Scott Brooks

Wizards Staggering to Start Season

otto-porter-v-mem

The Wizards fired Randy Wittman for this? Six games into the tenure of Scott Brooks, the team sits 14th in the East with a 1-5 record. Washington’s futility is comprehensive — they rank 23rd in both offensive and defensive efficiency.

There are four key team stats that determine who wins and loses in the NBA. Here’s where the Wizards rank so far on offense:

  • Shooting (eFG): 23
  • Turnovers (tov%): 26
  • Offensive rebounding (oreb%): 12
  • Free throw rate (FTM/FGA): 16

On defense:

  • Shooting (defensive eFG): 30
  • Turnovers (defensive tov%): 9
  • Defensive rebounding (dreb%): 12
  • Free throw rate (dFTM/dFGA): 14

Don’t get too encouraged by their top ranking in defensive turnovers. Forcing turnovers isn’t necessarily an indicator of defensive effectiveness. In the NBA, defense is overwhelmingly about shot defense. And the Wizards are dead last in that category so far.

Remember the old days when Wizards fans wanted Wittman fired because the team took two-point jumpers instead of threes? Welcome to the new Wizards, same as the old. So far this season, they’re 28th in three-point attempt rate, but have attempted the fourth most two-point jumpers.

On defense, they’re still keeping opponents out of the paint (they have the fourth lowest defensive at-rim attempt rate), but they’re allowing the second highest opponent three-point attempt rate, and the worst opponent 3FG%.

Back to those four key stats for a moment: while there are four, they’re not created equal. Dean Oliver, who first wrote about these factors in a comprehensive manner, determined these approximate historic weights: shooting 40%, turnovers 25%, rebounding 20%, free throws 15%.

In recent years, those values have shifted, according to my analysis. Last season, shooting was worth about 55%, rebounding 18%, turnovers 15%, and free throws 12%.

This is a long and tortured way of saying the Wizards are bad where it matters most. Being worst in shooting differential and 25th in turnover differential overwhelms their decent rebounding and break-even free throw rate.

Player Production Average

Player Production Average (PPA) is an overall rating stat I developed that credits players for things they do that help a team win and debits them for things that hurt the cause. PPA is similar to other linear weight rating metrics such as John Hollinger’s PER, David Berri’s Wins Produced, Kevin Pelton’s VORP, and the granddaddy of them all, Dave Heeren’s TENDEX.

PPA is pace neutral, and weighs a player’s performance per possession against the performance of his competitors season by season. While PPA falls into the category of linear weight metrics, the actual values for each statistical category floats a bit from season to season based on league performance.

PPA is pace neutral, accounts for defense, and includes a “degree of difficulty” factor based on the level of competition a player faces while on the floor. In PPA, 100 is average, higher is better, and replacement level is 45. Read more here.

Want some good news, look at Otto Porter, John Wall, and (to a lesser extent) Marcin Gortat. The first two have been highly productive so far. Gortat is the team’s only other above-average performer through six games — although his production has been markedly lower than it was last season.

Wall and Porter offer an interesting contrast. Wall’s high PPA is built on volume — he makes LOTS of plays, both good and bad. He uses more than a third of the team’s possessions when he’s in the game, and he’s racking up rebounds, assists, steals and blocks at a prolific rate. He also sports an astronomical turnover rate (7.8 per 100 team possessions).

Porter is all efficiency. He makes the few shots he attempts. He grabs rebounds at a decent rate, plays solid defense, and avoids turnovers and fouls.

The bad news: everyone else. Markieff Morris and Bradley Beal have been terrible, the bench just about useless.

Ernie Grunfeld’s Plan B offseason acquisitions are off to a rough start. Ian Mahinmi is sidelined with injury, Trey Burke has been the next Eric Maynor (but worse), and Jason Smith has been…well…Jason Smith. Tomas Satoransky needs more time to figure out the NBA game.

