NHL Possession Efficiency Ratings

NHL statistics, ratings, possession analysis

PIT@NSH Game 4: Microstats

leave a comment »

Pittsburgh at Nashville Game 4:  2017 NHL Playoffs
Individual Possession Analysis
Nashville 4, Pittsburgh 1

The below charts are the results of my individual possession tracking.  Essentially documenting what each player does per each possession they have at 5v5.  Corsi%’s were retrieved from firstlinestats.com  The below chart shows each player’s turnovers per possession, and which zone they occurred in.

Maatta and Rust struggled in the defensive zone in terms of turnovers, with defensive zone turnover rates per possession at .16 and .14 respectfully.  The team average defensive zone turnover rate was .059 (~1 in 17 individual possessions).

Matt Cullen struggled with turnovers in the offensive zone, but his line mates Kessel and Kunitz picked up the slack and had lower turnover rates relative to the team average.  Average offensive zone turnover rate was .094 (~1 in 11 individual possessions), average neutral zone turnover rate was .051 (~1 in 20 individual possessions), the average defensive zone turnover rate was .059 (~1 in 17 individual possessions), and the average dump in turnover rate was .065 (~1 in 15 individual possessions).

Screen Shot 2017-06-07 at 11.06.42 PM

The below chart looks at each player’s percentage of possessions where they completed a pass to a specific zone.  Phil Kessel had 39 possessions in game 4, with a .26 Offensive Zone Passes Completed Per Possession.  Meaning that of Kessel’s 39 possessions, 10 of those possessions resulted in him completing a pass in the offensive zone.  The team average was .13, so Kessel was twice the team average.

Crosby had the highest per possession rate at .32, so ~1/3 of Crosby’s possessions at 5v5 resulted in him completing a pass in the offensive zone.  Schultz with .46 Defensive Zone Passes Completed Per Possession was the highest on the team.  So almost half of Schultz’s 5v5 individual possessions resulted in him completing a pass in the defensive zone.  Given the fact that Schultz is a very accurate passer based on my tracking, he was able to obtain a Corsi% of 57.7%.

Screen Shot 2017-06-07 at 11.07.57 PM

The chart below shows each player’s Shot Attempts (SOG, Missed, Attempt Blocked) and Shot Attempt Assists per individual possession.  The average shot attempt per individual possession was .098 (~1 in 10 individual possessions).  The average Shot Attempt Assist per individual possession was .07 (~1 in 14 individual possessions).  Archibald and Kessel led the way with .25 and .21 shot attempts per possession respectively.

Kessel and Schultz led the way in Shot Attempt Assist per individual possession at .21 and .17 respectively.  Kessel Assisted on a shot attempt on ~1 in every 5 of his individual possessions.

Screen Shot 2017-06-07 at 11.26.40 PM

Rebound% shows which players are getting to loose pucks following a shot attempt in the offensive end or the defensive end.  The Crosby-Guentzel-Rust line made the most of their offensive rebound opportunities by obtaining ~27% of the offensive zone rebound available at 5v5.  The Crosby-Guentzel-Rust line offset their well above average offensive rebound% with high turnover per possession rates.  The line average at 30% turnover rate at 5v5, where the team average was 20.4% in game 4.

The team average Offensive Rebound% in game 4 was 10.9%, and average Defensive Rebound% was 10.6%.  The Crosby line was hemmed in their defensive zone, unable to obtain defensive rebounds, thus prolonging Nashville’s offensive zone time.  Maatta and Daley struggled to obtain defensive rebounds, combining for 0 defensive rebounds between them.  Hainsey managed above average offensive and defensive rebound%’s at 15.8% and 12.5% respectively.

Screen Shot 2017-06-08 at 12.15.24 AM

This chart looks at steals per zone at 5v5.  Looking at the defensemen with high steal counts (Cole, Dumoulin, and Maatta), you can assume that they have a lot of opportunities to attempt those steals, thus more offensive zone possession for Nashville while those players were on the ice.

Hagelin is so quick in the offensive zone, that he is able to apply instant pressure once Pittsburgh loses possession of the puck or fails to get a rebound in the offensive zone.

