Friday, November 8, 2013

The limitations of perception-based scouting

Using data-generated scouting to help quantify perceptions

Perception-based scouting is an important aspect of player-evaluation. There is no doubt that the only way to view a player's skill-set is to watch them play. Perception-based scouting sheds light on; (among other things) a player's speed, their strength, and their defensive-instincts.

That said, perception-based scouting can be limiting; even deceptive. As such, it requires a litmus test. That litmus test can be produced through data-generated scouting. Think of perception-based scouting as a way to establish what a player is capable of doing, while data-generated scouting tells us what the player is actually doing.

Montreal Canadiens winger Rene Bourque is an interesting example of where data-generated scouting can help quantify the perception of a specific player. The perception-based scouting report on Bourque would read along the lines of:
  • An above-average sized player
  • Good skater
  • A goal-scorers release

Granted, Bourque's scouting report has changed over the years. When he was racking up suspensions, you could add that he's a player that hits to hurt.

The data-generated scouting report could be used to either compliment the above report, or to contradict it. The important thing to remember though, is that there is no predisposition to do either. Data is data; it's not about bad, good, better or best.

Rather than attempt to communicate all of the data I've compiled for Rene Bourque over the years, I'll focus on relating this season's data to the above scouting report.

SIZE

The size of a player cannot be validated through data-generated scouting. That said, how a player uses their size can be evaluated.

A player must engage in events to use their size as an advantage. Rene Bourque has engaged in 3.55 events per-minute of even-strength ice-time this season. The average among Habs wingers in 4.05 events per-minute, while the team average is 4.29. Among wingers who have played for Montreal this season, only George Parros, Michael Blunden, Patrick Holland, and Daniel Briere have engaged in fewer events per-minute than Bourque.

Breaking those numbers down even further, we see that only Briere has engaged in fewer defensive-zone events per-minute than Bourque. However, in the offensive-zone Bourque has engaged in the sixth-most events per-minute among all Montreal players, and the fifth-most among wingers.


Another way for a player to use their size is by winning puck-battles. Data-generated scouting can quantify how a player uses their size by relating how they engage in puck-battles, how often they engage in puck-battles, and how successful are they when engaging in puck-battles. Puck-battles within this system are broken down in to two distinct events; removing possession from the opposition, and acquiring possession for your team.

Bourque engages in 0.230 offensive-zone puck-battles per-minute. Sixty-three percent of the time Bourque attempts to remove possession from the opposition in the offensive-zone he uses a stick-check. When attempting a stick-check, he is successful 48.6% of the time. Thirty-seven percent of the time, Bourque uses a body check; and is successful with 65% of those attempts.

Bourque engages in 0.122 defensive-zone puck-battles per-minute played. Eighty-nine percent of the time Bourque attempts to remove possession from the opposition in the defensive-zone he chooses to stick-check, and is successful with 57% of those attempts. The other 11% of the time Bourque attempts a body-check to remove-possession; successfully accomplishing that only 33% of the time.

Data-generated findings:
  • Bourque uses his size to engage in substantially more offensive-zone events than defensive-zone events; rendering him almost a non-factor in the defensive-zone.
  • Bourque is among the Habs more physically-active forwards in the offensive-zone
  • Bourque prefers to engage players with his stick rather than with his body; despite being substantially more successful with body-checks in the offensive-zone.


GOOD SKATER

Data-generated scouting can reflect a player's skating ability by relating how often they recover loose-pucks. A player defined as a good skater must use that speed to acquire puck-possession for his team. The only way to do that is to be the first on the puck.

Bourque has recovered 0.455 offensive-zone loose-pucks this season. Giving him the only eighth-most o-zone LPR's among Montreal wingers. In the defensive-zone, only Briere and Parros have recovered fewer loose-pucks than Bourque.

Good skaters can also use their speed to carry the puck through the neutral-zone. A way of quantifying this stat is to see how many puck-possession events a player engages in while positioned in the neutral-zone, and by viewing a player's tendency when in possession of the puck in the neutral-zone.








In terms of neutral-zone events while in possession of the puck; only Briere engages in fewer neutral-zone events with possession than Bourque. As we can see within this graph, Bourque is not a player who carries the puck through the neutral-zone. In fact, his tendencies point to a player focusing on keeping the game going north/south; dumping the puck into the offensive-zone 71.8% of the time he has possession in the neutral-zone.


Data-generated findings:
  • Bourque's speed does not allow him to recover more loose-pucks than the average winger
  • Bourque does not use his speed to carry the puck through the neutral-zone


A GOAL-SCORERS RELEASE

Anyone who has watched Bourque practice can speak to both the heaviness of his shot, and the quickness of his release. Data-generated scouting can take this perception and relate it to on-ice performance. A quick-release is a great asset, but only getting shots on net produces results.

Bourque uses his shot to attempt 0.259 shots per-minute of even-strength ice-time. The average amount of attempted shots per-minute among Montreal wingers is 0.274. Only 22.6% of his attempted shots are blocked by opposition defenders. This speaks well of Bourque's quick release, as only 3 other Habs wingers have had a lower percentage of their attempted shots blocked. A lower percentage of blocked shots also point to a player taking shots from in-close to the net.

Bourque has missed the net with 30.6% of his attempted shots. Among Habs forwards with significant ice-time, Bourque misses the net with the highest-percentage of his attempted shots. In total, Bourque is able to get 0.121 shots on net per-minute of ice-time. This place Bourque 10th among Montreal forward in shots on net PMP.

On the powerplay, Bourque has attempted 0.177 shots per-minute played, while the average among Montreal wingers is 0.326. He has been able to get 0.118 shots-through to the net per-minute; compared to an average among all Montreal wingers of 0.203.

Data-generated findings:
  • Low percentage of attempted shots blocked by the opposition point to a player with a quick-release and a willingness to shoot from close-in.
  • High-percentage of shots that miss the net demonstrate a lack of control and/or a propensity to shoot for the corners
  • Bourque shoots less-often, and hits the net less often than the average Montreal winger

This post is not meant to denigrate perception-based scouting. Nor is it meant to position one form of scouting over the other. Instead, it is meant as a way of showing how both systems can be used in tandem; reinforcing some perceptions, while forcing us to ask more questions about others.




No comments:

Post a Comment