The numbers:

PLAYER GMS MPG PPA
Sheldon McClellan 2 3.0 478
Otto Porter 6 34.3 173
John Wall 5 34.4 168
Marcin Gortat 6 36.2 135
Markieff Morris 6 34.0 67
Bradley Beal 6 34.8 66
Andrew Nicholson 5 14.0 33
Marcus Thornton 6 17.0 31
Tomas Satoransky 6 13.8 18
Kelly Oubre 5 15.0 18
Trey Burke 6 11.5 -48
Jason Smith 5 10.0 -93
Daniel Ochefu 1 4.0 -181
Advertisements

Wizards Remain Mediocre and Will Miss Playoffs for Second Straight Season

Oklahoma City Thunder v Memphis Grizzlies - Game Six

Yeah, I know the season is underway. Many teams have three games in the books; the Spurs already have four. This still serves as my Wizards preview, because while I’ve watched their first two contests (both losses), I’ve used nothing from the 2016-17 in the projection.

The approach this year is similar to the one I used for previous seasons: every player gets run through my statistical doppelganger machine, which spits out similar players from my historical database (similar production at similar age). There’s a process to weed out players with dissimilar career patterns — it makes no sense to compare a guy who stunk four years and suddenly had a terrific season to a guy like John Wall (for example) who’s been consistently quite good.

Once the list of “similars” is assembled, the system looks at the future of those players as a guide to the potential performance of the players being projected for the upcoming season. When the predicted performance (expressed in terms of Player Production Average — PPA for short) for each individual player has been completed, I estimate minutes (using an approach that must be similar to Kevin Pelton’s since the results were so similar). That gets translated into individual wins, which are totaled to team wins. Wins league-wide are capped at the number of wins available in a season (1230).

What’s new this year? Volume. For the first time, I projected the top 10-12 rotation players of every team. In previous seasons, I ran numbers for only the Wizards. This year — in a never-ending quest to make wrong predictions — I looked at everyone.

The Wizards

The Wizards spent two years hording cap space for an offseason in which nearly half the league would have sufficient room under the cap to pursue free agents with a maximum salary offer. The big prize was hometown hero Kevin Durant, who declined to even meet with the team. The team’s braintrust went after Al Horford (who signed in Boston) before managing to get Ian Mahinmi — a career backup coming off a career year who’s about to turn 30.

Their other roster moves were less inspiring: free agent deals for Andrew Nicholson and Jason Smith, and a trade for Trey Burke. They did manage to sign international guard Tomas Satoransky to a reasonable contract.

Here’s a quick look at what my projection system had to say about this year’s roster:

  • John Wall — Good news: Wall’s similars were a collection of very good players (albeit with a penchant for reputations that were better than their production). Bad news: half of the 10 most similar reached their career peak before age 26. More than half saw production declines following their age 25 season. Last season, Wall finished with a PPA of 144. Projected PPA: 130.
  • Bradley Beal — Beal’s persistent injury troubles overshadow what may be a bigger problem: his consistently mediocre play when he’s been on the floor. His PPA by season (average is 100 and higher is better): 92, 96, 99, 98. Players like Beal tended to peak at “decent starter,” not All-Star or All-NBA. The Wizards awarded him a max contract. Projected PPA: 108.
  • Otto Porter — Porter has improved during his career, and his future looks terrific (projected peak PPA would put him at All-Star level). But, the exercise in projecting the performance of individual players makes clear that it’s unwise to assume a young player will a) improve at all, b) that improvement will be linear, and c) that he’ll ever achieve imagined potential. Similar were useful defensive SF types who were also efficient on offense. But, there was no pattern of improvement after seasons most similar to Porter’s last year. So, Porter projects “about the same” as last year. Projected PPA: 127.
  • Markieff Morris — Last season, Ernie Grunfeld and Ted Leonsis swapped their first round pick in 2016 for Morris, who was deeply unhappy in Phoenix. What they got was a career mediocrity with little chance of getting better. The average peak of players like Morris (in Washington) last season was fairly low (acceptable starter level), and came (on average) at age 25.9. Morris is 27. Projected PPA: 95.
  • Marcin Gortat — The big man has been very good and consistent in Washington. He defied the decline I predicted for last season, and will have to do the same this year. At age 32, a drop in performance is probable — eight of the ten players most similar to Gortat declined the following season, and a ninth maintained. One oldster (Robert Parish) actually improved significantly in his age 35 season. I don’t anticipate something similar in Gortat’s age 32 season. Projected PPA: 147.
  • Trey Burke — The Wizards got him for next to nothing, which was the right price to pay. Burke started his career well below average, and has been less productive each year since. His comps were mostly backups who had short NBA careers. Surprisingly, Eric Maynor didn’t make the list. I’m actually predicting a modest improvement for Burke, although he’s unlikely to be close to what Ramon Sessions provided. Projected PPA: 67.
  • Tomas Satoransky — No comps for Satoransky since he didn’t play in the NBA last season. Although he has experience overseas, the NBA is the world’s most competitive sports league, and most players struggle to make the transition. Projected PPA: 65.
  • Kelly Oubre — The second year swingman seems to have abundant potential despite a horrific rookie season. Unfortunately, the history of players who performed like Oubre isn’t a pleasant one. Improvement was surprisingly modest (I double-checked the spreadsheet cells to make sure they were calculating correctly), and peaks were depressingly low. It’s worth mention that the same was true after Porter’s rookie year, although Porter had an injury. Projected PPA: 37.
  • Andrew Nicholson — The PF is coming off his best season (PPA: 81), which could mean he’s figured things out and is ready to become a useful backup, or…it could be the best he’ll ever play and he’ll recede to previous levels. His comps are useful backup types, and my projection suggests the latter. Projected PPA: 86.
  • Ian Mahinmi — When the Wizards whiffed on their other free agent targets, they turned to Mahinmi. It’s not exactly a bad contract under the league’s new financial realities, but it’s a #SoWizards kinda move. Mahinmi was a career backup who finally got a chance to start and responded with a career year. That’s good, right? Sure, except a) he’s going back to the bench in Washington (the team’s most productive player per possession the past few years (Gortat) plays the same position), and b) he’s about to turn 30. His “most similar” list is mostly journeyman centers. Some had high peaks, but few sustained it. What’s most likely is that he’ll be decent, but not nearly as good as he was last year. Projected PPA: 112.
  • Jason Smith — The decision to give Smith a multi-year deal was puzzling. He has a career PPA of 59, posted a 57 last season, and is 30 years old. It’s another #SoWizards move: no chance of meaningful contributions and no upside. It’s a nice lotto payout for Smith, though. Projected PPA: 50.

A potential wildcard: new head coach Scott Brooks. Previous coach Randy Wittman had his strengths, but would have ranked in the bottom third in the NBA. Brooks figures to be better, but the relevant research suggests the differences between professional coaches is pretty small. The exceptions are the very best and very worst coaches, but there’s a broad middle ground where coaches help a little or hurt a little, but don’t fundamentally alter their teams’ trajectories. While I think Brooks is an upgrade from Wittman, I also think they both occupy that middle ground.

Options

As I projected the entire league, I found that my process tended to push each team back towards the middle. The gap between the strongest team (Golden State) and the weakest (Phoenix) was about 26.7 wins. In recent years, the difference has been almost double that amount. So, I came up with an alternate method that ranked every team by their projected production, and then applied the average win total for that rank over the past five seasons.

The Wizards project to be ninth team in the East, and 19th in the NBA. Don’t go betting the mortgage, because my approach produced some results that are at odds with my gut and with predictions made by others I respect, such as:

  • My system likes Chicago and thinks the Bulls could finish as a top four team in the East.
  • Orlando projects to make the playoffs (7th seed).
  • Milwaukee and Atlanta both project to be worse than the Wizards.
  • In the West, my system likes Oklahoma City, Minnesota, Utah and Houston more than Portland.

For the Wizards, the win total from my projection system: 41.0. From the average record by league rank approach: 37.5. Take your pick.

My prediction: 41 wins and 9th place in the East.