Screen Shot 2017-06-08 at 12.23.10 AM

 

Advertisements

Written by RReed

June 8, 2017 at 12:32 AM

Posted in Uncategorized

PIT@NSH Game 3: Microstats

leave a comment »

Pittsburgh at Nashville Game 3:  2017 NHL Playoffs
Individual Possession Analysis
Nashville 5, Pittsburgh 1

The below charts are the results of my individual possession tracking.  Essentially documenting what each player does per each possession they have at 5v5.  Corsi%’s were retrieved from Hockeystats.ca.  The below chart shows each player’s turnovers per possession, and which zone they occurred in.

Overall Malkin had 19 individual possessions at 5v5, and committed a turnover 10 times (.53 or 53% of his possessions).  Kunitz did not fair much better, committing 7 turnovers on 14 individual possessions (.50 or 50% turnover rate).  Malkin had the highest offensive zone turnover rate at 5v5 at .37 offensive zone turnovers per possession.

The team average turnover rates at 5v5 per each zone for game 3 are as follows:
Overall  .27  (~2.7 turnovers per every 10 individual possessions)
Offensive Zone .124  (~1.24 offensive zone turnovers per every 10 individual possessions)
Neutral zone .04  (~1 neutral zone turnover per every 25 individual possessions)
Defensive zone .05  (~1 defensive zone turnover per every 20 individual possessions)
Dump ins .05  (~1 dump in turnover per every 20 individual possessions)

Screen Shot 2017-06-04 at 11.32.31 PM

As you can see above, the players that were able to reduce turnovers in the neutral zone, were able to achieve a higher corsi%.  Kunitz struggled in the neutral zone, committing a turnover on 43% of his individual possessions at 5v5.  Malkin and Kunitz had the two highest neutral zone turnover rates at .11 and .43 respectively.

The players with a sub 50% corsi% for the most part, were unsuccessful at dump in attempts at a higher rate than those with above average corsi%’s.  With most committing turnovers on dump ins at almost twice the team average of .05 (~1 in 20 individual possessions).   Hagelin at .10 (~1 dump in turnover per 10 individual possessions), Hornqvist at .10, Daley at .08, Cole at .10, Kessel at .07, and Kunitz at .07 dump in turnovers per individual possession.  Pittsburgh has far too much talent to settling for dump ins as a means to enter the offensive zone.

Below shows each players pass attempt success rate to each zone at 5v5 (Completed/Attempted).  The blank squares reflect that a player did not attempt any passes in that particular zone.

Screen Shot 2017-06-04 at 9.50.33 PM

The below chart shows each player’s shot attempts per possession(SOG, Missed, Att Blocked), as well as their shot attempt assists per individual possession.

The team averages for game 3 were:
Shot Attempts Per Possession: .11 (~1 Shot Attempt per 9 individual possessions)
Shot Attempt Assist Per Possession: .07 (~1 Shot Att Ast per 14 individual possessions)

Conor Sheary led the way in shot attempts per possession at .35, which was more than 3 times the team average of .11 shot attempts per individual possession.  Crosby and Cullen had the highst shot attempt assist rates at .20 each.  So 20% of Crosby’s individual possessions at 5v5, resulted in him passing the puck to a teammate who then recored a shot attempt.

Screen Shot 2017-06-04 at 10.09.22 PM

By looking at rebound%’s, one can see how possession spells can continue in the offensive zone, as well as relieving defensive zone pressure.  In game 3, the average offensive rebound% at 5v5 was 13.9%, and the average defensive rebound% was 7.6%.  As a defensemen, Olli Maatta grabbing 25% of the offensive rebounds available at 5v5 is very impressive.

The average offensive rebound% for players with a corsi% of 50% and above was 15.2%. The average defensive rebound% at 5v5 was 9.5%.  For players with a sub 50% corsi% at 5v5, the average offensive rebound% was 12.5%, and the average defensive rebound% was 6.5%.

Screen Shot 2017-06-04 at 10.46.58 PM

The chart below shows each players passes completed to a zone per each invidivudal possession at 5v5.  For example, Crosby had 30 5v5 individual possessions, of those 30, he completed 10 passes in the offensive zone (.33), 1 pass completed in the neutral zone (.03), and 2 passes completed in the defensive zone (.13).

