Tuesday, January 29, 2013

The Scoring "Maturation" of Carmelo Anthony

Having served a one-game suspension for confronting Kevin Garnett after a home loss to the Celtics on January 7, Carmelo Anthony and the Knicks were getting a lot of attention on how they would respond in their second match-up of the season against Boston last Thursday. The Knicks ended up winning their first game in Boston since 2006.  Questions remain regarding how far 'Melo can take his Knicks in the postseason and whether he should be a part of serious MVP discussions (his teams have advanced out of the first round only once out of nine playoff appearances).

Yahoo Sports’ Adrian Wojnarowski wrote a column shortly after the Knicks’ victory in Boston, addressing Anthony’s new-found composure and maturity, commenting that “no one and nothing should be able to deter him on his mission.”  Although Anthony's style of play is relatively similar with past seasons, his maturation and improvement are reflected in his statistical production this year.

Averaging 7.3 points more per game compared with last year and 4.9 points since becoming a Knick, Anthony has returned to the abundant scoring ways we saw from him in seasons with the Nuggets. His statistics this season would be even further from his Knicks career average had he not missed seven games earlier in the season due to injury and the suspension (Anthony has ten 34+ point games so far this season). Statistics unrelated to individual scoring (rebounds, assists, steals, blocks, turnovers, personal fouls) have hardly exhibited any variance between this season and his Knicks career, but Anthony’s stat line shows considerable differences in points per game average and field goal attempts per game. 

With the Vantage data set, we can analyze Anthony’s scoring tendencies to show how Anthony has been creating his shots this season compared to last. Diving into Anthony’s shot sequence data, we can see that there isn’t much variation in Anthony’s pre-acquisition characteristics (i.e., what Anthony does before he gets the ball). Comparing the frequency of Anthony’s pre-acquisition moves that lead to field goal attempts from last season and this season, we see that the majority of his field goal attempts still come from spot up or iso situations, and although the frequency of post ups has increased slightly this year (13.6% to 16.8%), Anthony’s change in distribution of pre-acquisition move frequencies is relatively similar to last season.

Although it’s not a dramatic change, Anthony has shown a slightly more balanced approach to his play away from the ball. Looking at Carmelo’s FG% when performing his five most frequent pre-acquisition moves, we see that his effectiveness has remained similar in spot up/iso and transition situations. However, there have been significant changes in his effectiveness in post up and cut/flash moves, moves that require speed and strength against their primary defender. Anthony’s 1 off ball screen drop off is representative of the decrease in frequency of those moves from last season to this season (last year, it was Anthony’s third most frequent pre-acquisition move, this year it is his fifth).
One refined area in the Knicks’ 2012-2013 campaign has been the frequency and success of their three-point shooting. The Knicks lead the league in average three-point attempts per game and rank fifth in the NBA in three-point percentage. Looking at only three-point field goals in the Vantage data set reveals no significant change in pre-acquisition characteristics in Anthony’s court position before a three-point field goal attempt (about 43% of the time his pre-acquisition position is beyond the three-point arch). This suggests that Anthony isn’t necessarily making more of an effort to attempt three-pointers.

Although the data suggest that Anthony hasn’t significantly restyled his game away from the ball this season, the most impressive insight is that his statistical performance this year has improved despite the fact that he has faced significantly more defensive pressure this season compared to last.

Anthony is receiving significantly more double and triple coverage than last year, suggesting that defenses are committing to limit his production without success. Looking at the shot defense that Anthony is facing as well, we see that individual player contested defense (defender within three feet of shooter and with a hand up) is more aggressive this season, with the percentage of shots contested against Anthony increasing from 49.61% to 57.65%. Even though his contested shots faced increased, he has managed to improve his contested efficiency from 42.3% to 49% shooting.

