In this FanPost, I used mathematical trend analyses of goalscoring for all of the forwards to predict their performance for the 2014-2015 season. Then, I looked at all of the defensemen, and computed how much the Devils can expect to score. Finally, I looked at goaltending, and I used this information to predict how the Devils, as a team, will do next season. All of my information came from Hockey Reference.
For those of you who just wish to know my results...
-Henrique is the leading scoring with 29, leading the Devils to 202 GF. Zidlicky remains the top offensive defensemen, contributing 11 goals.
-Lead by Cory's stellar goaltending, the Devils give up a measly 178 GA.
-This leads to an elite differential of +24 Goals.
My study is below. Thanks for reading up to here and beyond!
For determining goal contribution for the 2014-2015 season, I looked at the G/GP of each player over their career. I used this data and graphed a line of best fit to predict where we can expect the Devils' players to be for next season. The example of Jagr is below. Click on the picture to zoom in.
Graphing his G/GP over his career, I used a 2nd degree polynomial line of best fit to determine where we could expect him to be next season. His 2014-2015 expected G/GP is 0.175. I also looked at his average GP over his career, 73.65, and I used this number and other information to predict his GP for 2014-2015: 80 GP. I repeated this process for 16 of our forwards. The data from this analysis follows.
Forwards
Player | GP (ave.) | Formula for G/GP with variable year | Year of Q | Exp. G/GP 2014 | Exp. GP 2014 | Exp. G 2014 | |
1 | Henrique | 64.33 | y = 100.6ln(x) - 764.97 | 2014 | 0.367 | 80 | 29 |
2 | Ryder | 75.9 | y = 0.0003x^2 - 1.0298x + 1038.2 | 2014 | 0.291 | 78 | 23 |
3 | Cammalleri | 60.82 | y = -0.0034x^2 + 13.695x - 13755 | 2014 | 0.286 | 63 | 18 |
4 | Elias | 67.88 | Y=-.0025X^2+9.8413X-9862.9 | 2014 | 0.23 | 76 | 17 |
5 | Zajac | 68.88 | y = -6E-05x^2 + 0.2341x - 229.88 | 2014 | 0.175 | 80 | 14 |
6 | Brunner | 60 | 2014 | 0.22 | 62 | 14 | |
7 | Jagr | 73.65 | y = -0.0016x2 + 6.5521x - 6546.9 | 2014 | 0.175 | 80 | 14 |
8 | Rutuu | 59.83 | y = -0.0023x^2 + 9.0974x - 9130.1 | 2014 | 0.138 | 59 | 8 |
9 | Zubrus | 68.76 | y = -0.0019x^2 + 7.6941x - 7712.4 | 2014 | 0.079 | 71 | 6 |
10 | Havlat | 57.5 | y = -0.0037x^2 + 14.725x - 14760 | 2014 | 0.086 | 59 | 5 |
11 | Bernier | 60.22 | y = 0.0043x^2 - 17.122x + 17233 | 2014 | 0.081 | 62 | 5 |
12 | Gionta | 66 | 2014 | 0.07 | 68 | 5 | |
13 | Boucher | 23 | 2014 | 0.09 | 41 | 4 | |
14 | Clowe | 57.5 | y = -0.0058x^2 + 23.425x - 23517 | 2014 | 0.051 | 57 | 3 |
15 | Josefson | 29.5 | y = 0.0125x^2 - 50.309x + 50619 | 2014 | 0.073 | 31 | 2 |
16 | Matteau | 17 | 2014 | 0.06 | 17 | 1 | |
2014 | 2.472 | 984 | 168 |
Notes:
-Some players polynomial lines of best fit did not make sense, so I used other functions like 'log'.
-The predicted GP did not reach the expected (12*82=984) so I made the following GP additions: 2 to each player, Zajac: 5, Henrique: 3, Elias: 4, Boucher: 15, Jagr: 3.
-I removed the first year of data from the following players: Henrique, Elias, Clowe
-Some players had such small sample sizes, I had to use other methods to determine their prediction.
-I assumed 12 Forwards played a game.
Defense
Player | GP (ave.) | Formula for G/GP with variable year | Year of Q | Exp. G/GP 2014 | Exp. GP 2014 | Exp. G 2014 | Remove first Year? | |
1 | Zidlicky | 69.9 | y = 0.0022x2 - 8.9489x + 8990.3 | 2014 | 0.139 | 80 | 11 | |
4 | Gelinas | 60 | 2014 | 0.12 | 72 | 9 | Only 1 year data | |
2 | Greene | 59.7 | y = 4E-93e0.1044x | 2014 | 0.078 | 82 | 6 | |
5 | Larsson | 42.67 | y = 10.054ln(x) - 76.458 | 2014 | 0.033 | 80 | 3 | |
6 | Merrill | 52 | 2014 | 0.04 | 80 | 3 | One year | |
3 | salvador | 64.25 | y = 2E-05x2 - 0.0649x + 67.727 | 2014 | 0.012 | 67 | 1 | |
7 | Harrold | 28.88 | y = -2.627ln(x) + 20.016 | 2014 | 0.033 | 31 | 1 |
Notes:
-Additional GP: Zidlicky: 10, Greene: 12, Salvador: 3, Gelinas: 12, Larsson: 37, Merrill: 28, Harrold: 2.
-Some players had such small sample sizes, I had to use other methods to determine their prediction.
-I assumed 6 defensemen played a game.
This process determined that the Devils should score a pedestrian 202 goals next season. This is 4 more than last year.
Now, I used a similar process to analyze our goaltending.
Goaltending
Player | GP (ave.) | Formula for SV% with variable year | Year of Q | Exp. SV% 2014 | Exp. GP | Exp. SA/G | Ex. GA | |
1 | Schneider | 27 | y = 2.0139ln(x) - 14.393 | 2014 | 0.926 | 64 | 26 | 123 |
2 | Clemmensen | 20.2 | y = -0.0013x2 + 5.373x - 5398.9 | 2014 | 0.881 | 9 | 26 | 28 |
3 | Kinkaid | 1 | 2014 | 0.885 | 9 | 26 | 27 | |
82 | 178 |
Notes:
-I guessed GP for all three goaltenders.
-I guessed a SV% for Kinkaid of 88.5%
-Devils' average SA/G since 2009 is 24.13. I used 26 SA/G (0.5 more than last season) to be reasonable.
Findings and Disclaimer
This method shows that we should give up an minuscule 178 GA next season. This, combined with 202 GF leaves the Devils with an elite goal differential of +24. This is a fantastic result! I think the most interesting part of this is the impact that having elite goaltending (again) will have on the Devils.
I do not believe that this process is entirely mathematically sound. The idea of a line or curve of best fit is a good idea, but there is no way that it is a perfect predictor for goal scoring (this problem would be compounded more the further into the future one would try to predict). I also did not weight averages based on how many games were played by a player in a single season. Furthermore, TOI was not a factor in my calculations. All of this being said, I think that in principle, my analysis is applicable. At the least, it can be a cause for conversation.
Your Take
Obviously, Jagr being significantly out-performed by Ryder is concerning (and there are other problems unmentioned and mentioned in the previous section), but I think that overall this is a good way to analyze goal-scoring trends.
What do you guys think of this analysis? Are these predictions way off? Can Cory perform at that level? Which players will break away from their trends? Leave your comments below, and thank you so much for reading.
-Brogamesh