Advanced Faceoff Stats Volume 1: The Best And Worst NHL Faceoff Specialists Last Season.

What do we know about a faceoff? It is defined as the start or a restart of play, in which the referee drops the puck between two opposing players. The winner of the faceoff is whoever touches the puck first, subjectively decided by the arena scorekeeper. Since the stat is somewhat subjective, players playing at home often receive higher faceoff wins from the arena scorekeeper. This is common with other subjective stats like Giveaways, Takeaways, & Hits as well.As the hockey community continues to make advances statistically, it was only natural for someone to give the faceoff a facelift. It first started with looking at goals that occur within 20 seconds of every faceoff. The problem with raw goal based stats is there plain isn’t that many goals to go off of. low amount of goals leads to small sample sizes and a higher propensity of variance. That’s why shots are so often used to greatly increase the sample size of a stat which helps remove some randomness.

Craig Tabita of decided to take a shot based approach to faceoffs. He found that shots are most influenced within 10 seconds after a faceoff. Shots that occur AFTER the 10 second period are LESS likely to be influenced by a clean faceoff win and more a result of outside factors. This is why he only uses the 10 seconds following the puck drop to determine post-faceoff possession stats. 

Click here for Tabita’s explanation of his stat Net Shot Post Faceoff (NSPF) which goes in depth on how it is calculated.

I’m going to attempt to give a simplified explanation of the stat which may be tough. You see, I try to explain even the most complicated stats to be simple enough for someone that’s new to hockey to comprehend. This is difficult and sometimes takes work from the viewer to further research the topic. Also, like in sports where reps will make the player better, more reps will make people better understand stats. The more one looks at them, reads about them, the more they click and thus become easier to comprehend.

Therefore, I want to preface that before writing this, I read a chapter in Rob Vollman’s book StatShot about this stat. I followed up by reading Craig Tabita’s posts about his stat. Finally, I played around with the data to further grasp the stat. After all this work, I understand the stat. I get it. However, it still doesn’t come naturally to me, it is still difficult to try and teach something I just learned, and like I said earlier, it’s going to take some reps and some time before I feel totally comfortable with the stat. The point is, don’t expect to always 100% understand a new stat quickly. The next couple paragraphs might go over your head, but don’t worry about it. Don’t get down on yourself if you struggle to understand a stat. It takes time. It takes effort. 

Anyway back to NSPF. Remember, that’s the faceoff stat we are looking at, and the acronym stands for Net Shots Post Faceoff as we are literally looking at shot attempts post faceoff. We are examining shots for and shots against, so one could think of it as a post-faceoff shot +/-. To explain this stat, we must first look at it from each zone.

If Jonathan Toews is taking a faceoff in the offensive zone, Chicago is much more likely to have a shot attempt in the first 10 seconds than the opposing team is. Because of this, centers normally always have positive NSPF numbers in the offensive zone. For example, in the 2015-16 season, Toews took 374 even strength faceoffs in the offensive zone. In the 10 seconds following each faceoff, Chicago has 88 shots for and only 8 shots against. That’s a +80 shot differential in the first 10 seconds after a faceoff in the offensive zone. Good right? Actually Toews was below league average last season. More on that in a later Volume.

In the defensive zone then, a center is most likely going to have negative shot +/- numbers. If Jonathan Toews takes a faceoff in his defensive zone, it’s going to be much easier for the opponent to take a shot than for Chicago to gain the puck, travel the length of the ice, and get a shot off. Last season, Toews took 317 faceoffs in his defensive zone at even strength. In the 10 seconds immediately following every puck drop, Chicago had 6 shots for and 97 shots against. Again, this is to be expected as in the defensive zone it’s much easier for the opponent to take a shot than Chicago.

The point of me explaining faceoff shot +/- by zones is obviously that’s going to have a big influence on someone’s overall faceoff possession numbers right? If one player takes a majority of his faceoffs in the offensive zone, he’s going to have better numbers than a player that takes a majority of his faceoffs in the defensive zone. 

This stat covers for that so one doesn’t need to worry. However, this is where things get kind of tricky as math is involved. For the actual calculations, click the link in the third paragraph of this post. I’m going to attempt to simplify it and explain it without numbers.

So we know that the raw data is flawed. As I explained above comparing two players isn’t fair depending on which zone they take the majority of their faceoffs in. But what if we didn’t use the raw numbers? What if we compared to league average in a fashion similar to how Baseball’s WAR compares to a replacement level player. Therefore, in the stats we look at zero will equal league average. A positive number will be above league average. A negative number will be below league average. In this fashion, we can easily compare every centers numbers. Each zone will be separated, compared to the league average, and then added together to show how many shots above or below league average that center did for a season.

Here is how Craig Tabita’s describes the stat on his website: “Net Shots Post-Faceoff (NSPF) estimates a face-off taker’s contribution to puck possession by counting shot flow (shots-for minus shots-against) during the 10 seconds following a player’s face-offs in 5-versus-5 situations, and comparing it to a league-average face-off taker. NSPF is calculated for each of the offensive, defensive, and netural zones; the overall NSPF is their sum. 

