Guest Post: Kyle Lowry May Have Exposed Cleveland’s Achilles Heel

May 21, 2016; Toronto, Ontario, CAN; Cleveland Cavaliers guard Kyrie Irving (2) takes a shot over the outstretched arm of Toronto Raptors guard Kyle Lowry (7) during the second half of game three of the Eastern conference finals of the NBA Playoffs at Air Canada Centre. Mandatory Credit: Dan Hamilton-USA TODAY Sports
May 21, 2016; Toronto, Ontario, CAN; Cleveland Cavaliers guard Kyrie Irving (2) takes a shot over the outstretched arm of Toronto Raptors guard Kyle Lowry (7) during the second half of game three of the Eastern conference finals of the NBA Playoffs at Air Canada Centre. Mandatory Credit: Dan Hamilton-USA TODAY Sports /
facebooktwitterreddit
May 23, 2016; Toronto, Ontario, CAN; Toronto Raptors guard Kyle Lowry (7) dribbles the ball past Cleveland Cavaliers guard Kyrie Irving (2) in game four of the Eastern conference finals of the NBA Playoffs at Air Canada Centre. Mandatory Credit: Dan Hamilton-USA TODAY Sports
May 23, 2016; Toronto, Ontario, CAN; Toronto Raptors guard Kyle Lowry (7) dribbles the ball past Cleveland Cavaliers guard Kyrie Irving (2) in game four of the Eastern conference finals of the NBA Playoffs at Air Canada Centre. Mandatory Credit: Dan Hamilton-USA TODAY Sports /

On the verge of the NBA Finals, we bring you this guest post from Chistopher Pickard. Chris is originally from Phoenix, AZ, and is a diehard Phoenix Suns fan who is currently trapped in the confines of the Bay Area as a first year graduate student at Stanford University. As a former 4-year varsity swimmer, academic All-American and team captain for the Cardinal, Chris is passionate about finding the best ways to assemble teams intelligently and appreciates the finer things in life such as a good pick-and-roll defense. He is still waiting for David Stern to apologize for the 2007 Western Conference Semifinals. 


After the first two games of the 2016 Eastern Conference Finals, the Cleveland Cavaliers found themselves in a similar position as in their first two playoff series, ahead 2 – 0 with the early choke hold now placed on the second-seeded Toronto Raptors. Unlike the first two playoff series, the Cavaliers returned home after Games 3 and 4 on the road tied at two games apiece needing to push through a fifth and sixth game to close out the Raptors last Friday evening. Contrary to two dominant early round sweeps of Detroit and Atlanta, the Cavaliers struggled to demonstrate a similar strength against Toronto.

In large part, the struggles of the Cavaliers during coincided with more potent play from Toronto guards Kyle Lowry and DeMar DeRozan, most notably when Lowry had an increase in productivity. Until his 35-point outburst in Game 6, Lowry saw a substantial increase of 16.5 PPG between Raptors’ defeats (11 PPG) and victories (27.5 PPG) during the series. Lowry’s backcourt teammate DeRozan also saw a smaller increase from 18 PPG to 32 PPG between defeats and victories, but showed more overall consistency than Lowry during the ECF series. Despite DeRozan’s relative consistency, the Raptors’ success in the 2016 Playoffs was tied closely to the performance of Lowry. In their playoff victories, Lowry averaged 23 PPG while in their defeats he averaged 13 PPG. As a result, the success the Raptors had against the Cavaliers in correlation to the productivity of their point guard raises an important question: Was Lowry’s success a matter of circumstance, or is there a more significant underlying reason that the Cavaliers should be concerned about moving forward into the NBA Finals?

This question can be answered through a shot-distance-specific regularized adjusted plus-minus (RAPM) model that measures the likelihood that a given event occurs during a possession while a player is on the court and the subsequent impact on the outcome of the points expected from that event. Similar to the popular “score margin” RAPM model[2. One of the primary inputs in ESPN’s “Real Plus/Minus” metric.], the model used in this study considers a time frame when a player is on the court, but instead of measuring changes in point differentials, it estimates the likelihood that a given event occurs while a player is on offense or defense. Adding in he resultant points expected from each event, these expected point values are adjusted for each player’s percentage time on the court during a game. The model thus ultimately returns an expected point value per possession for a given player that has four average teammates playing against five average opponents[1. During the 2015 – 2016 NBA season the expected points per possession was 1.094 points.]. The model can be further refined to measure the expected value points per possession given a known shot distance attempt occurs. The model provides a comprehensive assessment of the efficiency a given player’s team is per possession in both offensive or defensive point expectancy while he is on the court.

Pickard1
Pickard1 /

Through the model, it can be better understood where a Kyle Lowry-led offensive team excels on the court, as shown in the purple columns above. The proverbial “eye test” of Lowry’s offensive play suggests he is an exceptional offensive talent, and this assessment is supported by the model. Compared to the league average, a Lowry-led offensive team is one of the more productive offenses in the league, specifically when shots are attempted within 12 feet from the basket. As with any RAPM-style analysis, this doesn’t mean Lowry is specifically responsible for these shot attempts, but while he is on the court these shot attempts returns the highest yield in points scored per possession. Lowry’s offensive style of play suggests he has a greater than coincidental impact on the results, but when Lowry offensive teams can produce short to intermediate shot attempts, the Raptors put themselves in the best position for offensive success.

