Passing Box Score: Kentucky vs. North Carolina

After reading A Sea of Blue contributor Anything But Gatorade's thought-provoking take on Assist to Turnover ratio and its (relative) usefulness, I was intrigued to dive in further. To that end, I rewatched Kentucky's thrilling victory over North Carolina and charted all the passes on Kentucky's offensive sets.

I was interested in understanding where Kentucky's offensive efficiency (or lack thereof) was occurring, not just in tangible results (namely points and assists), but in execution (in this case, passing). After many minutes of pausing, fast-forwarding/rewinding, and spreadsheeting, I present the following information for your digestion. (Oh, and I also have newfound appreciation for all the hard work JLeverenz puts into his Defensive Score Sheets.)

PASSING BOX SCORE (BASIC)

Definitions

• Total Passes: For each player, I charted the numbers of passes he made throughout the course of the game. Each pass was broken down by type: (1) assistable passes, (2) turnoverable passes, (3) regular passes, and (4) transition passes. I did not count any pass that came from a player on the sideline passing in after a dead ball (unless it was an assistable or turnoverable pass), nor did I count any pass coming from the defensive end of the court (unless it was an assistable or turnoverable pass or lead to a fast break).
• Assistable Passes: These are passes a player made that could have turned into an assist, whether the pass receiver ended up shooting and scoring or not. Consider these "hidden" assists. Not only does it account for assists that actually happened, but assists that could have happened. Consider the following example: Terrence Jones has the ball in the post and is double-teamed. Recognizing the double-team, he kicks it out to an open Doron Lamb, whose man has collapsed into the paint. Lamb catches and and swings it to an open Darius Miller, whose man has crashed towards Lamb as a help defender. Miller takes and misses the 3. In that scenario, there were 2 assistable passes (1 by Jones, 1 by Lamb) despite 0 actual assists.
  
• Turnoverable Passes: These are passes a player made that could have turned into a turnover, whether the ball was actually turned over or not. Self-inflicted turnovers (such as traveling) do not count as a turnoverable pass.
• Regular Passes: These are passes that are made in the course of the offense that are neither assistable passes nor turnoverable passes. One of the most common examples of the regular pass is how Kentucky initiates its offense. Marquis Teague generally brings the ball across halfcourt, then will pass to a forward ballhandler (usually Jones) while the offense starts its set.
• Transition Passes: These are passes made to initiate or during a fast break that don't lead directly into a shot attempt. Since they're not exactly assistable passes, yet are more "helpful" than regular passes, I created a separate category for them.
• Assists: These are passes that lead directly to a score or free throws. That means I counted passes that directly lead to a shooter being fouled as an assist, which obviously differs from the official definition. Like errors in baseball, assists are inherently subjective and are defined by the scorekeeper; thus if you compare my box score to the official box score, there are quite a few variances.
• Turnovers: These are passes that lead directly to the ball being turned over as well as self-inflicted turnovers. Unlike assists, turnovers are not subjective. It's either a turnover or it isn't. However, also unlike assists, turnovers do not have to be the result of a pass (the 8 to 8 ratio above is merely a coincidence). According the official box score, it appears I missed a turnover somewhere by Jones.

After the jump, I'll discuss some of the numbers presented above, and delve into some deeper statistics.

Analysis

• Taking Care of the Ball: One of the keys to victory in this game was Kentucky's ability to hold onto the ball and not be turnover-reckless. By my count, Kentucky had only 8 turnoverable passes, six of which came in the first half. 5 of those led to turnovers (3 others were self-inflicted). A key reason for that is stellar wing play by Jones and Michael Kidd-Gilchrist. Neither of them had a turnoverable pass in the game, and Kidd-Gilchrist's only turnover was self-inflicted (Note: I missed Jones' turnover somewhere). Against North Carolina, Kentucky was almost exclusively a perimeter team, hardly attempting post entry passes or dribble penetration. A big reason for that is that John Calipari feels secure with Jones ball-handling on the perimeter (more on that later).
• Teague vs. Lamb: The common consensus coming out of the North Carolina game was that Kentucky was a much more effective offense with Lamb at point guard over Teague (despite Teague out-assisting Lamb 3 to 1 by my count, and 4 to 1 by the official scorer's count). However, a closer look shows that Lamb was actually the more efficient passer. Lamb attempted 11 less passes than Teague, while still notching 2 more assistable passes and 2 less turnoverable passes. It's important to note that there was a stark dichotomy in halves. Teague picked up all of his turnoverable passes in the first half. In the second half, he had 6 assistable passes to just 3 in the first half.
• My oh my, Anthony Davis: While Davis had his issues with defending post entry and defensive rebound positioning, he was amazing in many of aspects of the game. Not unsurprisingly, given his high school days as a point guard, Davis is incredibly comfortable ball-handling on the perimeter. A little more surprisingly, Calipari calls quite a few sets for him to do so, as Davis caught multiple passes at the top of the key (Note: Luke Winn's always excellent Power Rankings for a pictorial example). He's a gifted passer: I had him down for 8 assistable passes, and 3 total assists (2 more than the official box score).
  
