clock menu more-arrow no yes

Filed under:

How does Deni Avdija compare with other young players, and what does that tell us about his future?

And a PPA update

Washington Wizards v Houston Rockets
Wizards forward Deni Avdija
Photo by Logan Riely/NBAE via Getty Images

One of the key questions the Wizards must answer is what they can reasonably expect from second year forward Deni Avdija. There’s considerable debate among observers of the team — in addition to being a likable personality who consistently gives an honest effort, Avdija has an army of passionate supporters who believe in his potential and his future.

The central thesis of making Avdija a central part of the team’s future (whether that’s as an All-Star level performer or a high-level role player) is that he’s playmaking forward and an elite defender. Among the arguments in Avdija’s favor:

  • He’s good for his age.
  • He’s good relative to his draft class and position.
  • His role in the team’s offensive is too small to show what he can do.

What I’m doing is turning each of these claims into questions, which I’ll attempt to answer using available data and analysis. Let’s start with those bulleted arguments.

How good is Avdija for his age?

Let’s start with the big picture. In his second season, Avdija is 21 years old. He doesn’t turn 22 until January 2023. While he did have two full seasons of professional basketball in Israel before coming to the NBA, he’s definitely a youngster.

Since the NCAA implemented the so-called one-and-done rule in 2005, my database has 971 seasons from players age 22 or younger who played at least 500 minutes. That group includes 564 player seasons from players age 21 or younger. I chose the one-and-done rule as the cutoff point to eliminate 21 and 22 year olds like Magic Johnson or Michael Jordan, who played two or three years of high-level college basketball before coming to the NBA.

Avdija’s second year score in my PPA metric (in PPA, 100 is average and higher is better — see below for more details on PPA) is 70. Keep in mind that PPA includes an accounting of defensive impact, as well as a player’s role on the team and his position.

The range of scores is broad. At the high end is Anthony Davis’ 266 PPA at age 21. At the low: -1 scores at age 20 for Dennis Schroder and Gary Harris. Those two at the very bottom underscore an important point to keep in mind — no matter how bad a young player looks early in his career, significant improvement is possible.

So, where does Avdija’s 70 PPA rank?

Among the 22 and under: 39th percentile. Other players at the same level: Lou Williams, Robert Swift, Sergey Karasev, Derrick Jones, Yi Jianlian, Dario Saric, Austin Daye, Luke Kennard, Jan Vesely.

If I limit the group to 21 and under, Avdija goes to the 42nd percentile.

Back to the point about young guys improving, here’s a sampling of 22 and under players who had seasons that ranked lower than Avdija’s 70 PPA so far this season: Victor Oladipo, T.J. Warren, Deron Williams, DeAndre Jordan, Derrick Favors, D’Angelo Russell, J.J. Redick, Marcus Smart, Terry Rozier, DeMarcus Cousins, Devin Booker, Jamal Murray, Jae Crowder, Aaron Gordon, Gordon Hayward, Kentavious Caldwell-Pope, DeMar DeRozan. That only gets us to the 25th percentile. In other words, a slow start to a young player’s career is not necessarily predictive of future success.

For this season, I have 78 players 22 and under who have at least 500 total minutes. Among that group, Avdija’s production ranks 52nd — about the 33rd percentile.

A similar analysis using other reputable all-in-one metrics produced similar results. The one outlier: 538’s RAPTOR, which ranks Avdija among the top 40 in each version of its model among 157 early career players (1-to-4 years of experience).

How good is Avdija relative to his draft class?

Avdija was the 9th overall pick in 2020. For this question, I’m focusing on second year performance for the draft class to eliminate bad rookie years (most rookies aren’t much good, most rookies improve from year one to year two, and Avdija has done both of those things). Lumping in rookie years would likely push Avdija down the list a ways — he had a near-replacement level 48 PPA last season.

Avdija and Detroit’s Saddiq Bey are the only second year players to appear in all of their team’s games so far this season. Avdija ranks 14th in minutes per game and 11th in total minutes this season.

Among second-year players with at least 500 minutes, eight rate average or better, according to PPA:

  1. Tyrese Haliburton, SAC/IND, 12th pick — 162
  2. LaMelo Ball, CHO, 3rd pick — 160
  3. Desmond Bane, MEM, 30th pick — 151
  4. Tyrese Maxey, PHI, 21st pick — 133
  5. Anthony Edwards, MIN, 1st pick — 125
  6. Jalen Smith, PHO/IND, 10th pick — 124
  7. Saddiq Bey, DET, 19th pick — 117
  8. Onyeka Okongwu, ATL, 110

So far this season, 32 players from the 2020 draft have gotten at least 500 minutes of playing time. Among that group, Avdija ranks 23rd in PPA. Ranking ahead of Avdija are 17 players selected after him. Only Killian Hayes was chosen ahead of him (7th overall) and ranks lower in production.

James Wiseman, the second overall pick, hasn’t played this season. He had a 53 PPA as a rookie.

If I look at total production, Avdija ranks 12th overall from this group. Isaiah Stewart and Kenyon Martin Jr. were picked after Avdija and are ahead of him in total production this season despite playing fewer minutes.

