Back in the day when I was still a capable guard and ballin’
, my center and I would orchestrate a method before tip-off to snag the first possession. I would stand outside the circle, and as the referee tossed the ball into the air, my center would intentionally graze the ball, barely touching it. The opposing center would easily tip it back, but I would rush into our opponent’s side, jump, and grab the ball. With my speed, I would dart though the surprised jerseys and race for a layup. It worked about 70% of the time, with the other 30% resulting in hilarious backfires.
This piece is obviously about the opening tip off in basketball games, and you’d think I’d come up with a better introduction since I’m literally talking about the first play of the game. But all I have is that anecdote, which never actually happens in the NBA.
Ultimately, the winner of the tip-off undoubtedly means nothing, because basketball is a game of equal, alternating possessions. Still, I decided to dive into play-by-play data provided by ESPN.com and discover who the best jump ballers in the league are. I’m interested in this data because tip-offs are generally an isolated play, free of noise. It’s a simple matchup between two bigs and their raw ability to time and leap.
There’s already some previous research on jump balls. My data will be somewhat different from Weakside Awareness’s study. I excluded middle-of-game jump balls to avoid mismatches. I only want to capture situations where the coach sent their perceived “best man for the job,” not happenstance jump balls when two players fighting for possession were whistled. This way, I’d omit matchups where a big faced off against a guard. Furthermore, such jumps would have a more random result due to the man-to-man positioning of the other eight players. In opening tip-offs, the point guard is stationed alone in his backcourt, where he can gain possession from his teammate’s successful rudimentary backward tip1.
So my data is essentially composed of center court tip-offs: those occurring at the beginning of the game and any overtime period. There isn’t much else, so let’s get right into 2012-13’s leaderboard (minimum 20 tip-offs, including March 15th games).
And here are the ten worst.
Position and height are loosely defined by Basketball Reference. There’s a minor element of bias in the second list; any team tracking their big man’s record should, hypothetically, cease such duties for the poor jump baller, which is why we might observe lower samples for these bigs. This is unless you’re the Sixers, and you’re stuck between three men without much vertical in Lavoy Allen, Kwame Brown, and Spencer Hawes.
Weakside Awareness did a great job detailing the effect of height in these matchups. I’d be interested in a few other variables to see if winning percentage can be modeled: weight and vertical. I’d posit that the latter is most significant; unfortunately I can’t prove that, as the data is scarce and limited to draft combines (more on this later). We can note though, that the top 10 list hosts the league’s more athletic bigs.
Luckily, there are some methods to find out whether these tip-off stats can be verified as meaningful. Vertical is also a basis for performance in sister stats like rebounding and blocks. I gathered that data from HoopData and ran some quick individual correlations (sample of 151 seasons, minimum 20 tips).
As predicted, there are positive, statistically significant correlations between tip-off success and rebound rate, as well as blocks. Keep in mind, these are two dependent outcomes I’m correlating (I’ll talk about how tip-off win rate could be interpreted as an independent variable later).
There’s a discussion to be had on how much jumping utility constitutes a player’s performance in these related stats. Tip-offs are an independent event, so height difference, timing, and vertical should be the major indicators. But for rebound rate and blocks, basketball aptitude and surrounding players would contribute a fair share too, especially in the continuous flow of a game. This is likely why our R-squared is small, but still statistically significant.
What’s also worth exploring to me is year-to-year correlation: is winning tip-offs a repeatable skill? I paired together data in consecutive years from 2010-13, and kept only players with 20+ tip-off matches in both seasons. I plotted those 68 player seasons’ year-to-year win rates2. My data can be found here.
We’re comfortably above the threshold of stability. Year to year, Anderson Varejao doesn’t waver far from 50%, Tim Duncan is holding consistently at 67%, and Al Jefferson will need springs on his sneakers to ever budge from 45%.
Though that reliability exists, the more compelling cases are those with larger residuals. What in fact, explains the variance? Well, it would be naïve to assume absolute consistency year-to-year. For an action that relies heavily on timing and athleticism (things we can’t quantify), I would speculate that tip-off performance is dependent on the improvement or decline of those native abilities.
