The overrated value of a faceoff


Every so often, an analyst will talk about how a certain player deserved a “three star” nominee because he won two or three big faceoffs. Even more so often a coach will send out two centremen in a high-leverage situation to win a key, late draw. The question is whether this tactic has a tangible value.

Advertisement - Continue Reading Below

Let’s use the Corsi number here. The Corsi number is, of course, an advanced +/- statistic that counts every goal, saved shot, missed shot and blocked shot while a player was on the ice and is a reliable indicator of which team had possession of the puck when a player was on the ice. Since a won faceoff essentially gives one team possession over the other, logically, faceoff percentage would correlate well to a player’s Corsi number.

In the beautiful, scenic spreadsheets offered at Behind The Net there exists data on 1799 players who have played 60-or-more games over the course of a season. I looked at a few, key bits of data from those players to determine faceoff value.

Microsoft Excel kept crashing on me, so I couldn’t label the chart. However, the Y-axis indicates a player’s Corsi number while the X-axis indicates the team’s faceoff percentage while he was on the ice. I’ve added a trendline, as well:

Advertisement - Continue Reading Below

With an r-squared value of .015, there is little correlative value between winning a faceoff and actually turning the possession into anything tangible.

I ran a similar correlation between where a player started his shift [ Offensive zone starts / Total Offensive and Defensive Starts ] and his Corsi number. Let’s see this result:

The r-squared value is .160. It doesn’t mean that there’s a determination in where you started your shift as to having tangible possession, but it does show us that the location of the puck is more important than who actually has the puck. This is where the dump-and-chase gets away with being a still useful, method of zone-entry. A team concedes possession for puck location and works to get it back in a similar spot.

Oilers blogger Tyler Dellow has looked at the value of a faceoff on the penalty kill recently and I have to add that I’ve come up with a similar conclusion at even strength. There’s a 60-40 split between the top and the bottom regular faceoff men in the league. At 10 draws a game (roughly) that equals two touches of the puck on your defenseman’s stick before anything can happen on the play.

I will add, however, that there are some faceoff specialists who double as strong defensive players: Manny Malhotra, Jarrod Smithson, Steve Ott, Paul Gaustad and [I guess] Selke winner Ryan Kesler had strong seasons on both draws and preventing shots. Players like Zenon Konopka and Jarrett Stoll had less generous defensive numbers, and Jonathan Toews, who came second in Selke voting purely by virtue of his faceoff skill was 169th in shot prevention among the 314 forwards who had played more than 600 minutes this past season.

Advertisement - Continue Reading Below

This site uses Akismet to reduce spam. Learn how your comment data is processed.

  • a lg dubl dubl

    I think a part of the problem the Oil had last year was the wingers lack of willingness to go help the centerman and dig for the puck, it cant always be left up to the center to win the battles on the dot.

  • Wax Man Riley

    Just by my “saw him good,” I saw the Oilers fail on the dot and saw the puck end up in the back of the net too many times in the last couple of years.

    I saw the Oilers dominate the draw in the ’06 run and remember how many times it was commented that they won another draw and have possession again.

    • Lofty

      Vancouver, San Jose, Detroit, Florida and Boston were the top 5 faceoff teams in the league last year. Aside from Weiss and Florida I think there’s some sort of correlation between being an elite team and winning faceoffs.

      I don’t know what the definition of a face-off win is? Is it just the draw or is possession factored in?

      Either way I think its a big advantage to have the puck more than the competition and it all starts at the drop of the puck. Without much evidence, I think its tough to be in the top 1/3rd of league without winning more draws than you loose. The Oil have been losing a lot of draws

          • Horcsky

            Correlation is not causation my friend, as my stats prof would say. Are elite squads good partially because they win faceoffs, or do they win faceoffs more so because they are good?

            Playoff teams that are below the ‘dreaded’ 50% in faceoff wins stat: Pens, Habs, Sabres, Rangers, Ducks.

            Also, this year, 2/3rds of the teams were above 50% in faceoffs, so let’s extend the above list to include teams in the bottom half of the league in faceoffs that still made the playoffs and we can add the following teams to that list: Bolts, Flyers, Preds.