Is Avdija’s role in the offense too small?

An interesting question that doesn’t have a direct answer in the data. We don’t have to rely on recitations of skills or imagination, available data can give us information on outcomes from which we can make inferences.

Let’s start with the most basic and important measures of a player’s offense — usage and efficiency. For most basketball players, there’s a tradeoff between the two. The more possessions a player uses by shooting, getting fouled and shooting free throws, assisting or committing turnovers, the less efficient they tend to be. Current Wizards assistant coach Dean Oliver calculated that the average tradeoff is 1 point per possessions per point of usage above a player’s norm.

So, a player with an offensive rating of 110 with a usage rate of 20% would be expected to have a 109 ortg at 21% usage, or 111 at 19%. This is a general rule of thumb and not an ironclad rule of basketball science. There are exceptions that run in every direction, and the game’s greatest players defy this “rule of thumb” almost entirely.

For the Wizards, Kyle Kuzma appears to be one of the exceptions. When I analyzed his career, his efficiency doesn’t fluctuate much based on usage. He’s been below average in efficiency whether he’s using 25+% of the possessions as The Man, and when he’s used 18% or less as a role player.

The theory behind this effect is straightforward. Low usage players deal with less defensive attention. They take open shots, make easier passes, and give the ball up to teammates when pressured. High usage players force the action. They attack the defense, even when it’s loaded up to stop them. They take contested shots, draw double teams, make challenging reads and difficult passes.

Think John Wall vs. Tomas Satoransky. Wall was far from a perfect player, but he was dynamic. He forced the action and demanded defensive attention. Satoransky is a ball mover and facilitator. He’ll drive or shoot, but only when it’s open. He won’t force much of anything, and therefore doesn’t draw the same type of defensive attention.

Back on topic, in both his NBA seasons, Avdija’s been low usage and low efficiency. Average usage is 20%. In his rookie season, I estimated his usage rate at 11.6% and his relative offensive rating was 9 points per 100 possessions below league average.

This season, his usage is up to 15.2% and his relative efficiency is about 5 points per 100 possessions below average. While his efficiency has improved and his usage is up, the core challenge is his shooting. His conversion rates are below average on twos, threes and free throws. That should have the team concerned because his low usage rate suggests he’s getting a relatively easy diet of shots.

The usage, efficiency and shooting data do not suggest ramping up his usage is advisable at this point. If his efficiency followed Oliver’s rule of thumb, we could expect his efficiency to drop as he takes on a bigger challenge.

Is Avdija a playmaking forward?

So far this season, Avdija is averaging 3.7 assists per 100 possessions with an assist-to-turnover rate of 2.0. My database has 29 forwards age 22 and under with at least 500 minutes played. Among that group, Avdija ranks 8th in assists per 100 possessions and 5th in assist-to-turnover ratio.

Averages among that group:

  • 3.3 assists per 100 possessions
  • 1.5 ast/tov ratio

That can be viewed as at least somewhat encouraging — he’s likely a bit better than average among his position and age group peers as a playmaker.

Since the NBA is not based on age-group divisions, we do have to compare him to the league. This season, there are 133 forwards (PFs and SFs) with at least 500 minutes. Avdija ranks 53rd in assists per possession and 38th in ast/tov. Average for the group: 3.8 assists per 100 possessions and a 1.8 ast/tov rate.

NBA tracking data indicates the Wizards are already looking to him for some secondary ball handling. The only frontcourt players on the roster with more touches per game are Kristaps Porzingis and Kuzma. The same two are the only frontcourt players with a higher time of possession per game, and only Kuzma averages more time and dribbles per touch.

On the other hand, he’s dead last among the team’s frontcourt players in points per touch.

Let’s take a quick look at NBA tracking data on passing. With the caveat that working with the NBA’s information is maddening, Avdija produces 7.0 potential assists per 100 possessions. He may have had some bad luck — teammates are shooting 53.0% on his potential assists vs. 58.0% on all potential assists from frontcourt teammates.

If Avdija’s teammates converted his passes into assists at the same rate they do for his frontcourt compadres, I estimate that would boost his assist total by 12 (from 125 to 137), his assists per 100 possessions to 4.1 (from 3.7), and his offensive rating to 108. His PPA would go from 70 to 72.

On the other hand, the team’s conversion rate on all potential assists — including from guards — is 54.5%. Using that figure means, Avdija’s teammates shorted him on just 4 assists total so far this season.

There’s another idea floating out there that Avdija is a “connective” player who keeps the ball moving and flowing to open teammates. If it’s meaningful, it has to show in the data somewhere. If “connective” matters, I would expect to see it show in team shooting efficiency and offensive rating. And it does.

With Avdija on the floor, the team’s effective field goal percentage this season goes from 52.2% to 53.6%. This isn’t a huge difference, but it’s there and is surely not being caused by Avdija’s own shooting. The offensive rating overall goes up by 2.8 points per 100 possessions, even though the team’s offensive rebounding, assists and turnovers don’t change whether he’s on or off. So, it could be that Avdija’s “connective” role is helping boost teammate shooting a little, which is improving the overall offensive output.

