Why is it so hard to evaluate individual defensive play At the individual level theres only so much a player can do to influence play on their own Players have to act as a unit gt Results are heavily influenced by teammates ID: 773398
Download Presentation The PPT/PDF document "Why is it so hard to evaluate individual..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Why is it so hard to evaluate individual defensive play? At the individual level there’s only so much a player can do to influence play on their own Players have to act as a unit -> Results are heavily influenced by teammates Player matchups can be more varies -> More influence of opponents Defensive systems can have a large impact on results -> Hard to compare across teams Granular data doesn’t exist for defence For offence we know the shooter, who had the assist on goals Goalies are voodoo and hate statistical analysts
What types of passes/shots are we interested in at the individual level? Metric Sh % Why do we care? Skill(s) Measured Entry Assists Against/60 7.2% Zone entry work has shown how important the type of entry is Zone Entry Prevention, Gap Control, etc. OZ Center Lane SAA/60 13.9% Expect this to be a skill – how well players control the middle of the ice Forcing play to the outside Behind the Net SAA/60 11.7% Higher danger shots against – expect that defencemen should be able to influence this by winning battles down low Winning puck battles in corners, maintaining position Royal Road SAA/60 27.6% Very high shooting percentage – could be influenced by ability to remain structured defensively or prevent rush attempts Preventing cross-ice passes Point Shots/60 2.6% Relatively low danger shots – could be influenced by keeping players to the outside Covering the point, keeping play to the outside
Defencemen - What’s repeatable? Split-Half Correlations (Does data in one half predict data in the other?) Metric Split-Half R^2 Shots Assists Against/600.54Entry Assists Against/600.53OZ Center Lane SAA/600.32Behind the Net SAA/600.39Royal Road SAA/600.01Point Shots/600.17 Which metrics seem to be most repeatable? Entry Assists Against/60In line with Eric Tulsky’s findingsOZ Center Lane SAA/60D need to control middle of iceBehind the Net SAA/60D need to control area behind net Players with >= 250 minutes in each half (N=48)
Defencemen – Are these skills related? Correlations Between Defensive Metrics A fair amount of correlation exists within defensive metrics Behind-the-net and OZ Centre-Line Passes are most correlated Both measure play in D-Zone Correlation isn’t perfect There can still be players who excel at one area but not the othersUseful for deployment within a teamPlayers with >= 500 total minutes (N=51)EntryOZ CentreBTNEntry10.660.55OZ Centre 0.661 0.78BTN0.550.781
Example: Randomly Selected Defencemen Metric P.K. Subban Shea WeberTOI338.6352.1Rate StatsEntry Assists Against/6015.4 16.7OZ Center Lane SAA/609.610.2Behind the Net SAA/60 4.1 2.9 Relative Stats Entry Assists Against/60 Rel 3.0 2.1 OZ Center Lane SAA/60 Rel 1.0 1.0 Behind the Net SAA/60 Rel 0.4 0.5
Forwards - Do they even matter on D? Split-Half Correlations (Does data in one half predict data in the other?) Metric Split-Half R^2 Shots Assists Against/600.36Entry Assists Against/600.48OZ Center Lane SAA/600.31Behind the Net SAA/600.19Royal Road SAA/600.04Point Shots/600.20 Forwards appear to have less control over defensive results Lower repeatability across each metricBiggest influence on Entry Assists Against/60Unfortunately, you actually do need to backcheckSome impact on Center Lane Shots Players with >= 250 minutes in each half (N=47)
Example: Buffalo Sabres Forwards – 2015/16 Player Name GP TOI OZ Center Lane SAA/60 Entry Assists Against/60O'REILLY, RYAN 18 275.811.115.0 GIRGENSONS, ZEMGUS 19 253.6 9.0 14.9 LARSSON, JOHAN 17 221.8 7.8 14.6 REINHART, SAM 19 252.3 9.8 14.3 MOULSON, MATT 20208.59.513.5EICHEL, JACK20299.37.213.0GIONTA, BRIAN19265.06.312.9MCGINN, JAMIE22252.16.412.9KANE, EVANDER16251.88.811.9 Noted Two-Way Center Ryan O’Reilly had the most Entry Assists Against/60 Noted Defensive Liability Evander Kane had the fewest Entry Assists Against/60