In short, NSPF is how many more shots-for a player’s team had, plus shots-against prevented, following face-offs he took compared with what a league-average face-off taker would have done with the same number of 5-versus-5 face-offs in each zone (offensive, defensive, and neutral).”

To get an idea, let’s take a look at the best and worst centers last year at NSPF.

Sidney Crosby took 1,374 faceoffs last season at even strength. His team produced 76 more shots than a league average center taking the exact same faceoffs. That was best in the NHL.

Artem Anisimov took 838 even strength faceoffs last season. Chicago produced 39 less shots than a league average center (in terms of faceoff ability) would have taking the exact same faceoffs as Anisimov did. That’s good for worst in the NHL.

The problem with this stat is that faceoffs aren’t equal. A players number will be exaggerated with more faceoffs and hidden with less faceoffs. It’s like goal scoring, if a player gets 2000 minutes, they will likely have more goals than a player who has 500 minutes. That’s when we use goals per 60 minutes of Icetime stats to make the minutes played equal and fair. Likewise, for faceoff possession, we can look at Post faceoff possession per a certain amount of faceoffs. On his website, Craig Tabita uses per each faceoff for this defining it as: “NSPF/FO is the NSPF per face-off, allowing for easier comparison among players with different numbers of face-offs taken.”

The only thing I don’t like about this is it creates such a small number for use in comparison. It’s similar to goals per 60. If we used goals per minute, it would be such a small number for each player. Averaging goal scoring out to per 60 minutes results in easier numbers to work with. Likewise, I will change that stat for NSPF per faceoff and instead use NSPF per 60 faceoffs.

Let’s take a look at the top-10 and bottom-10 last season in the stat Net Shots Post Faceoff Per 60 Faceoffs. This puts everyone on an even playing field. Think of this as simply faceoff ability, or Post faceoff possession.

Sidney Crosby was the most positive center overall, but he also had more faceoffs to enhance those positive numbers. When we put things on a per 60 faceoffs rate, everyone is on equal ground. 

Here we see Crosby drops to 4th best faceoff specialist last season with the number +3.32. This means for every 60 faceoffs Crosby takes, his team will produce 3.32 shots better than a league average center would in the same situation.

Likewise, Anisimov moves from the worst overall to the 10th worst in terms of rate. This is because he had a lot more faceoffs than those below him here. Anisimov’s negative number was exaggerated by his great number of faceoffs compared to others. Even when we make faceoffs equal, he was still one of the worst faceoff specialists in the league last season. Someone else we know makes the list as the second worst faceoff center in the league, Teuvo Teravainen.

Now, no one is saying this stat is perfect. But faceoffs in and of themselves are a flawed stat. Both stats can be influenced by teammates, competition, & coaching. I’ve already had someone tweet me this “So if he (Toews) wins the faceoff and the the Dmen turns over the puck that’s Toews fault? headshake.” Upon hearing a stat that tracks shots 10 seconds after puck drop, this person just immediately wrote it off as useless and illogical. Maybe in that scenario Toews didn’t win the puck clean enough. Maybe a better faceoff specialist wins the puck cleaner and the D-man doesn’t turn the puck over. How often does that scenario even happen to begin with?

When we look at regular faceoff stats, how often did the center lose the offensive zone faceoff on purpose just to make the opponent skate the full length of the ice before they can attempt to score. How often does a winger or defenseman deflect the puck, never getting an official “touch” and the center loses the faceoff. Yes there are problems with regular Faceoff stats. Yes, the advanced post faceoff possession stats aren’t perfect. But there are interesting. They do add more detail, more context, and a better understanding of being a true faceoff specialist. A cleaner faceoff win can that leads to a shot attempt is shown in these advanced faceoff stats. Likewise a sloppy faceoff win that doesn’t result in a shot attempt because the center did not win the puck clean enough can also be shown. We’re simply trying to get more in depth on Faceoff numbers. 

Never be afraid to open your mind to new thoughts and ideals. So many people in this world are close-minded, living in their own bubble, & unaware of the huge world around them. The same thing applies to stats. There are people out there like you reading this post, opening themselves up to a new idea, learning about a new stat. Growing! Congratulate yourself. 

I have plenty more I want to go over, but it seems like this is a good time to pause and turn this post into multiple volumes. This, Volume 1, was the best and worst faceoff specialist overall. I will post further in Volume 2 about the best offensive zone & best defensive zone faceoff specialists. Volume 3 will compare and contrast Chicago Blackhawks centers only. By Volume 4, I hope to look at Blackhawks players season-by-season Faceoff numbers. For example, over the years Toews has went from an elite faceoff specialist to below league average in this stat. There should be plenty of interesting things to come. I hope you’ll join me in the exploration.

If you would like to look at the faceoff possession database page to see the stats yourself, simply click here.