This provides a better visual as to how which players/lines were able to sustain longer attacks in Nashville’s end.  The Kessel-Malkin-Kunitz line struggled significantly to establish themselves in the offensive end.

Ron Hainsey had the highest defensive zone passes completed per possession at .29 (2.9 defensive zone passes completed per every 10 individual possessions).  Hainsey had 31 possessions for the game, so 9 of those 31 possessions ended with him completing a pass in the defensive zone.

Pittsburgh’s Defensemen; Dumoulin, Schultz, Hainsey, Daley, and Cole spent a significant portion of their individual possessions making passes in the defensive zone.  Maatta was the lone bright spot on the Penguins blue line in game 3.  Maatta had the highest Offensive Zone Passes completed per possession at .18 (so 18% of Maatta’s individual possessions were comprised of passes completed in the offensive zone at 5v5).

Screen Shot 2017-06-04 at 11.08.10 PM

Amongst other things, I also document steals by each player and the zone they occur in.  Players who may turn the puck over in any zone, or fail to get a rebound, may try and makeup for that by getting a steal.  Below is the raw count for each player at 5v5 during game 3.

Screen Shot 2017-06-04 at 11.30.04 PM

Written by RReed

June 4, 2017 at 11:35 PM

Posted in Uncategorized

WSH@PIT Game 6: Microstats

leave a comment »

Pittsburgh at Washington Game 6:  2017 NHL Playoffs
Individual Possession Analysis
Washington 5, Pittsburgh 2

The data presented below is the results of my Individual Possession Efficiency tracking.  Evaluating players on a per possession level at even strength (5v5).  I document how each individual possession ends for every player.  All corsi data was retrieved from Corsica.Hockey.

The chart below shows Turnovers Per Possession, by zone.  Bryan Rust led the team in corsi% at 5v5 with a 61.5% corsi%.  Rust’s overall turnover%(turnovers per possession) was 35%, which is higher than the team average of 30.6% for game 6.  These are very high turnover rates for the Penguins.  Based on my tracking in the past their average turnover% was ~23%.  So Washington is forcing ~8 more turnovers per 100 individual possessions over Pittsburgh’s normal average.

The Malkin line of Rust-Malkin-Guentzel led the team in neutral zone steals at 6.  The other 3 lines had 3 neutral zone steals combined.  The Malkin line was able to makeup for their offensive zone inefficiencies by regaining possession in the neutral zone.

Screen Shot 2017-05-10 at 7.35.32 AM

Kunitz and Kuhnhackl had two of the 3 highest overall turnover rates at 40% and 73% respectively.  Kuhnhackl had the highest offensive zone turnover rate at 36%, and the highest neutral zone turnover rate at 18%.

Although Sheary played on Pittsburgh’s top line with Crosby and Hornqvist, his Dump in turnover rate of 19% is far too high.  That’s 5 failed dump in for Sheary on 26 possessions.  That should never happen when playing on a line with Crosby.

The chart below shows shot attempts per individual possession and shot attempt assist per individual possession at 5v5.  Average shot attempts per possession for game 6 were .078 (7.8 shot attempts per every 100 individual possessions).  Average shot attempt assist per individual possession in game 6 were .049 (4.9 shot attempt assist per every 100 individual possessions).

Rust led the team in shot attempt assist% at 12%, where Guentzel was 2nd on the team in shot attempts per possession at 14%.

Screen Shot 2017-05-10 at 7.42.43 AM

The rebound analysis below shows that the Penguins had issues with defensive zone rebounding.  The team average for game 6 was 4.3%, where past tracking shows the average individual defensive zone rebound rate for Pittsburgh to be ~8.3%.  Pittsburgh giving up almost double the defensive rebounds per Washington shot attempt.

What stands out is the Crosby-Sheary-Hornqvist line did not grab one offensive rebound at 5v5 during the game.  This is usually a strong point for Crosby’s line, as offensive rebounding helps to maintain offensive zone pressure.

Screen Shot 2017-05-10 at 7.54.23 AM

Lots of red in the defensive zone passes completed per possession.  As Washington had Pittsburgh hemmed in their own zone quite often.  In Game 1 Pittsburgh’s average defensive zone passes completed per possession was .19 (19%), in Game 6 it was .18 (18%).

