It wasn't a fluke that Westbrook scored 41 against both the Lakers and Celtics in the month of February. Westbrook has scored 40+ on LeBron even in the playoffs. Westbrook dropped 40+ on lebron, dwade, Chris Bosh, Battier, Chalmers, and the entire Miami heat back in game 4 of 2012 nba finals!
We are going to play small ball until we get the doughboi out the front office or the greasy Sicilian off the floor. Illl hola
LeBron got benched late in the game because of load management. The last minute and 30 seconds of that game the media showed LeBron on the bench sweating like a villain and gased like he needed oxygen. Especially on the Westbrook dunk Westbrook took off from around the free throw line like Jordan in the 89 dunk contest!
[Kelly Iko/Seth Partnow] An analytical study of defense, shooting, and small ball, with an expert’s help Spoiler In case you haven’t been paying attention recently, the Houston Rockets have decided to downsize things a bit on the floor. Small ball has always been a love of Mike D’Antoni’s, but he was never truly committed to a long-term relationship, opting to flirt every now and then. That changed this February when the team traded Clint Capela to the Atlanta Hawks in exchange for respected 3&D stalwart Robert Covington. But how impactful can Covington truly be joining a contender in the middle of a season? What’s the best way to measure that, besides on and off splits and defensive ratings? Is he truly what Houston has been lacking? Offensively, we have a good idea of what that will look like for Houston on a game-to-game basis. But what about the quality of the shots they take? The three-pointers will come early and often, but what goes into a ‘good’ shot? Does it have anything to do with how fast or slow a team plays? The Rockets pace has skyrocketed since swapping Chris Paul for Russell Westbrook, but the team’s offensive ranking hasn’t wavered. Most importantly, is small ball the answer? To figure out the answers to these questions and more, we enlisted the services of Seth Partnow, our colleague and analytics guru, and former director of basketball research for the Milwaukee Bucks. We’ll find out the truths about small ball, the differences between pace and playing fast, what goes into a good shot, Robert Covington’s defensive impact, rim protection issues, and more. Kelly Iko: Before the hiatus, we had seen about 15 games of Houston’s new small-ball look (or old, depending on who you’re asking). But before we dive deep into the new approach, I want to understand the nuances surrounding the decision in the first place. Seth, the Rockets have been touted as one of the analytical pioneers for as long as I can remember. As it pertains to on-court aspects they’ve played smaller at times, but for the majority of the time, they’ve always employed a traditional big. Assuming you were working in the Rockets’ analytical department, what trends would you have been looking at (either this season or past) to make the case to go without a center? Seth Partnow: I think Daryl Morey has discussed this publicly, so I’m not exactly going to be splitting the atom here. One of the main differences in going from Chris Paul to Russell Westbrook is a sizable downgrade in outside shooting threat. It’s been reasonably well-established that “shooting” is an additive skill, in that the more shooters a team has on the floor, the better. Shooters don’t crowd each other out in quite the same way as on ball creators do, and in fact there is some evidence that each additional shooter on the floor is more valuable than the last. To put it another way, four shooters in a lineup is a bigger improvement on three shooters in a lineup than are the gains from going from two shooters to three. When the Rockets had lineups featuring Harden and three shooters, Capela’s roll gravity plus his defensive presence was beneficial. When this changed to Harden, Westbrook and two shooters, that equation changed. Plus, over the last two seasons the Rockets have gone from putting Harden and Capela into pick-and-rolls to simply isolating Harden. So, reduced spacing and declining use of Capela’s best offensive skills had turned his presence into something of a drag on the offense. Going pure five out is generally considered trading offense for defense, but Rockets defense was only 1.5 points allowed/100 possessions worse with Capela off the floor, while the offense was 5.8 points/100 worse with Capela in the game, that net swing of around 4.3 points/100 is not wholly attributable to Capela, but it’s a big gap and illustrates that something