On the positive side of the spectrum this year is Emeka Okafor, who went from winning 40% last year to 61%. Not quite obeying his aging curve is Roy Hibbert, who fell from over 50% to under 40%.
Here are the deltas of at least ten percentage points. Year represents the season’s end year.
My first reaction to these was noticing some correlation with overall player performance. LaMarcus Aldridge took a clear step forward in 2011, and Roy Hibbert has regressed this year, without question.
I thought about what might cause these residuals. Perhaps they correlated to the other athleticism related statistics above? I found small, positive correlations with the respective deltas in defensive rebound rate, total rebound rate, and percentage of own shots blocked. My sample size was too small (by pairing seasons together) to draw significant conclusions, though. By using these deltas, we’ve introduced lots of noise.
However, I wouldn’t discount the idea of change in tip-off performance being related to simply a player’s raw growth or decline. The hypothesized effect might just exist for some cases. Such cases aren’t necessarily outliers, either. Each player can be his own data set.
DeAndre Jordan, Serge Ibaka, and Josh Smith are young bigs who are rising toward their peaks. They’ve also remained healthy over their entire careers. Take a look at their year-over-year tip-off win rates.
It’s not surprising to see these three athletic big men improve on tip-offs each year. The more interesting question is whether we’ll see continued progress, or their peak. Two players in my four season data set have eclipsed Jordan’s 77% (minimum 60 tips): Samuel Dalembert (64-18, 78.0% at age 28), and Andrew Bynum (54-15, 78.3% at age 22). Both of these monstrous records occurred in 2010.
It’s harder to identify cases of deterioration. As I remarked before, a smart coach would pull their designated tip-off man when the symptoms of aging manifest, such as decreased vertical. This prevents us from witnessing a decline in the data.
This brings me to a question that’s been underlining my analysis. Given that timing and athleticism are skills that we can only measure with our eyes, can the rate of winning tip-offs be a reasonable proxy? Can I estimate a player’s vertical/jumping utility by their success on tip-offs?
Proving this would be difficult. Players perform a variety of tests at draft combines and camps, which we don’t always have access to. But suppose we did have a quantitative assessment of timing and athleticism. Could combining them with age, height, and weight provide us with enough input variables to model a player’s success on tip-offs? This would help us determine the potential causation between vertical and tip-offs.
(A few things: there would likely be Shaq-sized covariance in those inputs. And yes, we’d have to adjust for quality. And I could be totally wrong; height might ultimately matter more.)
I’m making some assumptions here. The primary one is that tip-offs solely rely on the raw ability to time and leap, as I mentioned at the beginning. Related to that assumption is the idea that other on-court players have minimal influence on the result3. Lastly, competing players must be jumping to their maximum vertical each time. I don’t think any of these is a stretch: as I said, tip-offs might be the most isolated play in basketball after free throws. That lack of noise can really be helpful in determining the factors that influence the result.
Unlike free throws however, tip-offs don’t monumentally affect the game. Evidently, I’m making quite a big deal about a minor basketball play that is literally complete in a second. In context of a game, it’s mostly irrelevant. But I think there are intriguing questions to consider regarding the players competing in tip-offs.
The minutiae of sports are fascinating, because you never know what insights can be gleaned from them. Perhaps my questions are overblown. But it won’t stop me from being curious and asking questions about basketball. Whether they’re trivial or provoking, we can always learn something new.
1. I’m going to be making reference to the receiver of the ball again, so here’s a chart listing the breakdown of who caught the tip-off, by position. These aren’t the positions assigned in the particular game – just what Basketball Reference deemed the receiving player for that season.
I’ve calculated these for sanity, rather than analysis. There is year to year consistency and no surprises, a good result to move forward with this study with.
2. Playoff data is excluded. I thought about this extensively, but decided that there could be some “matchup abuse” that would inflate and deflate the numbers, as well as unfair sample size skewing. For example, Serge Ibaka won every tip against the Mavericks’ Brendan Haywood in their quarterfinal series last year. If you’re keeping track, yes, Brendan Haywood is still terrible at just about everything.
3. Related to footnote one, I contemplated limiting the data to tip-offs where a guard received the ball. Such situations would be a clear indication of a direct backward tip (i.e. a clean tip-off win by the big). Doing so would cause sample size concerns, though.