Based off of last year’s playoff run for Pittsburgh, their average defensive zone passes completed per possession average was ~13.4%.  The neutral zone steals from the Malkin line helped reduce their defensive zone passes completed %’s relative to the rest of the team.

Screen Shot 2017-05-10 at 8.04.13 AM

Written by RReed

May 10, 2017 at 8:17 AM

Posted in Uncategorized

PIT@WSH Game 1: Microstats

leave a comment »

Pittsburgh at Washington Game 1:  2017 NHL Playoffs
Individual Possession Analysis
Pittsburgh 3, Washington 2

The data presented below is the results of my Individual Possession Efficiency tracking.  Evaluating players on a per possession level at even strength (5v5).  I document how each individual possession ends for every player.  All corsi data was retrieved from Corsica.Hockey.

The chart below shows Turnovers Per Possession, by zone.  Hornqvist led the Penguins in 5v5 corsi% at 48%.  He also had the lowest turnovers per possession at .11 (4 turnovers on 35 possessions= .114).  Average turnover rate for the Penguins in game 1 was 23%.  Conor Sheary had the lowest corsi% at 5v5 at 4.3%.  Sheary posted the highest turnover rate at .67 turnovers per possession.  Overall Sheary had 9 possessions at 5v5, 4 resulted in a turnover in the offensive zone (4/9 = .44), and 2 resulted in a turnover in the neutral zone (2/9 = .22).

Turnover Rates Per line at 5v5:
Guentzel-Crosby-Hornqvist:  17.8%
Rust-Malkin-Kessel:  33.8%
Wilson-Bonino-Sheary:  29.2%
Kunitz-Cullen-Kuhnhackl:  31.5%

Screen Shot 2017-04-29 at 11.53.40 AM

Clearance Rates Per Line at 5v5:
Guentzel-Crosby-Hornqvist:  11.1%
Rust-Malkin-Kessel:  6.8%
Wilson-Bonino-Sheary:  16.7%
Kunitz-Cullen-Kuhnhackl:  13%

Now I track clearances separately as they do not count towards turnovers.  Simply because I want to differentiate player intent.  Clearances are part of the possession totals for each player, but not included in the turnover total.  Looking at clearances, gives you a good idea as to which players were under duress at times.

Turnover rates per defensive pair:
Schultz-Cole:  13.2%
Dumoulin-Hainsey:  16.9%
Maatta-Daley:  20%

Defensive Zone turnover rates per defensive pair:
Schultz-Cole:  7.5%
Dumoulin-Hainsey:  9.2%
Maatta-Daley:  3.6%

The chart below presents shot attempts and shot attempt assists per possession.  League average per my tracking thus far for shot attempts per possession is .11 (1.1 shot attempts per every 10 individual possessions).  League average Shot Attempt Assists Per possession are .066 (1 Shot attempts assist per every 15 individual possessions).

Screen Shot 2017-04-29 at 12.29.42 PM

The chart below shows passes completed per possession by the zone they were completed in.  For example, Hornqvist completed 8 passes in the offensive zone (8/35 = .23 Offensive Zone Passes Completed Per Possession).  Judging by the high defensive zone passes completed per possession for Pittsburgh’s defensemen, you can see how the ice was tilted quite a bit in Washington’s favor.

Screen Shot 2017-04-29 at 1.05.46 PM

The chart below shows offensive and defensive rebound% at 5v5.  Team average offensive rebound% in game 1 was 8.7%.  Average defensive rebound% was 5.2%.

For reference, in game 2 vs Columbus the average offensive rebound% was 10.3%. Average defensive rebound% was 6.5%.
In game 4 vs Columbus, average offensive rebound% was 7.8%. Average defensive rebound% was 8%.