There are other factors that contribute to the rise in field goal attempts and ultimately the Knicks’ success. Despite ranking 28th in the league in assists and 24th  in total rebounds, the Knicks rank in the top 10 in points per game (100.3) and effective field goal percentage (50.5%) while allowing the fewest turnovers per game.  On top of that, the Knicks average the oldest roster out of any team in the NBA. The mission is on course and the improvement is evident by their record, and when it comes to 'Melo, the stats show that despite their best efforts, teams still are not “deterring him on his mission.” His ability to produce at a higher rate this season under more aggressive defensive schemes is reason to believe the MVP candidacy talk for the Knicks' star forward.  

The New New Kobe

The new Kobe was going to be a full-time defensive stopper with playoff-level intensity.  That was last week.  This week, the new new Kobe is a facilitator extraordinaire dishing out the spotlight as much as he can.  Kobe’s ability to be a facilitator has had some good initial returns with two wins in a row, but the question is again whether it can last.  Using stats in this case is a little difficult because passing is just not something he did very much.  Raw stats and pace-adjusted stats aren’t great at telling you how good a player can do something when that player doesn’t do that something very often.  Using the Vantage data set, however, we can dig deeper into Kobe’s passing past and see if there is anything interesting in those times he actually did pass compared with some other well-known NBA facilitators.  To do so, we’ll look at various ratios composed of assists, crucial passes (i.e., passes leading to an assist -- or the “hockey assist”), passes to missed open shots (no sense in punishing the passer for hitting a teammate who misses an open shot), passes to shooting fouls (many times just as good as an assist -- though maybe we should rethink this one in the context of certain Lakers), turnovers on passes, and passes deflected.

The Good News: Kobe's Assist+ Ratio is way below the other three players', which indicates that his assist numbers on the official stat sheet are under-representing his true facilitation value.  Driving this low ratio is the large number of open shots that his teammates are missing when he gets them the ball.  His ratio of Assists to Open Shots is half that of Rondo and even less than half compared with Paul.  This suggests that if his teammates start making more open shots (and there should be some regression to the mean here), Kobe will see an uptick in assists.  Finally, Kobe also leads the selected bunch in indirect facilitation, which is the touchstone of the crucial pass -- not surprising given Kobe's ability to track multiple defenders, which creates space and leads to easy buckets two passes away.

The Bad News: Unfortunately, Kobe also has the lowest ratio of Assist+ (i.e., all of the direct and indirect passing stats combined) to turnovers, and by a large margin compared with the elite comparison group (Nash is lagging a bit in this category too, but remember our sample doesn't include numbers from Nash's prime).  Both Kobe and Nash also have poor Deflected Pass%'s, which means that their weaker Assist+ to Turnover ratios are probably due more to bad passing than to bad luck. 

The biggest question regarding the new defensive Kobe was whether he could keep it up physically.  Regarding the new new facilitating Kobe, the stats suggest that Kobe may be able to be an effective facilitator, especially if his teammates can start knocking down open shots and he can cut down on his turnovers by making cleaner passes.  The question is whether he can keep it up psychologically. 

Wednesday, January 23, 2013

Selecting the Western Conference All-Star Frontcourt Reserves

Every year, the selection of All-Star Game reserves leads to debate. This year, choosing between a crowded group of Western Conference frontcourt players has caused much consternation. Luckily, Vantage's data can shed light on some of the hidden factors for these players. Let's take a look at how the All-Star candidates stack up in areas that aren't measured in the box score: one-on-one defense, help defense, pick-and-roll defense, and rebounding effort.

Shot Contesting
The first aspect of defense that we will examine is how well these players defend their man when he’s shooting. Prior research has confirmed that the stronger the defender contests the shot, such as getting a hand up in a shooter's face, the lower odds that shot will go in. The table below displays the players’ field goal percentage allowed and Contest+ percentage. Contest+ indicates how often the defender contests (being within three feet of the shooter with a hand up), alters, or blocks the shot.