This week I decided to take a look at players who rebound, but also score. I charted the total points each team received from players grabbing at least five rebounds in an effort to determine if teams with high scoring rebounders won more often than not.
I figured that winning teams would have a higher average point total from qualifying players, but never did I imagine the difference being this drastic. During the 50 game week, winning teams received an average of 50.8 points from players notching at least five rebounds, a 31.6% advantage over losing teams.
Interestingly enough, the scoring output was more consistent from the losing team than it was the winning teams. The losers had a range of 65 points from their highest scoring game to their lowest (Toronto totaled 75 such points while the Celtics managed only 10) and the victors had a range of 91 points (the Thunder managed 97 points while the Wizards notched only 6). Oklahoma City’s production on Friday night (97 points from players with 5+ rebounds) out did 44.4% of the winning teams total points for the night.
The Miami Heat recorded the second (21 points) and third (22) lowest outputs by a winning team. If you subtract these two games from the study, the advantage for winning teams increases to 37.2%. But they were the exception, not the rule, when it came to elite teams in this study. Oklahoma City more represented the norm, as they tallied high point totals in losses and wins. The Thunder had the second most points scored by their leading rebounders (71) in defeat and recorded the highest total in a victory (97).
Just another step in my effort to understand the game of basketball. Do you have a question you’d like answered? I’ll run your statistical inquiry through the gauntlet for the next seven days and provide you with a bit of insight. Tweet me your ideas @unSOPable23.
With all of that being said, here are your 35 stats from the week that was in the NBA.
This week I thought I’d take a look at the shot taking/making of offenses “at the rim” and behind the arc. Theoretically, offenses work hard to get a good look from one of these spots on very possession, and I was curious which had a greater impact on the game. My hypothesis was that winning teams would have the consistent edge “at the rim” while the three point shooting would be something of a crapshoot, an indication that a team can live/die by the long ball. I also wanted to see where the winning team gained the largest advantage on a per game basis. My thought here was that this study would prove that while three point shooting can win games, pounding the ball in the paint is the way to have consistent success in the NBA.
For the most part, my train of thought was on the money. Winning teams shot 69.3% “at the rim” and 39.0% from distance during the 54 game week while the losing teams shot 63.1% and 34.0% respectively. What surprised me about the results were the shots attempted at each location per game. The winning team averaged 25.5 field goal attempts at the rim while the losing team averaged 25.2. The results for three point attempts were nearly as symmetrical, with the winning team shooting 19.7 per game as opposed to 18.8 from the losing team.
For the week as a whole, the winning team outscored the losing team by an average of 3.6 points “at the rim” and 3.9 points from distance. There were a few outliers (the Bucks made 14 triples and 13 shots at the rim in a loss to the Cavs and the Knicks connected on a mere eight from point blank and 16 from distance in a win against the Hawks), but for the most part the data was pretty consistent. Teams that made 10+ three pointers won 60% of the time and teams that made 20+ shots at the rim proved victorious 75% of the time. My conclusion is that if you’re a good three point shooting team, let it fly, but if you’re an elite interior team, you will have more long term success.
Let your voice be heard and tweet me (@unSOPable23) your stat of choice for this week’s #StatStudy. You’ve got nothing to lose. This is your chance to uncover NBA data, don’t miss out! With that being said, here are the stats to amaze from the week that was in the Association.
1. Which All-Star selection fills you with blissful joy?
Kyle Soppe – @unSOPable23 – Jrue Holiday, for all the critics who say that the 76ers are a team without a true star player. This kid was a prodigy when he went to UCLA and has been as good as advertised in Philly. He already has 53 more assists than last season (27 fewer games played) and has seen his scoring average jump by nearly 50%. How many point guards in the league average at least 17 points and 9 assists? Only one.
Matt Cianfrone – @Matt_Cianfrone – Paul George. As I Bucks fan I should hate George but I just find it so hard. A superb defender, stupid athletic, great passing young guard who has carried his team minus what many people think is their best player. I am glad to see George rewarded even after his slow start. Also I already can’t wait for his dunks that will come in the game. It is going to be great.