Combined Offensive Rebound%’s per line:
Guentzel-Crosby-Hornqvist:  47.3%
Rust-Malkin-Kessel:  37.5%
Wilson-Bonino-Sheary:  0%
Kunitz-Cullen-Kuhnhackl:  16.7%

Combined Defensive Rebound%’s per line:
Guentzel-Crosby-Hornqvist:  0%
Rust-Malkin-Kessel:  16.1%
Wilson-Bonino-Sheary:  11.1%
Kunitz-Cullen-Kuhnhackl:  24.3%

Screen Shot 2017-04-29 at 12.57.33 PM

Written by RReed

April 29, 2017 at 1:20 PM

Posted in Uncategorized

PIT@CBJ Game 4: Microstats

leave a comment »

Pittsburgh @ Columbus: Game 4  2017 NHL Playoffs
Individual Possession Analysis
Columbus 5, Pittsburgh 4

The data below are the results of my individual possession efficiency tracking, where players are evaluated on a per possession level.  I track each individual turnover per possession by the zone it occurred in.  The efficiency rates are then compared to their Corsi%, which I retrieved from Corsica.Hockey.

The chart below shows Matt Cullen as having the lowest turnover rate per possession across the board.  Cullen’s overall turnover rate was .11 (1.1 turnovers per every 10 individual possessions).  League average per my tracking to date is ~.25 turnovers per possession.

Dump ins as a means of entering the zone did not appear to work out well for Hornqvist and Bonino.  In game 4, Pittsburgh’s average dump in turnover rate was .05 (1 dump in turnover per every 20 individual possessions).  Hornqvist and Bonino’s rate were ~3 times higher than the team average.  The dump in success rate for Pittsburgh in game 4 was 16% (4 dump ins recovered on 25 attempts).

Now I track clears and dump ins separately, so that if a player dumps it in then goes for a change, I document that as a clearance because there was no intent.

Screen Shot 2017-04-20 at 7.23.25 AM

Crosby and Guentzel had low turnover rates as compared to the rest of the team, at .17 and .21 respectively.  Guentzel also had the 3rd highest passing accuray% at 85.7%.  Sheary’s turnover rate was .42 (4.2 turnovers per 10 possessions).  Now this is extremely high, but Sheary makes up for it with well above average Offensive and Defensive Rebound%’s.

The team average in Game 4 for offensive rebound% was 7.8%.  The team average for defensive rebound% was 8%.  Sheary and Guentzel both well above the team average in rebounding, were able to sustain longer possession spells in Columbus’ zone.

Screen Shot 2017-04-20 at 7.57.27 AM

The chart below shows each players’ shot attempts (SOG, Miss, Blocked) per each individual possession, as well as their passes to setup players for a shot attempt (Shot attempt assist).  The league average per my tracking thus far for shot attempts per possession is .11, and .066 for shot attempt assist per possession.

Crosby had a shot attempt assist on 22% of his individual possessions, which is ~3 times higher than the league average.  Guentzel’s shot attempts per possession(.21) were ~2 times the league average.

Screen Shot 2017-04-20 at 7.54.35 AM

The chart below shows each players passes completed per possession by zone.  The Crosby-Sheary-Guentzel line had some success in the offensive zone.  Crosby completed an Offensive zone pass on almost 40% of his individual possessions.  The team average for this game was .17 (1.7 out of every 10 individual possessions).

The increase in offensive zone time relative to their teammates, as well as above average offensive rebounding, contributed to the Sid and the Kid’s success in this game.   Schultz’s high defensive zone passes per possession jump out as very high, but this is more due to him being patient in his own zone.  If the pass through the neutral zone isn’t there, Schultz will look to pass to his defense partner.  This delays the breakout to wait for another passing option out of the zone to potentially open up.

Screen Shot 2017-04-20 at 8.02.23 AM

Written by RReed

April 20, 2017 at 8:14 AM

Posted in Uncategorized

CBJ@PIT Game 2: Microstats

with one comment

Columbus @ Pittsburgh: Game 2  2017 NHL Playoffs
Individual Possession Analysis
Pittsburgh 4, Columbus 1

Below, are the results of my individual possession efficiency tracking, where I document what each player does per each time they have an opportunity to make a controlled play (Possession) at even strength (5v5).  All Corsi%’s were retrieved from corsica.hockey.

