Part Four: The Offense
Ah, the offense. The offense.
I have been waiting to get to the offense.
While I have no conviction that the defense needs to change other than adding some tactical options for special situations, I have rather strong convictions about the offense. What I want to do to the offense is:
Burn it to the ground, then hover over it in case any survivors try to return and rebuild it.
The offense does not need to be reconsidered, it needs to be scrapped. Unlike other observers, I do not believe Coach Bennett needs to bring in an offensive guru. It is my belief that Coach Bennett has all of the knowledge and teaching ability to install a different offense. He does not cling to this offense because it is all that he knows how to coach; he clings to it because he believes in it.
He is wrong.
However right he might have been in the past to believe in this offense, the NCAA legislated that basis away in spring 2015 and then executed it into oblivion with the way the rules changes consistently have been enforced. The mover-blocker offense - his variation of which he calls "Sides" because the two posts are based on the two sides of the lane - is a wonderful design, but it has been legislated into oblivion. Coach Bennett clinging to the Sides offense after the NCAA adopted The Virginia Rules is the same as if Detroit had continued to invest in leaded gasoline engines after the EPA mandated the shift to unleaded gasoline.
A part of me really digs Tony's allegiance to Sides. It makes him a rebel, a renegade refusing to knuckle under to The Man. The sheriff comes around and lays down the law, but Tony stands his ground: "This is MY team! You can't make me run a high ball screen offense!"
But much as I love his Rebel Yell, there is truth in a different song:
I fought the law and the law won
I fought the law and the law won
The law wins, Tony. The law says take that ball to the rack, and if you don't take that ball to the rack, the law is not going to help you. If you want that federal money (whistles), you can't stick to your old ways (off-ball cuts).
While Justin Robinsons jump into our chests and get whistles, and Tyus Joneses and Grayson Allens toss their heads back like women in orgasm and get whistles, London Perrantes and Kyle Guy get hogtied, tossed into burlap sacks and hurled into the river while the cops stand there and eat donuts. You have a better chance of wading through the swamp without picking up a leech than you have cutting through the lane without two hands latched on to your jersey.
Forget pace. Pace is a non-issue. It's not pace, it's design. I don't want to use the word "style" because that can be misconstrued to infer an issue of aesthetics when it is an issue of substance. I have the following substantive criticisms of the Sides system:
Its basis is incompatible with the FOM Regime.
The Sides offense is based on screening and cutting away from the ball. Two posts on either side of the lane set screens for the players without the ball, giving them options for getting open to receive the ball.
The Freedom of Movement regime is dedicated to encouraging dribble attack. While the rules promulgated also mentioned freedom of movement for cutters, it has never - and I mean never - been enforced that way. All manner of incidental and minor contact on dribble drivers gets whistled - even contact initiated by the dribbler - while grabbing cutters has been called if anything less frequently than it was before. I can think of two games where more than one foul for grabbing a cutter were called: a game at Pitt and a Miami home game. Other than that, even one call in a game draws my attention.
The way the games are called demonstrates the NCAA's intent, because the enforcement is so consistent and so dichotomous.
The table below shows how Virginia's free throw rate and fouls drawn on opponents have declined since 2013-14, with a precipitous drop in shots at the rim over the last two seasons:
Free Throws and Fouls Drawn
|Year||UVA FTr||Rank||% Change||D-I Avg||% Change||Opp F/Gm||Rank||% Change||%atRim||Rank||% Change|
This is important because the refereeing is an important part of the game. Not just in the fouls you can pile up on the opponent and the free throws you can amass, but also in the ability of the opponent to defend you. Because the referees turn a blind eye to grabbing cutters and holding players coming around screens, it makes it easier to defend that type of offense. The corollary is also true: because the referees punish minor contact on dribble drivers, it makes it harder to defend that type of offense. You're metaphorically swimming upstream. If you're a salmon trying to procreate, it's what you have to do; if you're a basketball coach trying to win games you don't have the same biological imperative.
It produces the wrong kind of shots.
For several years now, the data has been pretty consistent that the top offenses get a lot of their shots at the rim and at the arc and do not rely on a lot of 2-point jumpers. Bart Torvik pointed out that none of this year's Final Four teams shoots even 21% of its shots from the mid-range. The blend between at the rim and from the arc varies, but they all either shoot the three or go to the rim.
