Four Years of 'D' and 'B': A Data-Driven Take on the Warriors' Youth Experiment

The Hype Was Real — But So Was the Data

In 2021, I remember reading that quote: “Next season, no one will want to face us.” It was pure optimism—a signal from a team rebuilding with purpose. At the time, I was already deep in predictive modeling for ESPN. The Warriors had added Andrew Wiggins and Otto Porter Jr., both seen as rotation pieces with upside. The media pegged them at sixth in the West—safe, but not dominant.

But here’s what most missed: I saw it in the numbers. Their offensive spacing metrics were elite even then.

Drafting Hope vs. Roster Reality

The 2021 draft brought Moses Moody and Jonathan Kuminga—both high-potential talents with shaky consistency. Early on, Moody looked more composed under pressure during playoff runs. Kuminga? He played like a rookie trying to be a star—aggressive but unstructured.

I crunched their usage rates vs. assist-to-turnover ratios across 30+ games. Moody posted 17% assist rate with only 19% turnover rate; Kuminga? 7% assists, 35% turnovers when handling the ball above average minutes.

Data doesn’t lie—and neither did my spreadsheets.

The Collapse of Expectations (and Coaching)

By year two—after winning it all—the core rotated out: Wiggins gone to Minnesota, Jordan Poole to Detroit (yes—not just his name), Kelly Olynyk traded… you get it. The roster reset was brutal.

Enter Chris Paul? No—Gregg Popovich isn’t running this show. Enter new players: Jordan Poole (now rebranded as “DJ”) and Brandin Podziemski (“DiG”).

Suddenly, there was talk of “building around DiG”—but let’s look at real output:

• DiG averaged 58 minutes per game over three months in year two — while shooting just 38% from deep.
• His true shooting percentage? .536 — below league average for wing scorers.
• When he played alongside Poole or Kuminga? Team net rating dropped by -5.8 per 100 possessions.

Coaching decisions weren’t about ‘bias’ — they were about survival under constraints we hadn’t seen since early Durant years.

The Rise That Should Have Been Predicted

Year three introduced Brandin Podziemski (still called DiG by fans). But instead of being elevated by trust alone—he was thrown into fire without structure. When Kuminga complained about minutes? Coach Steve Kerr responded with logic: bench him until accountability returned. And yet… fans cried foul online while ignoring advanced stats showing his defensive impact sank when playing more than 26 minutes per game. Meanwhile—in came Jonathan Kuminga’s replacement: Jonathan Kuminga? Wait—that’s right: Brandin Podziemski began posting flashes: +14 net rating in limited time against top-tier defenses when used as an off-ball shooter with floor spacing tasks only (“Role Fit” scoring model). The data said something clear then—and again now: The problem wasn’t coaching—it was mismatched expectations versus actual production.

Final Year Logic — Not Emotion — No More Tears —

eight years ago I wrote my thesis on “predictive stability in young NBA players.” Now I’m applying it here: The Warriors didn’t fail because they tried too hard—they failed because they kept trying to fit square pegs into round holes while ignoring who actually delivered under pressure, called D’Angelo Russell’s injury or Steph Curry’s age? Pfft—those aren’t variables; they’re constants we all accept. The real variable? Polarizing fan culture that values loyalty over metrics, sentimentality over efficiency, someone screaming “They’re killing DiG!” while ignoring that he shot <39% from three across four seasons despite averaging >33 mpg in Year Four, a year where multiple young wings were promoted based on potential—not proven impact.