Another Look At Goal Differential

A couple of weeks ago Ryan looked at how improved the Oilers might be based on the improvements that have been made on defence and between the pipes. This morning we got an email in response to that post from a new member of the Copper & Blue community,  Gregarov, which has some interesting details regarding goal differential and is worth sharing with everyone.

I read your blog for a few weeks ago and was curious to look into the numbers a little bit more, so I put in the statistics into a spreadsheet for the past 3 years (extrapolating to an 82 game schedule for 2012-2013) and came up with some interesting results. Their were a few teams I didn't include in the stats for certain seasons as they only had won 0-1 more game than the previous year which would throw off the averages too much to include their numbers in my conclusions.

Mostly the average goal differential for winning additional games for teams that improved their records was 4.37 per game between 2011-2012 and 2012-2013 and 6.33 per game for 2012-2013 and 2013-2014. Conversely teams that lost more games than the previous year has average goal differentials of 3.60 per game for 2011-2012 and 2012-2013 and 4.67 per game for 2012-2013 to 2013-2014. Interestingly it seems you don't have to have make your goal difference much worse to lose more, and it is tougher to get more wins without improving that goal differential substantially.

Now it really gets interesting as teams that won or lost more than 8 games more than they did the previous year had quite different numbers. In 2012-2013 7 teams won more than 8 games more than they did in 2011-2012 and their average goal differential per win was 4.04 In 2013-2014 4 teams won more than 8 games more than they did in 2012-2013 and their average goal differential per win was 2.97 In 2012-2013 4 teams lost more than 8 games more than they did in 2011-2012 and their average goal differential per loss was 4.13 In 2013-2014 4 teams won more than 8 games more than they did in 2012-2013 and their average goal differential per loss was 3.88

So the overall conclusion is that if the Oilers are going to take a big leap forward at some point, chances are that they will probably not need anything near a +6 goal differential per win. If they are going to take incremental steps forwards over a number of years the number might be close +5 or more goal differential per win.

For all the teams that ended up with a goal differential of +30 or more the average was +5.5 goal differential per win. But that includes a lot of great teams that didn't win a lot more games but had a better goal differential like St Louis, Chicago, and San Jose. Perhaps the oddest situation was Tampa Bay in 2011-2012 having a -46 goal differential and a 38 win season, and the following year having a -3 goal differential (+43 better) but only had an extrapolated 31 wins.

This proves that bettering goal differential doesn't always have an impact on improvement in the win column although the vast majority of time it does. The statistics are not ideal as the interpolated 82 games from 2012-2013 will never be the same as a full season, but it gives an impression of what it takes to get extra wins season to season.

Disclaimer notice: I didn't double check the numbers I entered so I apologize if there are any errors within that may skew numbers.

If you want to look into the numbers in detail you can download into Gregarov's Excel file here.