The trend toward this rim-arc bipole has accelerated and it makes sense given the rules regime. TV wants exciting plays. The NCAA wants to keep TV happy. The referees call games to encourage driving to the rim. What's the best thing to do when the drive to the rim is cut off? Kick it out for a three. Michigan and Villanova do it very differently, but both offenses are geared toward getting players moving toward the rim and playing off of that. Michigan uses an offense based on cuts to the rim, and Villanova's offense is based on dribble penetration. Those are general principles; of course all offenses - including ours - use all different techniques.
The chart below shows UVA's split between at the rim, 2-point jumpers and 3-point shots from 2012 through this past season, compared to the D-I Median:
UVA Shot Distribution vs. Median
|Year||% AtRim||D-I Median||% 2Pt J||D-I Median||% 3FGA||%3FGA Rank||D-I Median|
This year's team shot the highest percentage of its shots as 2-point jumpers and the lowest percentage at the rim of any UVA team in the last five years. I was surprised to see that even 2017 shot more at the rim and fewer 2-point jumpers. You can see how UVA has bucked the national trend, and the chart above shows how the Hoos have paid for it at the free throw line.
One caveat to the at-rim vs. 2-point jumper statistic is that "2 point jumpers" is a very broad category that includes everything from runners in the lane out to those Shayok jumpers. So just because we shot more 2-point jumpers does not mean we shot as many long twos as last year. In fact, the Hoos shot 20% fewer long two-point jumpers than the previous season. The following chart derived from Synergy Sports data shows UVA's shot distribution for the last seven years. Because Synergy slices the shots up differently than Hoop-Math.com, the numbers in the categories won't quite sync, but the basic trends are still the same. I used the Hoop-Math.com figures above because the D-I median figures are easy to locate and show how Virginia compares to the NCAA general. Here we get more detail on how Virginia's shots break down:
Shot Distribution Over Time
|Year||Around Basket||Post-Up||Runner||Short 2 J||Long 2 J||3FGA||Total||Close||Mid-range|
The Cavaliers' offense has moved away from the basket dramatically. Interestingly, the trend started in 2016, with Gill, Tobey and Brogdon still on the roster. From 2012 to 2015, the offense moved closer to the rim, with the aggregate of "Around the basket" and "Post-up" shots going from 43% of shots to 50.7%. But then the trend reversed. In 2016, that number fell slightly to 47.4%, then tumbled to 38.5% and 34.9%. While reliance on three-point shooting has steadily increased in this period, it only absorbs half the decrease in power attempts. The combination of "Runners", "Short 2-pt Jumpers" and "Long 2-pt Jumpers" went from 21.2% of shots in 2015 to 23.4% in 2016 and then to 27.6% and 29.5% in 2017 and 2018.
Still, Virginia is taking way more of the lower efficiency mid-range shots than the prototypical championship offense. For comparison purposes, I looked at the four teams in this year's Final Four. While Michigan (29) and Loyola (68) had lower Kenpom AdjOEff ranks than UVA (21) entering the tournament, both teams shot the vast majority of their attempts from near the basket or beyond the arc. In fact, the Hoos took a larger percentage of 2pt jumpers than Michigan and Loyola combined!
Final Four Teams Shot Distribution
|Team||Around Basket||Post-Up||Runner||Short 2 J||Long 2 J||3FGA||Total||Close||Mid-Range|
The two-point jumpers have higher variance than at-rim shots and are worth less than 3-point shots. Even though this year's Cavaliers had a higher FG% on 2-point jumpers than the previous three UVA teams, it converted on barely more of them than of its 3-point attempts.
The two-point jumper is as susceptible to going cold as is the three-point shot, with only 67% of the potential reward.
The two-point jumper does not create contact that draws fouls.
The only value of a two-point jumper is that it can take what the defense is giving and can force the defense to cover more floor - if you hit it consistently. But it does not need to be more than a complementary percentage of shots. The best offenses keep that number below 25%.
And the two-point jumper is a natural output of the Sides offense.
It is too easy to defend without scoring bigs.