wasn’t working. The other thing that I would have brought up is the “let’s get weird” factor. As I discussed as part of breaking down the Capela/Covington trade with our Danny Leroux and Sam Vecenie, the Rockets were already experimenting with playing smaller for longer than any team in recent memory: Not to belabor the point, but the Rockets are entering largely uncharted territory in the modern game. Even in the downsized 3-point revolution era, only the 2013-14 Mavericks have played at least 10 percent of their minutes over a full season with the tallest player on the floor listed at 6-foot-7 or smaller. Only seven other teams, including Houston last year, have played even 1 percent of their minutes with those lineups over that span. Now, it’s not quite apples with the NBA’s crackdown on height inflation prior to this season, but the Rockets are already blowing that mark out of the water with around 27 percent of their minutes occurring with no players taller than 6-7. This is only going to increase with P.J. Tucker at 6-5 and Covington at 6-7 likely getting the bulk of the minutes as the “big guy.” With no real precedent to evaluate what this might look like I almost have to shrug my shoulders and say there are going to be a lot of apparent size mismatches on both ends of the court in Houston games the rest of the way. It wasn’t hard to see the Rockets weren’t good enough to really challenge for a title trying to play their version of traditionally, while we don’t really know what all small all the time looks like in the modern NBA. So if you’re championship or bust, embrace the variance of entering the unknown! Iko: Would it be fair to say this switch was a result of trade-offs and variance? Meaning the need for space, speed, and variance like you said. With that being said, does a move like that have to be an all-in approach from a front office? I remember Covington’s first game against the Lakers, a game in which D’Antoni compared it to taking a big test without studying the night before. It’s not everyday teams move a young, starting, productive center—this trade signaled there was no going back. Partnow: I’m going to cheat on the test and reject the premise. “All-in” or “Not” isn’t a binary choice. That said, the Rockets were already all-in across a number of dimensions, with their championship or bust mentality relying nearly exclusively on the Harden-plus-spacing strategy. In that context, moving Capela for Covington wasn’t “going” all-in so much as re-calibrating based on relatively small changes in the underlying mentality as we just discussed. Iko: Speaking of Covington, how would you quantify his potential defensive impact prior to his arrival? What stats are you looking at when it comes to acquiring that talent and raising a collective defensive ceiling? This season, Houston didn’t switch as heavily as in years past before the trade. Partnow: Covington’s defense is sort of a walking advert for the importance and utility of statistical analysis. From a traditional viewpoint, where our eye naturally follows the ball, he’s ok. He’s neither quick enough to guard the top scorers who rely on quickness nor stout enough to contend with those who use power. But his best attributes aren’t the sort of on-ball “lockdown” stuff which people tend to focus on. Rather, his length, anticipation, and versatility make him an extremely effective off-ball disruptor. Among non-centers, only Covington and Kent Bazemore have had Steal and Block percentages over 2.0 for each of the last 4 seasons. Covington is easily the better shooter once you did a little deeper than raw percentages. He has hit 35.6 percent for his career compared to Bazemore’s 35.0 percent. But only 27.5 percent of Covington’s career 3-point attempts have been uncontested, compared to 63.1 percent for Bazemore. Among players with at least 500 career attempts, only 10 have taken a larger portion of contested 3FGA than Covington. This willingness to let it fly at volume makes Covington such an ideal fit offensively with Houston’s overall ethos. But we’re talking about defense, where Covington’s nose for the ball has placed him among the league leaders in deflections year-after-year. Iko: So if we’re talking defense, what data effectively measures his presence and Houston’s overall acumen on that side of the ball? Certainly, there’s more to it than on/off splits and defensive ratings. Two areas they have struggled in since going small have been rim protection and corner 3s allowed. A big part of that is their “shrink the floor” strategy, but what are you seeing?