Unfortunately for Marc Gasol, he has bucked the trend of strong defense leading to poor shooting. Among this group of players, Gasol is playing the strongest defense against shooters, contesting two-thirds of his opponents’ shots, yet they still are managing to shoot 50 percent against him. If Gasol continues this level of defense, expect that field goal percentage to drop. 

Things make a little more sense among the poor defenders, where LaMarcus Aldridge and Zach Randolph do the worst job guarding shooters and also allow the highest field goal percentages.

Help Defense
One of the main defensive responsibilities of a big man is to back up his perimeter teammates when they are beaten off the dribble. The table below shows how effective the players are when they are playing help defense. The Keep in Front column indicates how often the player is able to stop a player from driving to the basket. Effective Rate shows how often the player made a positive defensive play, such as forcing a turnover, denying or deflecting a pass, taking a charge, or forcing a pass to a contested shot. 

Duncan and Aldridge have been most disruptive, with the highest effective rates to show for it. To go along with that, they also allow the lowest field goal percentage, too. The strange thing is that while Duncan has done a great job disrupting shooters (Contest+ % of 64.6%), Aldridge is the worst in this group at contesting shots. Luck may be playing a factor in Aldridge’s great field goal percentage, because opponents should be making their lightly guarded shots at a better rate. 

To no one’s surprise, Ibaka leads the way in shot defense, challenging shots nearly 75 percent of the time. Even prior to this season, when his second place finish in Defensive Player of the Year voting was called into question, no one doubted his ability to disrupt shooters in a help defense situation. When he’s not contesting shots, Ibaka is also doing a solid job of slowing down drivers, where he is second only to Gasol in Keep in Front percentage.

Ball Screen Defense
The table below shows the types and effectiveness of the ball screen defense for this group of players. The final three columns show the frequency of a particular defensive style when helping on a ball screen. The Play Behind column shows how often the defender hangs behind the screener, waiting for the ball handler. JaVale McGee demonstrates (poorly) here. The Show column indicates when the defender sets up higher next to the screener, but allowing the ball-handler to use the screen. The Hedge column displays when the defender moves perpendicular to the screener, in an attempt to force the ball-handler away from the screen, as LaMarcus Aldridge demonstrates here.  

Pick-and-roll defense is where Gasol shines. Although he may be the biggest player on this list, he is the best at keeping ball-handlers at bay when they try to attack him coming off of a ball screen. Gasol is staying in front of these drivers over 80 percent of the time. Duncan, too, despite his advanced age, has succeeded in staying in front of drivers. This plays into the Grizzlies’ and Spurs’ defenses, as these two players often playing behind the screener and wait for the ball handler to drive. 

On the other end of the spectrum, Al Jefferson has allowed drivers to beat him off the dribble half of the time. He is more likely to show or hedge on a pick-and-roll than Gasol or Duncan, but he has trouble stopping drivers regardless of his defensive tactics. Unlike Jefferson, Serge Ibaka has been successful at staying in front of a drive whether he plays a passive or aggressive style of pick-and-roll defense.

Defensive Rebounding
The table below displays the types of rebounds that the All Star candidates have corralled this season. The  Out of Area column displays whether a player moved from one area of the floor to another, toward the rebound, between the time of the shooting motion and the rebound. The Traffic column shows how often the players' rebounds came with an opposing player within a foot of them. 

Serge Ibaka leads the pack with about 70 percent of his rebounds coming in traffic. However, Zach Randolph may be working the hardest to get his rebounds. Almost two-thirds of Z-Bo's rebounds are tracked down outside of his area, and more often than not, they are in traffic. 

Although Tim Duncan is posting the highest defensive rebound percentage of his career, he rarely leaves his own area to get a rebound, and they are usually not in traffic. Duncan’s high rebounding percentage could be in for a decline if a few of those missed shots don’t bounce directly to him. David Lee doesn’t fare much better in this measure, as about 60 percent of his defensive rebounds are secured with minimal resistance from the opposition. It is worth noting that San Antonio and Golden State are the two best defensive rebounding teams in the league. As Alex of Gothic Ginobili points out, the Spurs stress boxing out and crash the boards as a team.

