As of Saturday night, all teams in the league have played at least half of their regular season games. I felt this would be an appropriate time to update the All-Strength Team PDO analysis that I have been doing throughout the year.
For starters, let us look at the raw numbers for the league through Saturday night's games:
|Team Name||All Strength
After climbing out of the hole earlier in the season, peaking at around 1.02 (a little better than mediocre), the Washington Capitals PDO has now fallen back to the realm of solidly mediocre. In terms of their standings points, they are still at a deficit of approximately 2-3 wins from where they should be with their PDO. This is the same deficit they had after 20 games into the season. The projected standings points for the end of the season is in the mid-90s which also does not come as much of a surprise.
As mentioned in my last installment, with the exception of last year's shortened season and the season in which everything went right (except the playoffs) of 09-10, the Washington Capitals have had a Team PDO ranging from 0.990 to 1.005. Therefore, their current PDO is not much different from most of their recent seasons and it appears the magic of AO77's first partial season may have vanished.
How are the Team PDO values for the league trending? The following plots show the running Team PDO through the first half of the season. Any PDO that falls outside the 2SD lines implies that the team is doing something that is significantly better (or worse) than what would be a variation due principally by chance.
First off, the Metropolitan Division:
As has already been mentioned, the Caps are within the "meh" part of the plot. In fact, most of the division is. Only the Pittsburgh Penguins show a significant positive trend and the New York Rangers and New York Islanders show a significant negative trend.
Next is a quick look at the Caps Conference rivals, the Atlantic Division:
This plot shows the Boston Bruins reaching into the exceptional zone (and the Florida Panthers doing likewise in the exceptionally bad zone). The Toronto Maple Leafs are still doing pretty well as their PDO has fallen back toward the center from their unsustainable start. The Montreal Canadiens and Tampa Bay Lightning have managed to maintain their above average PDOs for the whole season. So the Atlantic seems to have 4 decent contenders for the last half of the season.
Next, we move farther West and take a look at the other two divisions. First up, the Central Division:
The Central seems to have a few interesting stories. The St. Louis Blues have managed to maintain their PDO for the first half of the season and have fought against the pressure of regression to the mean. This has left them well above the 2 SD range at the halfway mark. The Colorado Avalanche has managed to stay high but has followed the SD lines while slowly regressing toward the mean. Meanwhile, the Chicago Blackhawks started out average and have been climbing steadily as the season progresses. While their numbers are not out of the ordinary yet, the trend seems to imply that by season's end, they will have proven that they will be contenders in the post season yet again. The rest of the division is below average but no one is atrocious.
Finally, we take a look at the Pacific Division:
This division is the western equivalent of the Metropolitan division. None of the teams seem to stand out. The Edmonton Oilers and Calgary Flames are struggling to stay mediocre, but otherwise, the division is mediocre.
As in the past, we now take a look at how the teams are fairing relative to the average Shot and Save %. The red lines marking the center point of this plot represent the season average shooting and save percentages of the current season (which is a bit lower shooting and high save than the average of previous seasons).
The circles in this plot represent the 1, 2 and 3 Standard Deviations of the PDO of previous seasons.
Interesting to see the St. Louis Blues and Buffalo Sabres have a fair amount of regression toward the mean in order to fall within the season ending 3 SDs of the average. From a Washington Capitals fan's perspective, it is disturbing to see that we are not all that different from the Columbus Blue Jackets or the New Jersey Devils. However, it could be promising to see that we are also not that much different from the Canadiens or the Lightning. We have a ways to go to catch up to either the Bruins or the Penguins, unfortunately.
As was pointed out in the comments of my last post, the ovals in the above plot lose the significance of the more useful information of the save and shoot percentages. In order to make a plot showing the standard deviations of these two data points, I needed to calculate those standard deviations. After combining the season ending, all-strength team shooting and save percentages since the 05-06 season (a total of 270 data points of each), I made the following plots to ensure the distributions were actually normal.
At first, one would think that this conclusively identifies the distributions as normal. However, the samples were broken into only 8 bins. Sturges' Rule implies that the data set should be broken into 9 and Rice's Rule indicates that the data should be broken into 13 bins. If we take a compromise between the two and break the data into 11 bins, we get the following plots:
These plots seem to indicate that there is some sort of bimodal or multi-modal distribution occuring. The explanation (if there is one) is not immediately forthcoming to me. Regardless, I think I will ignore that aspect of the data for the moment and assume for the following plot that the distributions are normal. Further investigation into what would cause a multi-modal distribution will have to wait for another FanPost.
Using the 270 data points for shot and save %, the Average Shot % is 0.009309 with a Standard Deviation of 0.009573. The Average Save % is 0.90691 with a Standard Deviation of 0.010082. Plotting standard deviations onto the Save vs Shot plot shown earlier while using this season's average shot and save % as the center point, we get the following:
This shows that although the Buffalo Sabres PDO was out of the range for the season, it does show that their shooting and save % are still within the 3 standard deviations of the average. However, the St. Louis Blues are still showing a shooting % that is outside the expected. I think we can safely assume that they are due for a drop in shot % in the near future.
I welcome others thoughts and suggestions as I intend to carry on with tracking team PDO for the remainder of the season.