Spoiler Partnow: In terms of overall effectiveness, we can look at some on/off derived measures from the Adjusted Plus/Minus family. If you’re not familiar, APM is essentially a regression model that looks at the 10 players on the floor on any given possession and estimates each player’s positive or negative impact on the scoreboard controlling for teammates and opponents. My favored flavor of these models is known as Regularized Adjusted Plus/Minus, or “RAPM,” which uses some fancier modeling techniques to better measure the effectiveness of low minute players while also allowing for separate calculation of offensive and defensive contributions. Other models have been shown to be somewhat more accurate, but I prefer RAPM because it has, for me, the best combination of explainability and accuracy. So, that being said, Covington has, prior to this season, but a DRAPM monster, rating in the top 5 each of the last seasons. Now, RAPM scores are better interpreted as broad indicators rather than explicit rankings, but even with that caveat, Covington’s results indicate that he has been a high impact defender to an overwhelming likelihood. So far this season, he has been less impactful, which one could interpret as backing up the notion he is a less effective defender of primary scorers than as more of a center fielder. For much of the season in Minnesota, he was forced to be the former. All that said, early (extremely small sample-sized) results in Houston have been mixed, with Houston giving up a fair proportion of shots at the rim (37th percentile of lineups in preventing corner 3s with Covington on the floor, per Cleaning the Glass) and a lot of corner 3s (21st percentile). They have defended the rim reasonably well on the shots they have allowed, as Covington’s mobility and length have made him reasonably able to rotate and contest from the weakside. The ability to scramble the game has shown in terms of turnovers forced, with 16.3 percent (88th percentile) of opponent possessions ending in turnover with Covington in the game. They have also done very well at avoiding fouls. Unfortunately, and predictably, Houston has suffered on the defensive glass, allowing opponents to grab 29.8 percent (5th percentile) of their own misses. Overall, this defensive profile is of a piece with going all-in on variance. The Rockets are in effect gambling that their peripatetic style will force turnovers and discombobulate shooters (opponents have been particularly inaccurate from the midrange with Covington on the floor) to the point where it offsets some of the interior play and shots off of penetration or ball rotation that this sort of aggressive style is bound to give up. Iko: To me, it seems like this strategy could yield mixed results, right? Of course, there’s a risk factor in any strategy, but is small ball’s mitigated compared to a more traditional style of play? Is there any data out there that suggests you could tweak things, as far as what “shrinking the floor” does? In the small sample size we’ve seen of this new look, they’ve had some impressive wins (Celtics twice, Lakers, Utah twice) and some bad losses (Knicks, Hornets, Magic). What does your data say to explain that? Is it simply the whims of an NBA season? Partnow: The Rockets have such a small sample in this configuration that I’d be hesitant to draw broad conclusions on what will and won’t work. You’re right to focus a bit on the ebb and flow of a season, which actually leads to me having a question for you. The high energy nature of the way Houston needs to play to execute on this strategy requires the team to be “up” for a game to really work. Looking at the lists of good and bad results, the disappointing losses sure look like a group of teams where a top team might struggle for total focus in the dog days of an NBA season. I know I’m asking you to cast your memory back over the decades it had been since March, but would your observation confirm the notion that the Rockets had brought less intensity to games against lesser opponents? Iko: I think that’s been their biggest problem during the season. The recent losses to the Knicks, Hornets, and Magic were really eye-opening, not because of the timing of the season, but because of how much of a constant theme it’s been for 2-3 seasons. There’s really no way of quantifying it either or any statistical trend you could analyze—the Rockets really don’t care about lesser opponents. You can see it in their body language on those East Coast road trips and even against teams in their own conference, they really just don’t seem that interested. You can chalk it up to sloppiness, defensive lapses, or missed shots, but the truth is that they just don’t bring that same intensity as against the best of the best. I think the only way you could get the best, or at least the most attentive version of the Rockets is to have them play the 2017-19 Warriors on a constant loop. There really isn’t another team where I can definitively say Houston sees them as equals, at least from their perspective. Maybe this year’s Lakers and Clippers teams, but there’s no real storyline there like there was with the big bad Warriors. But since we’re on the topic of ebb and flow, I wanted to get a better understanding of two things: pace and quality of shots. Mike D’Antoni has preached speed for as long as I can remember, urging his team to play even faster upon the return. How effective has that been in the past and is there any indication that it will see success? The second question is about shot quality. We see James Harden take a ton of stepback threes, something he’s good at. But what about the others? What constitutes a ‘good shot’? As it was explained to me when Westbrook arrived in Houston, the team wasn’t necessarily worried about his 3P% because at the end of the day it’s a 50 percent probability you’ll hit any given shot. Is that true? Partnow: Who said that about any given shot? That’s physically painful for me to hear. When we talk about shot quality, it’s usually meant generically. Stats that show up as “shot quality” try to identify the average accuracy of a shot with certain given characteristics. This can range from something relatively simple like finding league averages by shot zone and applying those to a player or team’s mix of shots to in-depth use of tracking data to include factors such as how many dribbles the player has taken prior to the