            So 8 of the 16 playoff teams were in the bottom half of the league in faceoff percentage during the season. That’s just this year, so maybe it’s an anomaly, but I would also point out that teams #3-#27 in faceoff percentage are separated by a lousy 4 percent (roughly), meaning that most teams tend to vary very little in faceoffs, even though they are very far apart in the standings. This all seems to support the article.

            What does it all mean? I think Mr. Charron might be onto something.

    • John Chambers

      It seems odd but consider this:

      But when team A wins a faceoff, team B is in an optimal defensive position. Very rarely does team A win a faceoff then develop a scoring chance on the same puck possession sequence.

      More often than not, a scoring chance gets created as a result of a break-out, turnover, or offensive-zone cycle. Now to your point faceoff wins are more valuable in the offensive zone, and especially valuable when on the powerplay. From a game logistics standpoint, that is undeniable.

      I think what Cam is suggesting is that faceoff prowess isn’t as valuable a skill in comparison with tools such as footspeed, hands, or the ability to dispossess an opposing player of the puck. Ergo, a guy who posts a 46% faceoff success rate can hardly be considered a liability, and I wouldn’t go crazy armchair GM’ing a team’s need to throw big money after Boyd Gordon.

      • I’m not in complete disagreement here but, scoring chances are created as a result of a break out and the cycle. You win the draw back your wingers break out, hence first pass defense man are so important to have. The offensive cycle is created by winning the puck back in an offensive draw which is then kept in play by the offensive team.

  • Aitch

    If your team can’t forecheck, backcheck or keep possession of the puck once it has it, faceoffs are important. Otherwise, it’s a nice advantage, but hardly the be-all and end-all that some folks make it out to be.

    (That said, I’m still glad that we picked up Belanger.)

    • John Chambers

      Interesting question.

      I bet Toews wouldn’t get much attention on a bad team. Like Shane Doan. That said, I think a player like him who matches against the other team’s top players is an invaluable asset to a winning team.

      Coaches love him. Poolies and fantasy junkies, not so much.

      • OB1 Team Yakopov - F.S.T.N.F

        The Toews “un-love” by some of the supposed stats guys has become beyond ridiculous. Really good player and really good advanced stats to back it up.

        And Toews was a deserving Selke nominee.

    • Hemmertime

      Yes. By some. I like Toews, he is a very good Center and a great leader. That being said without Patrick Sharp being as good as he is as a #2 (or 1B) Toews would not look near as good as he does and the expectations on him would be ten-fold. People would complain about how he only gets cracked 70 pts for the first time this season and say he’s disappointing. The person who mentioned Doan is a good example of good player bad team. Toews has his great reputation because of good player + great team.

      You know who puts up Toews point #’s (does it on Wing not Centre, and C’s should get more pts) and plays against the other teams toughs? Hemsky. Toews 0.89 PPG last 3 seasons, Hemmer – 0.91. Much much much worse team (Stanley cup vs Last) and yet most people regard 83 as a decent second line RW and praise the hell outta Toews. The east people are so uneducated about the Western Conference its laughable. We have the east force-fed to us on all channels so at least we know enough about the Eastern Conf. to usually wipe the floor with the East fans in the pools.

      Even more food for thought: Would you trade RNH and a 2nd rounder for Toews? Its tempting, but I think I’d lean towards no. Toews took a step forward this year but doesn’t strike me as a 100 pt man. RNH might be 60-90, but Ill take that risk over guaranteed 70

      • OilerLand

        There is NO WAY you can get Toews from the Chicago Blackhawks for Hemsky and a first round pick. Even a top pick.

        And Toews is on a Selke path not a Hart one. Eventually, I think you will see Kane and Toews split on two separate lines so each can go towards their respective “trophies.” Which is really saying contribute using their individual talents to their fullest.

  • John Chambers

    Hey Colin White is on waivers. Should the Oilers or Flamers pick him up?

    One year left on a contract that pays the 33-year old 3 mil. Last year was only -2 on a rotten Devils team.

    For your money and cap hit it’s a far better value than Kotalik wouldn’t ya say?

  • And then I was all like “check out the value of R squared!” and he was all like “DUDE, check out the sweet conditional formatting on this excel sheet.” and then I was all like, “NO WAY!!!”

