Embedded Files
*FIH PER adjustments can be found here
John Hollinger created a Player Efficiency Rating (PER) to use for the National Basketball Association (NBA). The stat looks at the efficiency of players during their time on the court, in other words: how effective are you while you are playing?
Ice Hockey, baseball, and soccer have adaptations that have been created over the years. With no higher level to achieve in field hockey, other than the FIH Pro League, no one has previously created a formula for use in our sport.
I decided to tackle this issue by creating a formula for use in field hockey with the most commonly produced statistics on collegiate field hockey websites.
The Formula
I wanted to take inspiration from the Basketball PER, however, basketball has more complex statistics such as 3-pointers, rebounds, and steals, making this task harder.
The first step was to determine the important factors and commonly produced statistics for field hockey.
Looking at this chart, there are two main flaws: a lack of negative statistics and a lack of defensive statistics.
For negative statistics, there is not much that can be done, unless you have someone (or AI) to hand-count all of the passes that have been received and missed throughout a match. Respectfully, as someone who has done this before, counting passes is not a useful way to spend your time.
For defensive statistics, the FIH Pro League also has steals, blocks, and interceptions. These have not been brought down to the collegiate level at this point, although steals have been adopted in some high schools. As the sport grows, hopefully, teams will take into consideration the benefits of using these measures, but for now, the formula will no include these.
Another issue that occurred while collecting data, was the lack of standardization in posting statistics. 20 teams did not post player's minutes, while a large number of teams had minutes that were not accurate (think all starters have 60 and substitutes are listed as 0). While collecting data, two programs did not have statistics publicly available and three programs had not published statistics for their freshman classes.
The Formula:
(2*G)+(A)+(0.6*SOG)-(0.5*GC)-(YC)-(2*RC)+(DS)
Okay, so I will admit that it looks a little boring.
A goal (G) being worth 2 points and assists (A) being worth 1 was determined by the standard points system the NCAA uses.
For shots and shots on goal (SOG), I ran a correlation coefficient test for SOGs compared to Goals scored. The resulting correlation was 0.626373748, which I rounded to 0.6 for the sake of simplicity.
For cards, a correlation coefficient test showed that a yellow card (YC) had double the negative correlation on the result of a game compared to a green card (GC). Therefore, I made a green card equal to ½ and a yellow card equal to a whole. The red card correlation was unclear given the limited amount of red cards (RC) this season, so given the correlation between green and yellow cards, a red card is worth double a yellow.
For defensive saves, since they are the only defensive measure and are a relatively rare occurrence, I determined the weight to be 1.
*Note* While I did not weight these results by minutes due to the data gaps when determining this metric for your team, include
(1/MP)
In front of the above equation.
Then to weigh the data, I calculated the average unadjusted result for each player, which resulted in 0.61313438.
Then I ran
unadjusted*(5/0.61313438)
The 5 in the equation represented the average PER for each player. The NBA used 15 for this equation, however using a number that high would create outlandish numbers, so I adjusted it accordingly.
Team Results
I was surprised to find that Maine (shown in graph below) was found to be the most efficient team with an average team PER of 10.2.
Maryland followed closely behind with a 9.58 and UNC was in third with 8.0.
One trend that I noticed is that teams that tend to play with more defenders on the field received lower PERs, which makes sense given that the formula weights offense so heavily. Teams that utilized their defense in offensive corners had an edge since those players were attributing to the offense as well.
Another factor that lowered the team’s PER was when they had many substitutes. Teams that played their entire rosters tended to have lower PERs because they had more zeros in their averages.
At the end of the day, this is a player ranking system, so let’s get into the players.
Players
I think it is a surprise to no one that Erin Matson was the most efficient player this season, with a 43.8. For those of you unfamiliar, a 35+ is considered an all-time great season in basketball.
The top 10 most efficient players during the 2022 season were:
Erin Matson (UNC)
Annick van Lange (Monmouth)
Poppy Lambert (Maine)
Allison Smisdom (Albany)
Sophia Gladieux (Penn State)
Bente Baekers (Northwestern)
Valeria Perales (Rider)
Beth Yeager (Princeton)
Eva Valentini (Saint Francis)
Lily Neilson (Bucknell)
The mix of conferences is very interesting, with only 1 ACC school, 2 Big 10, 2 NEC, 1 Ivy, 1 CAA, 1 Patriot, and 2 American East.
Matson’s most efficient game had a resulting PER of 102.7 which came against Liberty. Van Lange’s most efficient game had a resulting PER of 88.1 against Northeastern, and Lambert’s was 94.6 against Merrimack.
Overall, most of the players fell in the 0 to 5 category, which is expected since the average PER was 5 for the formula. 719 players fell into this category.
82 players had negative numbers due to cards, and 326 players had a PER of 0.
Only 16 athletes reached a PER of over 25, and 20 players received between 20 and 25.
So What Does It Mean?
The implications of this calculation are really up to the user.
From a coaching perspective, you can see how efficient players are with the minutes they are given. This can be used to make decisions regarding playing time. When looking at opposing teams, you can also determine which players will need to be more heavily guarded and if your opponents have any secret weapons.
Otherwise, this is a fun way to see how you or your favorite players rank up against the best in the league.
Data Requests
If you would like to request the PERs for you or your team, please fill out the request form.
If you have any further questions, please feel free to contact me either via the form on my website, my email, or Linkedin/Twitter and I will get back to you as soon as possible.
Happy Holidays!
Page updated
Google Sites
Report abuse