more than four.  The following chart gives us a pretty good idea of what's normal:

Excluding the year after the lockout, it looks lik most teams wil employ one or two rookies, though 0, 3 or 4 are not uncommon.  The Edmonton Oilers total of seven is massive.  The team in 2006-07 is the Pittsburgh Penguins with 6.  In 2005-06 the New York Rangers employed six rookies and the L.A. Kings employed five.  Now, two of those teams were pretty good and the other two, not so much.  And that's where the results surprised me.  With the criteria I've chosen there doesn't seem to much correlation between how many rookies a team has in the lineup and their regular season point totals:

You could maybe make a case that the teams with no rookies do substantially better.  In some ways this is probably a selection bias since teams that end the year with no rookies playing 60 games have probably been very healthy and I would imagine that good health is probably the bigger factor. You might even be able to argue that teams who employ only one rookie over the course of the season tend to do better but the difference really isn't very big and there are really good teams across the board.  It looks to me like "they're starting the year with a lot of rookies" is not a good way of showing a team is of poor quality.  Although the number of rookies may have a small impact on overall performance, it looks to me like, so long as the rest of the team is quality, you can break in rookies who are ready for the NHL and still win.  Indeed, breaking in rookies doesn't seem to cost much at all compared to other bottom-of-the-roster options.