The chart below presents each players’ Turnovers Per Individual Possession at 5v5, broken down by the zone they occurred in.  For example, Phil Kessel had 18 individual possessions at 5v5 in game 2, and committed .39 Turnovers Per Possession.  Phil had 18 Individual Possessions at 5v5 and committed 7 turnovers. (7/18 =.39 Turnovers Per Possession)

The average turnover rate for the team was .23 turnovers per possession.  Based on my tracking to date, the league average individual turnover rate is ~.25 (2.5 turnovers per every 10 individual possessions)  Matt Cullen had the highest overall individual turnover rate at .50, so half of Cullen’s individual possessions resulted in a turnover.  Hornqvist at .42, and Hainsey at .35 turnovers per possession.

Screen Shot 2017-04-16 at 9.01.19 AM

The passing analysis below shows each players’ passes completed per possession at 5v5, and the zone which it occurred in.  Schultz is really good.  While considering his high individual possession count (29), Schultz was very much involved.  Schultz completed 17 of 18 passing attempts for a Passing Accuracy% of 94.4%.  Kessel and Hainsey posted the two lowest passing accuracy%’s of 30% and 50% respectfully.  More reasons as to why Kessel should shoot even more.

Screen Shot 2017-04-16 at 9.12.40 AM

The data below shows each players’ shot attempts (SOG, Miss, Attempt Blocked) per individual possession.  Shot attempt assist document each players pass completed in the offensive or neutral zone, where the receiving player attempts a SOG, Miss, or attempt blocked.

Per my tracking to date, the league average for Shot Attempts Per Possession is .11, where the league average for Shot Attempt Assist Per Possession is .066.  So ~11 out of every 100 individual possessions result in a shot attempt, and ~6.6 out of every 100 individual possessions result in an assist on a shot attempt. For example, Kessel had 18 possessions at 5v5, where 4 of them resulted in a shot attempt (.22).  Kessel’s 30% passing accuracy contributed to his 0.0 shot attempt assist per possession.

Screen Shot 2017-04-16 at 9.24.01 AM

The ability to obtain rebounds can be very important to maintaining pressure in the offensive zone, as well as relieving pressure in the defensive zone.  This can also highlight a player’s quickness, and their ability to anticipate the play.  The visual below shows each player’s Offensive and Defensive rebound%’s at 5v5.  Per my tracking to date, league average offensive rebound% is ~8.2%, where the defensive rebound% is ~8.8%.

Conor Sheary posted the highest offensive rebound% at 33.3%, meaning he literally grabbed 1/3 of the offensive rebounds available at 5v5 (4 rebounds/12 available = 33.3%).  This is a crucial aspect to sustaining long possession spells in the offensive zone.  Hainsey   and Dumoulin did not fair so well in the rebounding department.  Of the 22 defensive rebounds available at 5v5, they were not able to get one rebound between them.  This highlights a reason as to why Columbus was able to sustain an offensive zone pressure with the two of them on the ice.  This is also reflected in Hainsey’s 30.2% corsi% and Dumoulin’s 28.2% corsi%.  More rebounds for Columbus, mean more opportunities for shot attempts against.

Screen Shot 2017-04-16 at 9.52.28 AM

Written by RReed

April 16, 2017 at 10:07 AM

Posted in Uncategorized

2017 NHL Playoffs log5 Analysis

leave a comment »

2017 NHL Playoffs log5 Analysis
The log5 analysis was developed by Bill James. This method of evaluation is used to determine the predicted winning probability of one team over another. The input for the formula is derived from the final 2016-17 NHL regular season Possession Efficiency Ratings.

As you can see below, the log5 analysis measures the difficulty of each team’s path, taking into consideration potential match ups that could occur. For example, the result of the log5 analysis below is showing that Washington has a 62.4% chance of defeating Toronto in the first round and advancing to the 2nd Round.

Washington then would have a 33.2% chance of defeating Toronto in the first round and then defeating the winner of the Pittsburgh-Columbus series to advance to the Conference Finals.

From there, depending on who Washington would meet in the Conference Finals, they would have a 19.2% chance of winning the first 3 rounds and advancing to the Stanley Cup Finals.  Washington would then have an 11.3% chance to win the first 3 rounds and the Cup final.

Screen Shot 2017-04-10 at 9.40.25 PM

Written by RReed

April 10, 2017 at 9:41 PM

Posted in Uncategorized