Myles Ma – @MylesMaNJ - Tyson Chandler. Yes, this is a total homer pick. But this selection absolutely fills me with blissful joy. Tyson Chandler has finally made an All-Star team after serving his time as the lynchpin of a Knicks defense whose perimeter defenders volunteer as traffic cones at the DMV. It’s his first All-Star game, and it comes in the midst of one of his finest seasons. Over the past three years, Chandler has decided to limit his offensive game to just dunks and free throws, with spectacularly efficient results. This year, he’s perfected the art of the tap out, turning a lot of J.R. Smith bricks into the midpoints of extra-long possessions instead of the unhappy endings they usually are. He even made No. 8 on GQ’s 25 most stylish men of 2012. Even with that scraggly-ass beard. It’s definitely his year.
Kris Fenrich – @DancingWithNoah - David Lee (I almost typed “David Curry”) with Jrue Holiday coming at a close second. I often refer to Lee as the modern-day Bob Pettit and I’m only partially joking. He scores with ease, rebounds well, has well-above-average vision for a four man and passes well. And none of this is new, it’s just the guy’s never been in a winning situation before. Good to see his multiple skills acknowledged among the league’s best.
Michael Shagrin – @mshaggy -Kyrie Irving. When it’s all said and done, I think this kid will have the last laugh. He’s a Chris Paul look-alike with more size and a smoother J. And he’s only 20 years old! Classic Kyrie outing: the night he returned after breaking his finger, the Cavs played a nail biter against the Lakers with Kyrie going for 28 points. As Kobe tried to wrest control of the game from him in the final minutes, he cooly steered Cleveland to victory. His absence from the starting unit was almost my answer to the following question…
This week I had a stat request from Jesse Silverman (@JesseBeau)
, a Cortland intramural champion and defending March Madness king. He wondered if the number of double digit scorers was directly correlated to victories, guessing that winning teams would have more double digit scorers in each game.
As it turns out, there was not a considerable difference in the number of players in double figures for winning and losing teams. Over this holiday week (46 games), winning teams placed an average of 4.67 players per game into double digits while losing teams averaged 4.11 such players. For this small sample size, a remarkable 81.5% of teams had at least four players tally 10+ points, a number much higher than I speculated based on the emergence of the “Big Three” era. The Heat and the Bucks proved to be outliers rather than the rule, as their success directly depended on star players/volume scorers.
In addition to tracking total players who eclipsed 10 points, I tallied the number of players who did so in a reserve role. The average losing team had a higher percentage of their double digit scorers come from their bench than the winning teams. For the week, 31.6% of double figure scorers from losing teams didn’t start as opposed to 29.8% for winning teams. My initial guess was that teams who had multiple double digit point producers come off the bench (Clippers and Spurs come to mind) would be awfully successful, but such teams won only 56.3% of the time.
Curious about a stat of your own? Think you’ve got a trend that predicts victories but don’t have the patience to research your gut feeling? Looking for an edge in your fantasy basketball league? Whatever the case may be, tweet me your #StatStudy ideas (@unSOPable23) and I’ll make it my upcoming research project.
Here are the week’s best stats and a look at some odd trends to consider the next time you flip on an NBA game.
This week’s study was suggested by college basketball guru and FSWA (Fantasy Sports Writers Association) gate keeper Perry Missner (@PerryMissner
). He is a passionate basketball fan and his suggestion led to some very interesting results. Perry wondered which was more directly correlated to success: quantity or quality when it comes to field goal attempts. I ventured a guess that FG% would be a better predictor, but figured it would be a close study. After a full week of research (55 games), I can safely say that I was wrong.
To my surprise, only 49.1% of teams that attempted more field goals than their opponents won the game. On the other side, an amazing 84.9% of teams that shot a higher percentage from the field won their game. The link between FG% and victories is high, but understandable. The fact that losing teams were (on average) taking more shots than winning teams is what surprised me. I can account for a slight difference by saying that winning teams are getting fouled down the stretch and thus attempting fewer field goals, but that isn’t going to explain a 35.8% difference.