shot, proximity of defenders, the speed and angle of their movement relative to the basket, similar velocity-angles of closing defenders, time remaining on the shot clock and so on to get a much more specific measure expected accuracy, to everything in between. For example, just looking at defender proximity, here are average efficiencies on shots at the rim, from mid-range and from beyond the arc separated by into 2-foot ranges of defensive attention: So by this accounting, if a player attempted half of their shots completely uncontested at the rim, and the other half were highly contested midrange shots, their average shot quality would be (93.1% + 35.5%)/2 or 64.2%. It sounds simple, but it can’t be emphasized enough that this measure assumes a perfect average shooter – in other words a myth. Every player in the league is better in some areas relative to average than others. So while a wide-open 3-pointer has an expected accuracy of 57.6 eFG% (which translates to 38.4 3FG%) over his career, Westbrook has hit only 34.6 percent of his uncontested 3s for 51.3 percent eFG%. Which illustrates the broader point which also encompasses Harden’s step-back threes: high-quality shots are ones the shooter can make at a high rate. While this borders on tautology, it needs to be kept in mind when we discuss shot quality or expected FG% or anything like that. With me so far? Iko: Yep Seth, I’m following! (Let it be known that I never talk with exclamation marks)
Spoiler Partnow: So to complete the thought, general measures of shot quality are a useful indicator but need to be considered in conjunction with who is shooting those shots. Unfortunately, very few players take enough shots across a range of situations to derive any sort of player-specific shot quality model. To account for this, I prefer talking about a players’ shot profile across two dimensions: shot quality and shotmaking. “Shotmaking” is simply the difference between what that hypothetical average player would shoot on a given set of shots and what a player has achieved. To continue the Westbrook example, his shotmaking on uncontested 3-pointers listed above would be 51.3%-57.6% = negative 6.3, whereas Eric Gordon (41.7 percent career on uncontested 3FGA, 62.6 percent in eFG% terms) could be described as 62.6% – 57.6% = plus 5.0. This kind of accounting allows you to identify both the kinds of shots a player gets and how well they make them. For example, though Capela has always been an efficient scorer from the floor, this is in large part because he tends to take only high-quality attempts, thus his raw shooting numbers overstate his scoring ability to a degree. By comparison, Harden is efficient on an extremely difficult mix of attempts. If you don’t want to do the math or reference tracking data, Basketball reference has recently introduced a quick and dirty measure called “adjusted shooting” which doesn’t include any sort of tracking data, just general shot location info. On that measure, if an average 3-point shooter is a 100, Westbrook has been a 71 this season. Not great. Finally, to circle back around, “make or miss league” does NOT imply each shot is a 50/50 proposition. Iko: It’s possible that whoever said it meant it in a joking manner but I didn’t mean to trigger you! Partnow: I think they did it just to hurt me…Do you still want to hear about pace or I have numbered you into submission? Iko: Of course not, this is brilliant stuff. You have the green light, the floor is yours. Partnow: Pace is actually a pretty big topic, but I’ll simplify. It’s not that playing fast as in running up and down the floor and maximizing possessions in a game is necessarily good. Genuine fast break attempts are good, but that’s more about being opportunistic in terms of turning offense into defense. Very few teams are consistently able to generate a lot of transition play off of opponent makes. The small ball Rockets are well equipped to play on the break in this way, as few players in the league have been as consistently impactful in terms of creating open floor chances as has Westbrook over his career. But what D’Antoni is talking about isn’t that, or isn’t just that. He spoke on this very point at the MIT Sloan Sports Analytics Conference a few years back and I’ll paraphrase his main thrust as saying that playing with pace doesn’t mean sprinting the ball up the floor every time, but it does mean getting the ball out of bounds, and getting it moving quickly. Get into an offense quickly. Make the defense work for more of the shot clock. Regardless of how fast a team plays in terms of possessions per game, playing quickly works. In the past, I’ve found that every second earlier a team can get into offense on the shot clock is worth about a point of offensive rating or 1 extra point per hundred possessions. NBA teams tend to operate in a narrow range in terms of average offense initiation, but there is still a second and a half to two seconds difference between the fastest and slowest teams. Two points of ORTG is a big deal, worth probably five wins a season! So, I’m not sure if that’s exactly what D’Antoni means by the importance of playing fast, but it’s why he’s correct to emphasize it. Iko: Assuming people’s eyes haven’t bled with all the basketball brilliance you’ve just spewed, I’ll leave you with this: Does this ultimately work? All the moves we’ve seen this team make in the last few months are to win a title this way. If we’re being realistic, the question can be changed to ‘Does this work against the LA teams in a seven-game series?’ That’s what this team will be judged on in the end. I know not to take much stock in regular-season games, but the 120-105 Clippers drubbing right before the shutdown didn’t look particularly uplifting for them. Partnow: Probably not? But at least given how it had looked at the time they made the Capela trade, playing “normally” was even less likely “to work” against the Angelinos, red or purple. This falls into that category of things which are destined to annoy analytics types, in that a strategy that makes a longshot win less of a longshot but still unlikely is seen as a failure if it doesn’t completely upend the laws of probability. But I’ll stop yelling at clouds and wrap this up. Iko: Thanks for delving into this with me, Seth. Any parting shots? Partnow: I think I’ve done quite enough here, let me stop while I’m ahead. Thanks for the convo!