    So what you’re saying is that all these smart hockey people like controlling the puck for no reason because excel said so? Must be true. I read it on the Internet, and nobody lies on the internet.

  • John Chambers

    And yet every coach in the NHL game plans based on winning the face-off and will tell all who ask that it is one of the most critical points of the game.

    Maybe someone should e-mail them the corsi to show them they are doing it wrong.

  • Lawndemon

    Ok, for one thing, R-squareds are more relevant to time-series data. You’re using cross-sectional data which always has a low R-squared, so you can’t just downplay faceoffs because of an indicator that favours other forms of data.

    Second, whoever said Corsi numbers were the be-all and end-all of statistics? I’m sure every coach in the league is putting his best faceoff man out there because he’s thinking “man, I need to get my Corsi numbers up!” How about doing a correlation between team wins and team faceoff percentage?

    Lastly, Sam Gagner is terrible at faceoffs and should be traded at the end of this year, but that should go without saying.

    • ubermiguel

      For the simple reason that in order to score, you need to shoot and goals represent a small % of the shot attempts.

      And I’m pretty sure that coaches are putting their best face-off man out there to either help the team generate a shot (offensive zone draw) or prevent a shot (defensive zone draw), especially in the late stages of a hockey game.

  • Lofty

    At some point…the mathletes are going to figure out that “Time of Possession” is more useful in measuring the worth of a hockey team (and the players who are on the ice at any given time) than Corsi or FO winning%.

    Thing is, it’s pretty easy to measure, and yet none of these guys have figured that out yet.

    • ubermiguel

      Of course possession time is more useful, and while it’s easy to measure for an individual team, it’s much harder to measure over the the whole league.

      This is why, in lieu of TOP, things like Corsi/Fenwick exist. If there was a way to get the TOP in the various zones, it would be used far more.

  • stevezie

    No one is saying faceoffs don’t matter, just that they don’t matter as much as is often claimed. I thought the linked Dellow article did a very good job of showing that you’re better off with a good player who is bad at faceoffs than a bad player who is good at them.

    Maybe faceoffs are in the same category as size, fighting, leadership etc.; a really great thing to have but pretty useless without the solid hockey skills. Let’s not confuse the icing with the cake.

    • paul wodehouse

      That’s some ace research right there.

      How about a good player who is young vs a bad player who is experienced or a good goalie who is small vs a bad goalie that has size?

      Why not compare apples to apples instead?

      Obviously going with the best players available is where you start but in the games within the game further skillsets and player attributes do matter. And in this case where the coaches are outright telling you that it’s vital then it’s pretty much a non-question.

      • stevezie

        Experienced hockey guys agreeing on something doesn’t make it a “non-question.” The establishment has been wrong plenty of times before; one reason the game today is different than the one played in 1950 is a series of people realized that “common knowledge” was wrong and exploited widely used but flawed tactics. Lafleur’s Canadiens couldn’t beat many teams in the alleys, but they won a lot of cups.
        Again, no one is saying faceoffs don’t matter, but maybe Yanick Perrault was a waste of a roster spot for that last year. Maybe the ducks were right to look past Cogliano’s faceoff problems. Maybe when comparing apples to apples you’re putting too much value on shininess when you should focus more on taste and texture.
        This actually was some ace research, though not by me. I thought Dellow really nailed it.

        • stevezie

          Nah it’s still pointless research.

          Only someone looking to justify a spread sheet will compare different player types with different skills to make a point.

          Anyways, the comment about coaches wasn’t a suggestion that old timey hockey guys are always right, it was pointing out that these guys put an emphasis on the importance of face offs and they are the ones that are currently running the game.

          Faceoffs are important because the coaches put an emphasis on it, rightly or wrongly.

  • SurfacetoAirMissile

    Stats refer to an average over time. In an average game, at an average point in it, no, face offs don’t matter much.

    But as time runs out, each event becomes more significant if the game is still on the line.

    Nobody sweats it in the first period of the first game of the year, but they really care at any point in game 7 of the finals.

    The world cannot correctly be viewed through numbers that have no correlation to it’s complexity or it’s tendency to do enormous things in very short time spans. Like hockey.