This study tells me more about defenses than offenses. Successful teams in today’s NBA are capable of forcing their opponents to take bad shots on a consistent basis, allowing them to win even on nights where they don’t force many turnovers. It is also reasonable to think that winning teams play a patient style of offense, forcing the defense to buckle down for longer periods of time, and thus increasing the opportunity for a breakdown.
Moral of the story: less is more when it comes to offense in the NBA. This could be a indicator that the run and gun offense could be a thing of the past, as methodical teams are having success at routinely getting the shots they want.
Have a stat you want reviewed and interpreted by me in a weekly column? Don’t be shy and tweet me @unSOPable23 your idea.
Here are some interesting numbers from the rest of the week:
Thanks to the innovation of Hoopdata
, and the creative work of Kirk Goldsberry
and others, shot location information has become a bigger and bigger part of the general basketball discussion. Even the most casual fan can now identify long-two pointers as the least efficient shots in basketball and corner-threes as an important offensive weapon. The ideas haven’t just permeated the consciousness, the data has as well. With a simple search at Hoopdata, Basketball-Reference
or several other sites, you can find information on who shoots from where and how effective they are.
Over the past few years I’ve used this data for some offensive analysis that I called Expected Scoring. By looking at shot location data for all players in the NBA we can calculate the average points scored on shots from different locations. Obviously each shot can only result in 0,2 or 3 points, but by dividing the total number of points scored by the total number of shots from each location we arrive at an average per shot value. Using this expected point value for shots at the rim, three-pointers, and everything in between gives us a different dimension to measure the offensive efficiency of a player. In the past my Expected Scoring work used those expected point values to calculate how many points per 40 minutes a player should be expected to score given their shot selection. We could then compare that expected total to their actual total to see how reliable their overall shooting was compared to the league averages.
I want to use that same basic idea today, but to take it in a different direction. What I’ve calculated is something called Expected Points Per Shot. I’ve used the same techniques for my calculations but I’ve dialed it down from per 40 minutes to per shot. What Expected Points Per Shot (XPPS) gives us is an objective way to evaluate and compare a player’s shot selection to that of his peers.
XPPS is calculated by multiplying the number of shots from each location by the expected value of shots from that location, adding that product to the product from all other locations, and then dividing by the total number of shots. A player who takes a lot of their shots at the rim, or from behind the three-point line, will have a higher XPPS because those shots have a higher expected value. I’ve taken my underlying statistics from Hoopdata so I’m using their breakdown of shot locations – At the Rim, <10ft. , 10-15ft. , 16-23ft. , Three Pointers. I’ve also included free throws in my numbers. I used the standard modifier of 0.44 to arrive at the number of free throw attempts and made free throws that resulted from shooting fouls, eliminating and-1s and free throws from technical fouls.
I’ve created a Tableau visualization to display the results:
This week I took a look at which statistic had a greater impact on a game’s outcome, transition points or points in the paint. I charted the fast break points and interior points for each of the 49 games this week and broke them into two separate groups: winning and losing teams. I then divided the totals of the losing teams by the number of points scored to determine what percentage of points were scored in transition/paint. I repeated that process for the winning teams, and compared my findings.
The results weren’t overly shocking, as winning teams outscored losing teams on the break by a higher percentage than they did in the paint, but the wide disparity caught me a bit off guard. Winning teams scored an average of 14% (14.37 points per game) of their points on the break while losing teams scored 11.6% (10.53 points per game) of their points in transition. The 3.84 point difference per game reflects that winning teams outscored losing teams by 26.7% on the fast break.
When it comes to scoring in the paint, the results weren’t as definitive. In fact, losing teams actually accounted for a higher percentage of their points (42.7% compared to 40.97%) in the painted area. Victorious teams averaged 42.04 points per game in the lane while losing teams tallied 38.76 points per game in close, a 3.28 (or 7.8%) point advantage for the winning teams.
Could this explain the struggles of the Lakers, a team who has a “fast break” oriented coach but a “points in the paint” oriented roster? It may only be a week long study, but teams that excel at running have a better chance at winning games than teams that slow it down play a bruising style on the interior.
Remember, if you’re curious about any stat, tweet me @unSOPable23 and I’ll do a weekly study on your stat. Let’s team up and uncover some of the most unique stats/trends that this game has to offer.
Here are some of other interesting numbers from the past week.