Does anybody dispute that a team has to get into the paint consistently to be able to be successful consistently? Creating three-point shots off of paint touches is a vital part of an efficient offense. Getting scoring from around the rim is as well. Scoring at the rim can only happen in one of three ways: from posting up, from penetration, or from offensive rebounds.
If you have bigs who can establish post position and score on post moves, or face up near the lane and attack off the bounce, your task of getting paint scoring is accomplished. If not, you need to get your guards into the paint or dominate the offensive boards (for putbacks!).
In Sides, when the big man is a threat to catch the pass and score, his defender has to maintain a good defensive position on him. If not, the post defender can help off the screener to cut off the curl. That allows the guard defender to lock into chase mode and run the cutter off the three-point line. With non-scoring posts near the lane setting screens, help defenders are nearby when guards try to penetrate.
[Having two non-scoring posts on the floor at the same time complicates success in any offense. Moving a non-scoring post farther from the lane on the weak side does nothing. The defender will sag. At least when a non-scoring post is setting a ball screen his defender has to come with him or the dribbler will be able to get great looks for three off the screen.]
Mike Tobey and Anthony Gill demanded defensive coverage. Jack Salt and Isaiah Wilkins did not. Mamadi Diakite might. I expect Jay Huff would.
It inhibits drawing fouls.
This one needs little expansion here, as it has been discussed at length above. As we saw, Virginia has been drawing progressively fewer fouls and shooting fewer free throws since 2014. I will just add that drawing fouls is a useful tactical objective. Pin a couple fouls on a big, or get a team near the bonus early in a half, and you impact their behavior. You can force a team out of aggressive defense and make them more tentative on their own offensive end. Our offense's inability to draw fouls takes away the ability to put this kind of pressure on defenses.
Players don't like playing it.
This last critique is more important to between seasons program building factors but can also impact quality of play. If you enjoy something, you're going to do it better. If you don't enjoy it but you have to do it, you approach it like a job or a chore. I hope it's not too controversial to assert that we tend to perform better when doing something we enjoy.
It is, however, important for recruiting and retention. If you and I are competing for the services of a network security genius, and you want him to do cutting edge stuff that poses a challenge and engages creative mind, and I want him to install virus scanners and troubleshoot firewalls, I'm going to have to offer him a hell of a lot in other areas to overcome that. Why is it so hard to accept that basketball players are going to be more likely to want to go where they'll be playing a style they enjoy? These kids have choices. And if a current player is not having fun playing your offense, if he starts to feel discontented about other things, he's more likely to decide he would enjoy himself more elsewhere.
I know, I know, "how do you know players don't like playing it?" I can't prove it, but I can say that when all three of the perimeter players who left your program in one year are quoted a few months later in the newspaper about how much they're going to enjoy "the freedom" of their new home's offense, I consider that a pretty good indicator that they were not having the time of their lives playing the offense they just left.
Then there are the recruits. Listen to the interviews hundreds of them give and you will rarely hear a recruit say "I want to walk the ball up and run back and forth along the baseline" or "I want to slide up and down the lane setting screens and occasionally get a pocket pass for a dunk." I can't say for certain that the lack of first-choice recruiting success is because of the offensive system, but I can say that recruits are not flocking to play in it. I am going to guess that it is a factor, only because there is so much negativity surrounding the system in the non-UVA-fan basketball world.
What to replace it with? I would love to see Tony use that offensive creativity his father told us a few years ago that Tony has and design an offense to rival the PacklineTM - a base pattern on coherent principles with a couple of special situation wrinkles. It doesn't need to be revolutionary; it just needs to be based on the right core principles. The five criticisms of Sides above show what I believe to be the right core principles (not to mention my article from October From Packline to Blitzkrieg):
Open the lane
My original draft said "spread the floor," but "open the lane" makes the point clearer. You can "spread the floor" and have a clogged lane. Put three great shooters on the arc and have two posts near the lane, and you have spread the floor but still crowded the lane. What you've done is open up the mid-range spaces on the floor. Since that is where the space is, that is where the path of least resistance is going to lead the shooters - especially when that is where the defense would prefer to see you shoot the ball. When you don't have post players who are threats to score in the post, you need to get them away from the lane. They are not helping you there.