  • ubermiguel

    Over the last few years I’ve watched the Oilers lose too many defensive zone draws only to get pinned in their own end for a couple of minutes to say the “faceoffs don’t matter” theory passes the sniff test.

    Maybe it’s something about players that can’t retrieve the puck or defend in general.

    Still, I’d rather start a shift with the puck than without.

  • Quicksilver ballet

    Winning the faceoff is overated.

    Just be the better team getting off the bus. Teams have possession of the puck 70+ times a game. It’s what you do with it while you have possession of it that matters. Starting with possession isn’t as important as finishing the play with it.

    • book¡e

      Well, you may be LYFAO, but while doing so, you are identifying yourself as someone who would have trouble thinking more than one move ahead in a chess game.

      Winning faceoffs would clearly have some value in NHL hockey, however, this value may be very very small given the number of times that possession changes and the nature of the advantage of that possession. The positioning and nature of immediate possession after the drop of a puck may have very little influence on the outcome of games. The math provided in this analysis suggests that may be the case. So, it may be that team building strategies that emphasize winning faceoffs (and selecting personnel based upon that) may be a poor strategy.

      • Or maybe the stats and formulas used for the study arent the best way to interpret whether winning a faceoff is important.

        Considering the great disconnect between the focus of professional coaches and management’s desire to win faceoffs with regularity and this study’s result showing that “Winning faceoffs is over-rated”, maybe the results and how they were contrived should come into question before the concept that winning faceoffs is important comes into question.

        • book¡e

          I can fully agree with that – its always wise to take a critical approach to such things. It was just that the LMFAO statement sounded a lot like blatant and arrogant resistance to any notion of challenging the ‘obvious’. There are many instances where prevailing common knowledge has been successfully challenged by careful analysis.

          Now, with that said, at a quick glance, the argument made here does appear to be well put together.

      • You clearly identify yourself as the man that points out the obvious. Maybe I am playing presumptuous after you led the way with your enigmatic chasm. Maybe I have the attention span of a gnat, maybe I think this topic is useless. Maybe my opinion doesn’t weigh on you because I provided no back-up for my laughter. It appears I don’t need to when you so eloquently spell it out for me.

        When a team is 3pts back of a playoff spot, a few more faceoffs won on the PP, and who knows. The analysis would have no bearing on an organization in a league that with such intense parity, requires every inch

  • SurfacetoAirMissile

    As important as faceoffs are and having great centers, I think we may have a more pressing problem on the left wing in 2012/2013. Where is MPS gonna play with Smyth, Hall and Parise ahead of him?

  • OB1 Team Yakopov - F.S.T.N.F

    This just seems like a shot at justifying Connolly’s losing face-off % over Bozak’s Winning face-off %. There are a lot more face-offs than goals. There is where you’re low numbers come from. This should be broken down more into face-off locations and even strength vs. PP vs. PK on a team to team player to player basis. A lot of work to prove anything can happen. Strategies and ability of the opposition to read plays is another tangible. The fact of the matter is you don’t win without possession and the face-off is where it all starts.

  • Death Metal Nightmare

    hockey is chaotic.

    thats basically the point of all these lame stats. its chaotic. its not baseball and its not rigid football. small, micro-events always happen that can keep trends from surfacing.

    are faceoffs important? depends.
    is shooting important? depends.

    can you still play sick defense and have the puck go off your skate or body into your own net? sure.

    did Vancouver have the best numbers of the regular season? yes. did determinism dictate they win the Cup? nope.

    numeric determinism in hockey makes hockey boring. chaos and chance is why we watch it. keep winning faceoffs the best you can.

  • Lawndemon

    It isn’t much of a logical stretch to conclude that faceoff losses are more significant to bad teams (teams that can’t regain possession). When the Oilers lost a draw it was balls-deep panic time.

    Meanwhile, I recall a number of instances where the Oil won the faceoff and still couldn’t get the puck out of their own zone or get a shot on goal. Why? Because the other team just took the puck away from them anyway.

    Bad overall hockey players are for more significant to team success than bad faceoff artists.

  • Wax Man Riley

    i’m sure it has already been said, but some stats must be taken into context. An offensive zone faceoff in the first 5 minutes when the game is 0-0 does not matter as much as an offensive zone start with 30 seconds left when the offensive team is on the powerplay and down 5-4.