Goals by Expected Points
A FTF late night special.
I usually look at predictive stats because I'm mostly interested in trying to evaluate true talent. Every now and then I'll take some time to look back at the season through descriptive stats. The descriptive stats don't tell us what is likely to happen, but what has happened. Using descriptive stats allows us to analyze who has contributed the most to the Sharks successes and failures thus far. Recently I've worked with what I define as 'Expected Points.' A more detailed explanation can be found here. I'll save you from reading that by summarizing EP as the amount of points an average team would accumulate at a given point in a game.
I've mostly used EP to look at in-game changes in momentum, but tonight I wanted to take a peek into individual performance. I first grabbed all the goals that were scored for a player while he was on the ice, and then calculated the expected points gained from that goal. Not all goals are the same value, as a goal in the dying seconds of a game is a lot more valuable than 1 scored when your already ahead by 3. I did the same for all goals against. I then added these two values to see how many 'Expected Points' a player contributed to the team. This is really a watered down version as goals are not the only contributor (albeit probably the biggest ) to EP. For example, taking a penalty, or killing time at the end of the game while ahead are all important in calculating EP. I think in total this method accounts for about 60 points, the other 30 are mostly due to changes in time, and strength. Using goals gives us a rough estimate of total player contribution, although it may be slightly biased toward goal-scorers.
|Num||name||Pos||Home EP+||Home EP-||Home EP diff||Away EP+||Away EP-||Away EP diff||EP+||EP-||EP|
EP+ represents the sum of all EP gained from being on the ice for a goal, while EP- is the sum of all EP lost from being on the ice for a goal against. The last column is the overall EP differential.
As expected the big guns accumulate the most EP, none other than the big cat, JT. The interesting thing I found while compiling this data was the wild swings in EP. I think Pavs deserves a much higher rating but he's scored a lot of goals early in the game. For example a goal when the game is tied in the 1st nets the players on ice about 0.3EP. A goal in the dying minutes of the 3rd when down 1 nets the players on the ice about 1.5EP. Thus a small set of games can really determine EP. Also, the guys that are leaned on for eating time at the end of the game aren't getting any credit for that, only suffering from goals against. So it's a little offensively biased but there's also some value in that. We can see who really potted the clutch goals.