Ideally, you only have as many posts on the floor as are threats to score. If you have a post who can shoot the three and a post who can score near the basket, you can play a 4-around-1 with them. If you only have one post who can score near the basket and none who is a threat from the arc, then just play with one post and put four perimeter players around the 1 post. If you have no posts who can score and you have to have one for defense, then that one post should be all over the floor setting screens and diving to the basket.
Getting non-scoring posts away from the lane opens it up for both penetration and for players who can effectively work the post. Even a mediocre post scorer can be effective if he can work one-on-one in space. Keeping the other four offensive players away from the lane makes the help come from farther away, which gives the post player more time to respond, and makes it easier for him to find an open man.
Getting non-scoring posts away from the lane also helps your shooters. If the opponent is going to run them off the arc, they can put the ball on the deck and not run into this:
The Michigan offense and the Villanova offense are built on radically different designs. Yet they have one critical thing in common. Michigan's offense is a derivative of the Princeton offense and is built on off-ball screening with movers cutting toward the basket. The Michigan offense and the Virginia offense share an emphasis on screening action away from the ball, but the difference is that in the Michigan offense the cutters generally cut toward the basket when they use the screens, while in the Virginia offense, the cutter is generally moving away from the basket. Please note the use of the modifier "generally," as each offense has elements common with the other (as do all offenses). We're talking about the most common pattern.
The Michigan offense, therefore, gets players moving "downhill" toward the basket as they attack. They also will use high ball screens and dribble attack (roughly 21% of attempts). And that brings us to the Villanova offense, which is a four-out 1-in motion offense using two post players. The Villanova offense relies on dribble drive to get downhill and either get to the rim or kick out for the three point shot. The Wildcats also feature the high ball screen (roughly 15% of attempts this season, 20% last season). Villanova's attack this season was built around the talents of Jalen Brunson, who can shoot the three, get to the rack at will, and also post up. The offense affords long-armed wings the chance to drive baseline, post-up, or use ball screens. Brunson and Mikael Bridges both excelled at pick-and-roll and post-up plays, giving Villanova a diverse offense.
After last season, Coach Bennett reportedly studied the Michigan offense. It makes sense, as it, too, is an offense based on off-ball screening and cutting. It, too, uses two post players. With its better spacing and use of downhill cuts, it could get Virginia more shots around the basket while opening up the arc.
The Villanova offense, with one post working the low blocks and the other post the "slot" area on the perimeter on the opposite side would open the lane. De'Andre Hunter fills the role of Bridges and Ty Jerome has much of Brunson's offensive versatility. Jay Huff has the skills to be the perimeter post player in this offense, and Mamadi Diakite could as well. Early in his freshman season he showed a confident three-point stroke. Hunter, obviously, would be well-suited to be the 4-man in this offense provided Virginia has three other perimeter players to put on the floor with him.
Whether Virginia chose to use an offense based off one of these (the Hoos have a 4-man motion package that they used at times this season), or some other, building the offense to feature downhill action should be one of the prime directives.
Emphasis on matchups and attacking
Ty Jerome knew that Donte DiVincenzo could not guard him. Recognizing that he had a matchup he could win, he forced an isolation and took him. If Wilkins had cleared out to the corner, Ty would have gotten an even easier shot, because Zay's man would not have been in position to cut off the drive.
Pro offenses are designed to find and use these mismatches. Good college offenses will do the same. When Jalen Brunson has a defender he knows cannot stop him, he forces the issue possession after possession.
identifying favorable mismatches and exploiting them can be a powerful weapon in an offense. It can create consistent offense and force an opponent to adjust in ways that open up opportunities for other players. Sometimes the mismatch is in the lineup, where the opposing player who will be assigned to defend your player cannot properly defend him. These mismatches can be identified in advance and game planned. Other times the mismatch will arise situationally - such as when a team switches on screens. These mismatches need to be identified on the fly.
Virginia's offense, by my observation and by the comments of other observers like Brad Franklin and Jamie Oakes, does not make use of matchups in any systematic fashion. The players have the freedom to use them, but the offense does not seek to identify or create mismatches to be exploited one-on-one. It's another area where the offensive system is not taking advantage of the talents of the players. Many times this season I saw a Virginia guard wind up with an opposing big man guarding him. Only occasionally did Virginia do what other teams automatically do when that situation arises: clear out and let the guard take the big man one-on-one. Sometimes they would, and the success rate was high, but far too often, the mismatch was allowed to pass.
