In last week’s post, the amount of production expected for a top four defencemen was discussed. While doing so, defensive point shares were brought up. Since point shares have been mentioned other times in my posts, I thought it was important to test their validity instead of treat it like a valuable stat everyone should accept. Granted, hockey sabremetrics have not hit a certain mainstream since 2008 and baseball has been able to parade their WAR and many other numbers much earlier, but come on! How has point shares not hit the mainstream?
For one, it could be as a result of too many constants and thus, positional stereotypes that lead to possible defensive point shares given to defencemen instead of forwards. After reviewing hockey-reference.com’s explanation on point shares last week, the positional adjustments actually make sense. If you add three forwards with positional adjustments of 5/7, two defencemen with positional adjustments of 10/7, you get 5: the total number of skaters that are on an ice at the same time. However, since defencemen play 20-25 minutes a game while forwards play 10-20 minutes a game, shouldn’t the positional adjustment be a little more even amongst those types of players. Yes, but no could also be an answer if you take into account that 6 defencemen and 12 forwards are played every night.
With one constant out of the way, how about that 7/12 constant when we need to calculate marginal goals? Let’s have hockey-reference.com explain this from their point shares explanation page.
Why 7/12? At even strengh a team has six players on the ice, five skaters and one goalie. Imagine each of these players having two chips to contribute to one of two buckets: offense and defense. Collectively the skaters will contribute five chips to the offensive bucket and five chips to the defensive bucket. However, the goalie will contribute both of his chips to the defensive bucket, giving the defensive bucket seven of the twelve chips.
Since that is straightened out, let us analyze offensive and goaltending point shares. We will use Washington’s finest to calculate and review these statistical categories: Alex Ovechkin and Braden Holtby.
Offensive Point Shares
First, we calculate the number of goals Alex Ovechkin has created. The objective for goals created is to combine goals and assists in order to make one common sabermetric stat for offensive productivity. Add all the individual goals created from each team and one should be able to get the total goals that particular team has scored. The current leaderboard in goals created shows that in order to do well, one must have a balanced number of goals and assists and not skew towards one or the other.
- (27 goals + 19 assists/2)*(126 goals/(126 goals + 220 assists/2)) = 19.5 goals created
Before we continue our next step, we will also calculate the league-wide goals created among all forwards. This will be used to calculate marginal goals for Ovie. Kids, don’t try this at home as this too way too much messing around on the hockey-reference.com season finder.
- (2,693 goals + 3,960 assists/2)*(3,378 goals/(3,378 goals + 5,689 assists/2)) = 2,536.8 goals created
Next, we use that league-wide goals created among forwards to calculate the marginal goals for Ovechkin. This takes into account how many goals are created that are above the league average.
- (19.5 goals created) – (7/12) * (53,006 seconds played for Ovechkin) * (2,536.8 goals created by all NHL forwards)/(#13,412,647 seconds played for all NHL forwards) = 13.64 marginal goals
- # = a rough guesstimate. If you thought figuring out league-wide goals created among forwards only was a giant pain in the ass, try going one-by-one and adding every second an NHL forward has played. You really want me to add all the seconds of all 537 NHL forwards that have played this season AND keep every other stat needed for these point shares calculations up to date? No chance!!!
- Thus, here is how I came up with my estimation: (3,378 league-wide goals during regulation and overtime)/((2.72 goals/game)*(60 minutes/NHL game))*(6 players per team)*(60 seconds/minute)*(1/2 players on the ice are forwards) = 13,412,647 seconds. If you have a better idea as to how to calculate time on the ice among all NHL forwards or defencemen, I would like to know please.
Next, we calculate goals per standings point.
- 3,378 league-wide goals during regulation and overtime/1,383 total standings points as the beginning of April 15th = 2.44
Finally, we have all the ingredients needed to calculate offensive point shares.
- 13.64 marginal goals/2.44 = 5.6 OPS
- This is not what hockey-reference.com says, but it is pretty damn close considering the circumstances
All in all, I love the calculations except for one thing. Instead of making the statistic related to all forwards, why not relate it to all players within that player’s of his own position? As mentioned in the top-six forwards post, centers have higher standards than wingers now and wingers might be left in the curb as a result of it. However, of the 20 best players in goals created, 10 are wingers and 11 wingers would get into the top-20 in OPS. So maybe this should not be a big deal at the end of it all. Now, on to the Holtby Shuffle and goaltender point shares.
Goaltender Point Shares
First, we calculate the shots against adjustment for Braden Holtby. I love the start of this formula because it takes into account how tough or easy a workload a goalie is dealing with every game.
- (959 shots/1,786 minutes)/((29.0 league-wide shots/game)/(60 minutes/average hockey game)) = 1.111
- In other words, Holtby is seeing 11.1% more pucks than the league average. That certainly makes sense if you’ve watched every Caps game this year.
Next, the marginal goals against is calculated.
- (1+(7/12)) * (1.111 shots adjustment) * (1,786 minutes) * ((29.0 league-wide shots/game)/(60 minutes/average hockey game)) – (81 goals against) = 51.46 marginal goals against
Next, calculate league-wide marginal goals/standings point
- 3,378 league-wide goals/1,387 total standings points as the beginning of April 15th = 2.44
Finally, we get Holtby’s point shares.
- (2/7)*(51.46 marginal goals against/2.44) = 6.0 GPS
- That puts Holtby 9th among 23 NHL goaltenders that will play over half the season.
One thing to keep in mind about goalie point shares is that of those same 23 goaltenders, 21 would be the among the 100 most valuable players in hockey in terms of point shares. You can look at it as more than one-fifth of the list in a position that represents one-sixth of the team or you can look at it as 90% of all starting goaltenders are considered valuable. Does that all make sense, perhaps. At the end of the day, goalies are the only ones on the team that play 60 minutes a night and are the last lines of defense. As a result, you can tell when watching a hockey game that a goalie always can not take a break has to make all his plays when counted on more than anyone else on the team. Thus, that statistic makes sense as well.
In conclusion, outside of maybe finding a better constant for defensive point shares and comparing amongst people within the same position for offensive point shares, overall point shares has been a valid statistic.
Many thanks go to hockey-reference.com and nhl.com’s standings and stat page for helping me on this piece.