The next place I’d anticipate to see “connective” data to show is in the secondary or hockey assist (this is the pass leading to the pass that’s tallied as an official assist). Avdija produces about half a hockey assist per 100 possessions and less than 1% of his overall passes. This is about the same rate as Thomas Bryant. In other words, the “connective” argument isn’t supported by these data.

That doesn’t mean this hockey assist information refutes the claim. It merely means it doesn’t support the claim. It could be that hockey assists aren’t a good measure of being a “connective” player. It could be that a connective player is making passes and movements that keep the team playing together in ways that show up in the team-level data and not in the player’s individual numbers.

Overall, NBA tracking data has Avdija ranked as follows on the team (stats per 100 possessions, unless otherwise stated):

  • Assists — 11th
  • Potential assists — 9th
  • Adjusted assists — 8th
  • Hockey assists — 11th
  • Assists / passes — 11th

I’ll definitely revisit the passing topic after the season when I have more time to delve into the NBA’s poorly designed tracking repository.

One last point on this topic: a friend suggested that Avdija could be a more productive passer if his teammates cut instead of standing around the perimeter. NBA tracking data says the Wizards are 8th in cuts possessions this season.

What about his elite defense?

The available data and metrics aren’t in full agreement about the quality of Avdija’s defense. 538 has him rated as one of the league’s better SF defenders. On/off data shows the Wizards have been 5.6 points per 100 possession better defensively with him out there — bad without him and ever so slightly below average with him.

ESPN’s Real Plus-Minus metric rates him a net negative on the defensive end.

NBA tracking data says opponents shoot 42.6% from the floor with Avdija identified as the closest defender vs. 45.8% in general. That’s a 3.2% difference, which is the best mark on the team. That’s a good number, period — it ranks 48in the league in a tie with Giannis Antetokounmpo.

He’s the team’s top non-center in contesting shots — only Daniel Gafford, Thomas Bryant, Montrezl Harrell, Kristaps Porzingis and Anthony Gill challenge more on a per possession basis.

The defense part of PPA says he’s a well-above average defender, though a bit short of “elite” status. I think he can become elite by cutting back on the fouling (his rate is down just 0.3 fouls per 100 possessions this season). Defense is surely the strongest part of his game at this point in his career.

Conclusions?

It’s too early in his career to say anything definitive, especially in light of the look at younger players through history. It’s fair to characterize him as behind his age-group and draft class peers, and behind the developmental curve of players his age in the one-and-done era.

That doesn’t mean he’ll stay there, of course. Maintaining the defense (and perhaps improving by fouling less often) while improving his shooting and ball handling is a clear course to becoming a quality NBA player. As is the case with many young players, the issue is whether he’ll put in the necessary work to improve those skills.

Player Production Average

Player Production Average (PPA) metric credits players for things they do that help a team win (scoring, rebounding, playmaking, defending) and dings them for things that hurt (missed shots, turnovers, bad defense, fouls), each in proper proportion to how much it contributes to winning or losing.

PPA is pace neutral, accounts for defense, and includes a “degree of difficulty” factor that rewards playing more difficult minutes. There’s also an accounting for role/position. In PPA, 100 is average, higher is better, and replacement level is 45. It usually takes a score of 225 or higher to be part of the MVP conversation.

The PPA score is not saying one player is “better” than another in terms of skill, ability, athleticism, or replaceability (if players hypothetically switched teams or were placed on a hypothetical average team). Rather, PPA shows production so far this season in terms of doing things that help teams win NBA games.

Wizards PPA through 71 games

PLAYER GMS MPG PREV PPA
PLAYER GMS MPG PREV PPA
Kristaps Porzingis 8 26.3 295 185
Daniel Gafford 62 20.2 148 144
Bradley Beal 40 36.0 117 116
Kyle Kuzma 66 33.4 114 107
Tomas Satoransky 11 15.5 133 97
Kentavious Caldwell-Pope 68 30.1 96 97
Rui Hachimura 31 20.2 80 90
Corey Kispert 66 21.9 74 82
Ish Smith 17 21.4 111 80
Raul Neto 62 20.8 65 74
Deni Avdija 71 23.4 73 70
Non-Rotation/No Longer with Team
Brad Wanamaker 1 27.0 188 186
Montrezl Harrell 46 24.3 130 130
Greg Monroe 2 9.0 131 129
Spencer Dinwiddie 44 30.2 102 101
Thomas Bryant 22 17.5 88 94
Anthony Gill 34 8.1 75 74
Aaron Holiday 41 16.2 68 67
Craig Sword 3 6.3 61 60
Isaiah Todd 10 4.1 59 59
Cassius Winston 5 5.4 32 32
Davis Bertans 34 14.7 32 32
Alize Johnson 3 6.0 24 24
Tremont Waters 1 8.0 15 15
Joel Ayayi 7 2.9 10 10
Jaime Echenique 1 3.0 0 0
Vernon Carey Jr. 1 1.0 0 0
Jordan Schakel 2 7.0 -107 -104
Jordan Goodwin 2 3.0 -217 -212