The only time I can recall seeing Virginia consistently attack a mismatch was when Isaiah Wilkins took advantage of Stanford Robinson in the Rhode Island game. Other times when it would appear a player had a mismatch, the Hoos would not consistently go after it. I would have liked to see if Diakite could keep scoring in the post against UMBC. Couldn't Nigel Johnson consistently take some players to the hoop? De'Andre Hunter was a walking mismatch.
Hunter, in fact, was as we would expect an excellent isolation attacker. He scored 1.1 PPP in isolation plays, which put him in the 91% percentile for an "Excellent" rating on Synergy Sports. And he wasn't even best on the team at it. That distinction would go to Ty Jerome, who converted 1.14 PPP in isolation for a 93% "Excellent" score.
If those statistics surprised you, it also might surprise you to learn that Kyle Guy had an "Excellent" rating in the pick-and-roll (0.93 PPP for 83%ile), another situation the Hoos could look to exploit systematically. The pick-and-roll with Guy would be an excellent way to use Jack Salt in the offense. Why? Because Salt setting a screen for Guy would require Salt's defender to come all the way out to the arc. If he doesn't, Guy is going to come off the screen and pull up for a three every time. If he does, Guy still might have the open three, but he has other options. The lane will be opened for other players to make back door cuts. Driving off the screen could force a switch, then Guy finds himself at the top of the key with a center guarding him - one of those mismatches. Salt would wind up with a guard on him. Could he set up under the basket for a catch-turn-shoot?
The other area where Virginia has player excellence is on "Cuts." Mamadi Diakite was in the 90% range with 1.435 PPP on cuts and Jack Salt - yes, Jack Salt - converted 1.298 PPP for a "Very Good" rating. Open the lane (or the baseline), get the guards driving toward the hoop and those two bigs can be relied on to make the cut and convert.
Finally, while the "Post-Up" is one of the least efficient plays in basketball (as a finish - actions off the post-up can lead to other, more efficient plays), but Diakite showed improvement and promise for more improvement in this area. While Virginia in general has been terrible at the Post-Up the last two years, Diakite converted 0.72 PPP this season for an "Average" rating. This is up from his 0.368 PPP as a redshirt freshman. Given space in a 4-around-1 offense, Diakite could create favorable mismatches against some teams Virginia will face.
Some miscellaneous principles
Virginia's offense is guard-dominated. In many games, the guards seemed to monopolize the ball. In others, they appeared to make a concerted effort to pass the ball into the post. In these games, the offense appeared to work the best. Good things flow from passing into the post. It allow for several actions that Virginia often used to good effect. One of my principles for the offense would be that unless in a transition situation, no jump shots until the ball has been passed into the post once. Make sure the ball gets passed into the post every possession. Keep the post players involved and give those actions a chance.
Another principle would be when a player has a mismatch, isolate. Everybody else just get away from him, get out of the path between him and the basket, and let him work.
Another one I like is "no jump shots until the ball touches the paint." This one, like "pass it into the post" is designed to force good offense.
Take the baseline. Baseline drives lead to so many fouls that they are worth trying on every possession for that reason alone.
You get the idea.
Sides is an offense that works great against defenses that make bad decisions or don't want to work. Elements of it appear in all functional motion offenses. But it can become predictable, and like all "read-and-react" systems I've ever seen in any sport, it can be manipulable. If you as an opponent know the reads the offense (or a defense in football) is going to make, you can force the offense to do what you want it to do.
We've been arguing for offensive changes for three years now. Two years ago, Tony expressed admiration for the Golden State Warriors offense and hinted at more offensive creativity using spread offense. Last year, Tony admitted that the offense needed to improve and the staff hinted at VAF meetings that changes would be made. He reportedly studied the Michigan offense. The offense has remained largely the same, with the changes that were made gradually falling to the wayside as the season progressed. It is time while these players are here to change the offensive system for the present and future. It is time to install a championship offense. It is a fallacy that only this offensive system can succeed with the players we can recruit. Villanova and Michigan recruit in the same pool as we do. Loyola not even. They all are able to run offenses that avoid reliance on mid-range jumpers. There is no reason Virginia cannot do the same.