• Etc.: The small forward position is clearly the forgotten man in this offense. Kidd-Gilchrist was hardly ever a primary ball-handling option, and Miller appeared to see more of his touches as a two-guard. Eloy Vargas is an offensive non-entity. Interestingly, in limited touches, Kyle Wiltjer basically replicated Jones' role in the offense, and was comfortable both on the perimeter and in the paint. The results aren't yet there for Wiltjer, but he looks like he'll blossom into a good offensive threat.

PASSING RATIOS

Definitions

• Total Pass %: This is the percentage of passes each player had out of the team's total number of passes. Teague distributed 49 of Kentucky's 176 passes; that's 27.8%.
• A/TO Pass Ratios: The assists to assistable pass ratio is exactly what it sounds like. Teague had 3 assists and 9 assistable passes, so only 33% of his assistable passes were directly turned into points (or free throw attempts). As mentioned earlier, every assist comes from a pass, so this ratio can never exceed one. The turnovers to turnoverable pass ratio is the same concept. This ratio can exceed one, since turnovers aren't necessarily the result of passes (see Lamb's ratio, for example). The assistable pass to turnoverable pass ratio is like the assist to turnover ratio, but better.
• A/TO Pass %: These percentages are a player's individual percentages. Teague had 9 assistable passes out of 49 total passes; that's 18.4%.

Analysis

• Ballhandling: The passing percentages are a rough indicator of how involved a player is in the offensive set. As evidenced by Teague's percentage, he's the primary ballhandler on the court. I don't think I can recall half-court possession where Teague was on the court but didn't bring the ball up. Lamb, as Teague's backup, has an uptick in his Total Pass %. However, Jones is actually the typical secondary ball-handler in offensive sets. As a 2-guard, Lamb sees a lot of his touches later on in the possession, either as a motion ball-handler or catch-and-shoot receiver. Jones actually will often get the second touch (and multiple) touches on the perimeter as Kentucky executes its motion offense. I'd argue that he needs to show in the post more, although it's understandable why that wasn't part of Kentucky's game plan on Saturday (because of John Henson's length).
  
• A Quick Diatribe on Types of Assistable Passes: One thing I haven't yet brought up. For each assistable pass, I spent time charting the type of pass it was. For example, if you refer way back to my assistable pass example, there were two pass types. The first is Jones' (1) "Kickout" pass from the post to find the open man on the perimeter. The second is Lamb's (2) "Swing" pass to the eventual shooter. In the course of play, I also identified the following: (3) "Basket Cut" - finding the cutter going toward the basket; (4) "Penetration" - like a kickout pass, but done by a dribble penetrator instead of a post player; (5) "Post Entry" - a direct feed into the post where the post player doesn't have to make his own basketball move to score; (6) "DDM" - a natural consequence of Kentucky's motion offense whereby the person receiving the ball can shoot; (7) "Curl" - similar to DDM, except the handoff occurs as the motion receiver is heading to the basket; and (8) "Lob" - a fan's favorite type, often resulting in some awesome slam dunk. Obviously, all of this is a work in progress and negotiable with you, dear reader.
  
• Assistable Pass Analysis: As I brought up earlier, Kentucky is pretty comfortable living on the perimeter. 15 of their 41 were charted as "swing" passes, and an additional 6 were "DDM" passes. I counted just 1 suitable post entry and 1 lob (a satisfying Davis jam). Here's where Teague's inexperience comes into play: 7 of his 9 passes were "swing" or "DDM", meaning easier passes stemming from offensive flow rather than creation. I charted 5 of 41 as "penetration" passes, and 3 of those belonged to Lamb. At this point, Lamb is simply much more comfortable driving to the hoop to create offense. And without a credible post presence to help draw attention and provide a target for Teague, he's struggling. "Lob" and "basket cut" passes indicate vision; had I charted North Carolina, I imagine I would have seen passing witch Kendall Marshall pick up a few of these.
  
