Using Goal Differential to Predict the Future

Scott Reynolds
November 15 2012 04:26PM


Photo by Jeff Nelson via Wikimedia Commons

 At the start of each season, all teams are equal in the standings. A team that was among the league's worst a season ago is tied for first at the beginning of October December, and hope springs eternal for all but the most browbeaten fanbases. Of course, those particularly unfortunate fans remember something important that many fans like to forget: every team has a past, and that past matters. Sure, teams change in the off-season but there's a lot of continuity too, which means that, even when we confine ourselves to a simple thing like the previous season's goal differential, we can get a pretty decent idea of what can be reasonably expected in the year to come.

Since the NHL expanded to thirty teams for the 2000-01 season, there have been eleven complete seasons (or 330 team seasons). That gives us 300 pairs of seasons to look at if we want to use one season's goal differential to predict what might happen in the next one. When we look at those pairs, these are the results:

One of the things that I was impressed with is the symmetry: 74% of teams will have a change of less than 40 goal differential from the season before and teams are just as likely to crater (13%) as they are to flourish (13%). I don't know why, but I found that result a little bit surprising. I was expecting to see more complete tank-jobs than giant leaps, but that doesn't seem to be the case. Setting that aside, the amount of change is also impressive. Over half of the teams will see a change of 20 goal differential or more and we can expect about a quarter of the league's teams to have a shift of 40 or more. Half of that is in the wrong direction, but that still leaves three or four teams taking a leap forward each and every season, providing hope for next year to everyone else in the process.

Another thing we might be able to do with this data is find historical precedents for current teams. The Edmonton Oilers, for example, improved by 49 goal differential (-76 to -27) from 2010-11 to 2011-12. If we look at teams who made a similar jump in the past, what might that tell us about Edmonton's prospects for 2012-13?

The first step is identifying comparable teams. To do that, I looked for clubs who had one season with a goal differential of -50 or worse in Year One, and then saw that improve by 29 to 69 to end up somewhere between -37 and -17 in Year Two. I then looked at how those teams performed in the subsequent season (Year Three) to see if that might help give some indication of what to expect out of the 2012-13 Oilers. Here are the results:

I was surprised to find that there was no consistent pattern in this group. Usually when a team takes a big jump like this, most of us regard it as a step toward better things, and while that group of teams is over-represented in this sample, it's not exactly exclusive. I would suggest that, while there seems to be some basis for the conventional wisdom in this case, there's also reason for caution. Some teams that look like they're taking a big step forward aren't able to sustain that improvement, and others regress badly. Which will the Oilers be? God, I hope the lockout ends so that we get a chance to find out.

 

Previously by Scott Reynolds

 

889688bd2efc9140bec34eeba34eacde
Comments are closed for this article.