(Photo: Sportsnet.ca)

by Mathieu Chagnon, Staff Writer, All Habs Hockey Magazine

As promised, I have continued to develop the eSAT hockey metric.

First, you might have notice that I’ve renamed this tool from SATSE to eSAT. I did it for few reasons: it’s shorter, it is less overpowering visually, it more closely resembles the existing  SAT (shot attempts) and most importantly, it better represents the metric.

In addition to the name change, eSAT has been further developed to incorporate the quality of opposition faced by a player, and the measurement of how they performed against that opposition.

As the project unfolds, I will be able to do player profiles based on these statistics in the future.

First let’s catch you up on exactly what this measurement captures.

In brief

First, eSAT is an acronym for effectiveness of the Shot ATtempts. Basically, the performance of a player is measured on the calculation of shot attempts and shooting percentage, when that player is on the ice.

In the past, a correlation of the standings was made by doing the differential between the on-ice goal for and on-ice goal-against, but this couldn’t provide a weight of the contribution from the players.

Now with this metric, we can not only measure the contribution of a player to his team’s  success, but it has a demonstrated success rate of 92.4 percent for a team’s ability to earn a playoff spot.

Using a simple deductive process, you could conclude that to have more goals, you just need to make more shot attempts, but this simply isn’t very accurate. A team can record twenty shots and scoring four times, while the opposition fires from every part of the ice registering thirty shots and is blanked.

So we know that shots location is important along with a number of other factors. They affect the shot efficiency.

For instance, here are just a few situations:

  • winning the race for the puck in the attacking zone, keeping the pressure in the zone, where it will ultimately lead to a goal
  • bodychecking the defender, and helping the recovery of the puck for his team, leading to a pass in the slot, and a dangerous shot taken
  • fighting in front of the net, and obstructing the view of the goalie before a shot is taken
  • being forced by the defender to shoot from a sharp angle and becoming an easy save

This is why eSAT is an algorithm that uses the Corsi For (CF) and Against (CA), the On-Ice Shooting percentage (OiSH %), the On-Ice Save percentage (OiSV %) and Time on Ice (TOI).

For a more detailed perspective, I highly suggest reading my  introductory article on eSAT.

Quality of Opposition

While we could satisfy ourselves with the global performance of a player, we must not forget that the opposition on the ice is not always the same. Therefore, we must look at the quality of defensive and offensive opposition faced by a player.

For the evaluation of the defensive opposition while our subject is on the ice, we must extract the data of all the players he skated against.

For example:

VS TOI Against Season eSAT A/min eSAT A Opp
Aaron Ekblad 5.65 0.094 0.5
Aaron Ness 6.21667 0.087 0.5
Adrian Kempe 2.53333 0.049 0.1
Alec Martinez 14.5167 0.094 1.4

 

The first column represents all the players he played against. The second column is the amount of time he spent against them. The third column, is their respective performance in eSAT against per minute for the current season. The fourth column is the multiplication between the second column (TOI Against) and the third column (Season eSAT A/min), creating the eSAT against that the opposition should have allowed (eSAT A Opp).

The fourth column is added up, and then divided by the total of the second column, this will give us the average eSAT against that the opposition is usually allowing (Opp eSAT A). That number will be evaluated against the league average (Opp eSAT AAL). If the number is higher than 50 percent, it means that the opposition doesn’t allow many efficient shot attempts.

Now that we know the strength of the opposition, we want to know how our subject is performing against them.

First, in the table above, we learned that our player played 14.1567 minutes against Alec Martinez. In the hidden columns, we that he has a Corsi For of 14 while playing against him and an on-ice shooting percentage of zero percent, which give us a eSAT for of 0 against him.

That process is applied to all the players he played against and this whole column will be added, subsequently divided by the total of second column of the table, therefore giving us how many efficient shot attempts per minute the player we are evaluating obtained against his opposition (eSAT F).

That eSAT for is then compared to the Opposition eSAT Against Above League (Opp eSAT AAL) creating the eSAT For Above Opposition (eSAT FAO). It is to be noted that due to special units, those two numbers cannot be compare as is, therefore a technical adjustment is done to put them under the same common denominator.

And because the eSAT FAO is now measured to the Opp eSAT AAL which was previously compared to the league average situated at 50 percent, the floor of the eSAT FAO cannot stay at 50 percent. It must follow the variation of the Opp eSAT AAL, if not, those two numbers put side-by-side could look counter-intuitive.

Are you following? Let me clarify this for you, by way of illustration. If our subject has an Opposition eSAT AAL of 55 percent (which is five percent above the league average) it is expected that he will produce a lower eSAT for, because the opposition is stronger. Therefore, if he produces the same level, the calculation will show a eSAT FAO of 50 percent, which means he’s not doing better or worse than his opposition.

But looking at those numbers in a graphical illustration, it would look as if our evaluated player is not as good as his opposition. To resolve this, the formula moves the floor higher or lower depending the opposition.

Data extracted on December 14th 2017

An interesting option can be done to see if a player performed better against the weaker or the stronger of each category. The same process is done except that we separate the over-the-league average (o50) and the under-the-league average (u50). Some players will be constant, other will show a sizable difference.

Data extracted on December 14th 2017

But one of the most important indicators would be to look at the situational department. Sometimes the all-situations data could be biased by the fact that the player is playing the  penalty-kill a lot and he will have less success than he’s having at five-on-five or having a lot of success on the power-play.

Data extracted on December 14th 2017

And all this wonderful process is applied to see how well our player is performing against the opposition offense. I’m not going to describe it, but you will understand that the methodology just inverts the “against” and the “for,” and it is going to be call, eSAT AOS, short for effectiveness of the Shots ATtempts Against Opposition Suppression.

 Quality of Teammates

Calculating the quality of teammates is a lot simpler. In fact, the fundamentals are the same as the individual eSAT, but the only difference is that we just get to look at the statistics separately between the teammates. Doing so we can observe with which players our evaluated subject is having the best performances.

Data extracted on December 14th 2017

In the future, I’ll be doing pieces on players with those kind of graphics and giving more information about it. At this time, they are presented as examples, so that you can visualize them a bit.