Within the context of the 2016 NBA Eastern Conference Finals, Lowry’s most common defensive matchup by position was the Cavaliers’ Kyrie Irving. Once again, using the shot-distance-specific regularized adjusted plus-minus model the defensive impact of an Irving led defensive team can be measured and is shown in the gold bars. Compared to league average, a defense featuring Irving performed at below-average levels in terms of expected points allowed per possession; allowing more points per possession than what is typically expected specifically in short to mid-range shot attempts. As described by the model, the most glaring weakness for an Irving-led defensive team stems from shot attempts taken within 12 feet from the basket, most notably within three feet, indicating that Irving led defenses are most exposed and least effective when shots are taken from these specific ranges.

The model highlights a glaring and significant mismatch at the point guard position in this year’s version of the Eastern Conference Finals swayed heavily in favor of the Raptors. As shown above a Lowry-led offensive teams excels in terms of expected points scored per possession where Irving-led defensive teams struggle the most. It’s the proverbial mismatch that teams look for in a long seven game series and could have been a small but critical element that helped propel a team like Toronto towards an upset over, or at least competitiveness with, the top-seeded Cavaliers. The model does not claim that an exclusive Lowry-Irving matchup is a mismatch, but given an average set of teammates, an already potent Lowry led offensive team should not be expected to be limited by any substantial margin by an Irving led defensive team.

Pickard2
Pickard2 /

Fortunately for the Cavaliers, they have two viable alternative defensive options at point guard in Matthew Dellavedova and Iman Shumpert to minimize the mismatch. The impact of a Dellavedova or Shumpert defensive on a Lowry led offensive team can be estimated as shown above. The bars indicate how many points expected per possession a Lowry-led offensive team would have given the opponent is an Irving-, Dellavedova- or Shumpert-led defensive team. In comparison to the league average, Dellavedova- or Shumpert-led defenses bring the point expectancy of a Lowry-led offense closer to the league average, specifically in Lowry’s most effective shot distances between 0 – 12 feet of the basket. Where Irving struggles greatly, minimally impacting the Lowry led offense, Dellavedova and Shumpert are valued as above average impact defenders and their effect is noticeable on Lowry. As a result, a switch between Irving or either a Dellavedova/Shumpert defense can minimize the impact of point expectancy for a Lowry led offense and provide the best answer to Irving’s weaknesses for the Cavaliers.

The model’s suggested impact of a Dellavedova/Shumpert defense in exchange for an Irving-led defense manifested itself during the fourth quarter of Game 4. For nearly the first six minutes of the quarter, Irving sat on the bench until the 6:39 mark of the quarter while Shumpert and Dellavedova remained on the court. During this period, the Cavaliers closed the initial quarter deficit from nine points to a one-point margin. While Irving sat, Lowry was limited to 3 points on two shots, 2 fouls drawn and a turnover. With Irving back in the game and Shumpert subbing out, Lowry accumulated 6 points on three shots, 3 fouls drawn and a turnover. Comparing over near equal stretches of time, Lowry’s effectiveness doubled in points scored and aggression as he increased the number of fouls drawn while Irving was on the court compared to when Irving was on the bench. This is not surprising as the the model suggests that Lowry should be about twice as effective when Irving is on the court as opposed to Dellavedova or Shumpert and speaks to the relative strengths between Lowry on offense and weaknesses of Irving on defense as described by the model.

The defensive liability that Irving brings to the court was also observed by  Zach Lowe, who noted how Toronto “ruthlessly hammered” the Irving/Kevin Love pick-and-roll defense with great success. As Lowe points out, on possessions where the opponent creates a situation with Irving and Love defend a pick-and-roll Cleveland gave up 1.207 points per possession, significantly higher than the league average and ranked worse than 119 qualifying defensive pairs in the league this season. This data provides a more definitive reason as to why Irving led defenses struggle greatly when shots are attempted within 12 feet of the basket. Poor pick-and-roll defense allows a dynamic playmaker such as Lowry to position himself or teammates in the short to mid-range shot attempt ranges, which creates multiple scoring opportunities for Lowry or the screener rolling back to the basket. Where the model lacks the ability to account for specific actions by individual players on the court, its findings are revealing themselves on the court.

While the Cavs have some effective options to counter a mismatch presented by a player such as Lowry, but the correlation between breakout performances by Lowry in Games 3 and 4 of the series that resulted in Raptors’ victories should not be viewed as an aberration, but an underlying core problem that should have the Cavaliers worried moving forward. Looking ahead to the NBA Finals, Golden State’s Stephen Curry is just this sort of dynamic playmaker, even more effective offensively than Lowry, especially in Irving’s weakest defensive ranges. Finding ways to avoid this direct matchup of Irving guarding Curry might be critical to Cleveland’s success. It could require lineup alterations, or more high trapping of Curry’s dribble[2. Which the Cavs did effectively in Game Five against Toronto, not to mention the first several games of last year’s NBA Finals.].

One thing is certain: If long term solutions are not found, it should be expected that the Cavaliers will be susceptible to breakout performances by opposing team point guards moving forward. This weakness could ultimately prove to be an Achilles heel to their ambitions of winning the franchise’s first championship.