It's not small-ball. It's good ball. If there is an awesome 7-0 Hakeem that walks through that door, DM will be more stoked than anyone. Until that time comes, we are going to play shorter guys that can actually play, rather than tall scrubs. GSW was playing Iggy and Draymond at the 4 and 5 and working teams over. We were playing Ariza and Tucker, and looked the best we have since 1994. Now we have Tucker and Lord Covington, and hopefully DM can add another skilled, tenacious front court player come next year. PLAY GAWDAMN THABO SEFOLOSHA. The little losing streak coincided with Mike playing JEFF GREEN over Thabo, despite Thabo being a key to our big Lakers victory. 8 rebounds, 2 steals and 1 block per36. We NEED that guy on the floor when Rob or PJ sit. He can still HOOP! His defense is NASTY.
Skill Ball is more accurate than small ball. Russ/Rivers/McLemore Harden/EGo RoCo/House/Demarre PJ/Thabo Green/Bruno/Tyson Clemons has no role on this team. Tyson barely has one -and even saying barely is generous. And Thabo cannot shoot anymore, not even a threat on offense. Replace with 3 guys who can shoot, dribble, pass as well as switch D. 7'0 Dragan Bender 6'9 Jonah Bolden 6'6 Justin Anderson 6'6 Theo Pinson Is it too late to replace our 2 way contract guys?
Micro ball will continue until the team philosophy changes, we get a new coach, or a big that is too good for MDA to not play shows up.
I have seen the Rockets use Small ball lineups with the success. A Micro ball lineups where they have used four players 6'4" and under on the court at the same time is going to be problematic against many of these playoff teams. In the playoffs when either Tucker and Covington get a breather, House, Green, Carroll and Caboclo are going to be key in bringing energy, hitting the open threes, providing some weak side help, getting the occasional block and of course helping to improve the teams rebounding. I am hoping both or one of Carroll or Caboclo can get into the rotation to join Tucker, Covington, House and Green stretching the floor and guarding multiple positions. Like the idea of trying to get either Jonah Bolden or Justin Anderson to get some time with the team and see if they will be able to contribute next season, as the Rockets may lose some of this season's rotational players in free agency at the end of the season.
this should be the last year of it. no way this team will be able to play a full season of it + playoffs next year.
It's frustrating to watch because the whole point of small ball is supposed to be that it's hard to defend. We've got playmaking guards who can beat people off the dribble and are great passers. But this strategy presumes that open 3s on every possession will be an unstoppable offense. And for whatever reason, for us, it very often is not. We are seeing teams use the defensive strategy of "help off shooters, pack the paint, don't allow layups, watch them shoot open 3s over and over and over again." And it's working. We were REALLY lucky to beat the mavs, and we were fortunate to beat the bucks in a close game. We could easily be 0-3 right now. But again, if teams are defending us this way, we're supposed to be blowing teams out. It's implied in our strategy that "watch them shoot open 3s over and over" is not a defense that works. I don't know what the answer is. I don't know if there is an answer other than changing the personnel. But teams giving us this many open 3s is just not supposed to work. With the amount of wide open looks, and especially with the amount of "standing in the gym practice shot" types of shots we're getting, this team should have a 125+ offensive rating every night. The offense was great against Dallas so I'm probably overreacting to just 2 games but whatever, just venting I guess.
harden was running on fumes before the restart, the whole team was. we were losing to the knicks, and getting blown out at home to the pistons and magic
It is already astonishing that mediocre career shooters like Jeff Green and House can shoot and have shot that well. They were basically picking up 'shooters' on the cheap with Tilman just giving out minimum funding for that. They just look like practice shots but if your adrenaline is running in the 4th,.......
Small ball works as an alternative but it can't be your only way to win a game. This is not basketball, it is a three point shooting competition without any other strategy. This way you can win some matches, but it is impossible to win a championship.
Not impossible if you have the right personnel. A Warriors team with Steph, Klay and Durant could certainly do it. Rockets just don't have the most optimal players to shoot 3s all the time.