Tim Duncan and Marc Gasol have been getting a lot of All Star love, partly due to their strong defensive reputations. The data back up those reputations, as these two big men are among the best at guarding both their own assignments as well as their teammates'. However, that third frontcourt spot may need reexamination. LaMarcus Aldridge is a frequent selection, but his defensive weaknesses are glaring when compared to the rest of the group. He has done a poor job contesting shots, both in one-on-one and help situations, and has not been very active on the defensive boards. Although it may not show in conventional statistics, Serge Ibaka's development has led to great shot contesting, pick-and-roll defense, and overall rebounding activity. This superiority over the field on the defensive end more than makes up for his shortcomings on offense.

Tuesday, January 22, 2013

Chris Paul, Clippers Rank First in Steals. Coincidence?

The Los Angeles Clippers gave us another reason last week to be impressed.  Playing without Chris Paul due to injury, the team went 3-0, all road wins against Western Conference foes. And these were pretty convincing wins at that (average margin of victory was 15.67 points).

The idea that “this is Chris Paul’s team” has been thrown around by commentators and analysts throughout the season because of his "leadership" that has changed the "culture of the team."  All these words words and phrases sound nice, but here at Vantage, we like to translate soft sentiments into hard numbers.  If Paul is changing Clipper culture, the stats should show it.

One of the first statistics that come to mind when you think of Chris Paul is steals. Paul has led the league in steals in five of his first seven seasons in the league, and is leading the league this season. When leading the league in steals, Paul’s teams have also been ranked in the top 10 in steals. So how much has Chris Paul contributed to steals during his tenure with the Clippers? In the 2011-2012 season, 27.33% of the Clippers’ steals were Paul's. So far this season, only 22.84% steals have been his, the lowest percentage of any season in which he has led the league. And not only does he lead the league in steals this season, but his team does as well, an accomplishment he has not shared yet in his NBA career. 

Another league leading statistic is the number of turnovers that the Clippers create. The team leads the league in creating turnovers this year, compared to 16th in the league last year. Of all Clipper opponents' turnovers, 61.64% are from steals.  “Lob City,” it seems, was built on fast breaks off steals (the Clippers rank sixth in the league in fast-break points).

Using the Vantage data set, we can see what types of steals are most frequent in helping the team create turnovers, and what proportion of the team’s steals is contributed by Chris Paul. 

Our sample shows that the ratio of off-pass steals to off-dribble steals/strips has increased this season. This suggests that there is an effort on being more aggressive in the passing lanes. However, the deviation in the rest of the team’s steals from Chris Paul’s steals is seen significantly in the off-dribble steals/strips statistics. This suggests that another focus on defense for the Clippers has been given to their on-ball defense. The ratio of off-ball to on-ball steals for the Clippers in our sample is 1.04 for this season, compared to 1.40 from last season. This suggests that more steals are coming from the on-ball defender. These findings also support the notion that more steals are coming from players not named Chris Paul.

The increase in steals is also prevalent in different locations on the floor, specifically on the perimeter. In the Vantage data set, 76.53% of plays where pressure was applied by Clippers defenders were perimeter pressure plays last year, compared to our data for this season, where nearly 88.7% of pressure applied was in perimeter defending situations. Our data also show that a greater proportion of the Clippers’ steals this season take place outside the paint.

The upswing in steals on the perimeter could suggest why the team has had more opportunities to score on the fast break throughout the season. The decline in steals in the low post and hashes are likely due to the effort applied to executing a strong defensive presence on the perimeter. This allows a better chance of stealing the ball in situations that have better chances of leading to fast-break opportunities.

Holding opponents to 92.7 points per game, the Clippers are 4th best in league, and this is the culture Chris Paul is helping to change. The focus on defense is even helping the Clippers win  games when their cultural guru (yeah, it's an official title in Lob City Hall) is not on the floor.