• Turnoverable Pass Analysis: There weren't many of these, and I still want to get to usage, so I'll skip my detailed analysis of these until my next Passing Box Score. What I will point out is how safe our big men were, combining for just 2 turnoverable passes total.

PASSING USAGE RATIOS

Definitions

• Total Possession %: I also tracked each possession individually, allowing me to break down passes by possession. The total possession % is the percentage of offensive possessions that a player was on the floor. Kentucky had 64 total offensive possessions. Teague was on the court for 49 of the 64; that's 76.6%.
  
• Usage %: This isn't the same as the regular usage rate you see in advanced statistics. Instead, I charted not only all the passes, but also the shots and offensive rebounds each player took. If a player did any of those three actions, he was a counted as "used" in a possession. He touched the ball, basically. Usage % is displayed out of both total possessions and individual possessions. Teague touched the ball in 44 of Kentucky's 64 possessions; that's 68.8%. He touched the ball in 44 of his 49 possessions; that's 89.8%.
  
• Passes Per Possession: These percentages show a player's passes per the number of offensive posessions he was on the court. In 49 possessions, Teague made 49 passes (1.0 passes per possession), 9 assistable (0.18), and 3 turnoverable (0.08).
  

Analysis

• Possession/Usage Split: The possession/usage split only confirms what I've been discussing in this post (and what others like Ken Pomeroy track). Teague, while seeing less possessions on the court than the other starters, touches the ball almost all of his oncourt time. Jones and Lamb are the next to in line, with Jones as the secondary ball-handler and Lamb as the catch-and-shooter with backup 1 duties. I actually thought Miller played more possessions in this game than he did, but apparently not. When he's on the court, he's involved and gets some point forward time. Davis actually posted the lowest usage percentage of the regulars; I suspect Kidd-Gilchrist's 4 to 2 edge in offensive rebounds was the difference.
  
• Who's Doing the Passing: Again, more confirmation. Teague averaged 1 pass per possession, the only Wildcat to near that water mark. Lamb and Jones ranked 2nd (0.70) and 3rd (0.59). Kidd-Gilchrist is generally a black hole as a finisher or offensive rebounder. Davis is a passer on the perimeter but doesn't really know what to do in the post. From an efficiency standpoint, Lamb shows to be the more efficient passer over Teague (0.21 to 0.18 in assistable passes per possession). As I mentioned before, Davis is a rather effective passer as well (0.15 assistable passes per possession). My plan is to do a few more of these throughout the year, whereby I'll hopefully be able to start doing some trending when it comes to per possession statistics. Right now, Teague's 1.0 on its own is relatively meaningless to me. I'm not sure if that's good or bad.
• What Worked, What Needs to Improve: After seeing the numbers and rewatching the game, I think it's obvious that Calipari went with a conservative, perimeter-oriented game plan that exploited North Carolina's defensively-challenged guards (Marshall and P.J. Hairston, most notably) while avoiding the aircraft carriers (Henson and Tyler Zeller) like the plague. However, Kentucky absolutely needs to establish an inside-out passing game, and not just the finishing lob plays that blitzkrieged St. John's earlier this year. And for all the teeth-gnashing about Teague, his development will be critical. Lamb is just so good as an off-ball guard, and while Miller has been a pleasant surprise as a point forward, asking him to run more of the offense than he does right now may be a stretch. Handoff and swing passes only go so far, especially when you shoot 4-17 from deep like the Wildcats did against UNC. Sometimes you need a magician passing the ball, and Teague has the best chance to be that guy. Finally, limiting turnovers were good. Low turnovers means more shooting opportunities means more potential points. That's a good start to the recipe for success.

If you've made it this far, thanks for reading. I hope it offered you some additional insight. Any and all feedback is greatly appreciated. As I just mentioned, I hope to do this some more and see what additional information can be gleaned. If you think I'm missing something or want to see the information presented differently (I originally had an idea for some Luke Winn-esque charts and graphs wizardly, but realized I had no interns to help me carry out my vision), let me know. I'm all ears.