We created the worst basketball box score the NBA has ever seen

The main controversy we’ve had this NBA season, and it’s a foolproof method that we all agree on, judging by the inconvenience it’s caused me, is that ESPN changed its ratings format. At the start of the season, anyone with an ingrained habit of checking the ratings that prevents them from changing the site will have seen ESPN’s new display, which moves the scores next to the minutes, pushes the percentages to the end, and generally reframes the whole thing into an incomprehensible mess. This is an anger. Look at this monster:

This went on for about a month. Every time I clicked on the rubric, I would get this jump scare, forcing me to ask existential questions like “What the hell is a rubric?” Yet time and time again, I fell out of the habit, shamefully proving that I was no better than those cheese-loving lab rats. I’ve used just about every stats site; some numbers require Basketball Reference, some require RealGM, and some are best found on niche sites like pbpstats.com. But damn, the stats have always been for ESPN. For me, it’s always been like this.

advertise

Thankfully, sometimes, truly tragic stories like this end up with a hero. This is Tim Legler, our beloved co-host of the ALL NBA Podcast, who was the first podcaster and journalist to righteously speak out against these changes. As Legler said, “Why did they do that?! Why do I have to scroll through the entire page to find out which players were on the court shooting?!” These are valid and pertinent questions that bother me too.

Legler’s request worked, and ESPN quickly restored the score data to the old format. Naturally, Legler was hailed as a hero: Once again, both scoring and field goal percentage could be viewed simultaneously. According to his ESPN player page, Legler shot 43.1 percent from three-point range during his 10-year career. But in my eyes, he never failed.

But this brings me back to that hyperbolic question: What is a box score? Of course, there is another, more sinister one: How could we do something worse?

advertise

What exactly is a box score?

Harvey Pollack is considered the creator of modern scoring. He was a sportswriter and later public relations officer for the Philadelphia Warriors of the National Basketball Association. The team merged with the National Basketball League in 1949 to create the National Basketball Association as we know it today, and remained employed by Philadelphia until his death in 2015. As Pollack once told the Philadelphia Inquirer, “When I started working, the statistics were field goals, free throws, free throws, personal fouls and points. That was it.” Pollack held the scoring for Wilt Chamberlain’s 100-point game and was responsible for Chamberlain’s subsequent hoisting of the “100” sign, creating one of the league’s most iconic images.

Oh, if only Pollack knew what was coming. Today, every column can be betted on, “box score watcher” is a pejorative term for “numbers nerd,” and subscription sites like Cleaning the Glass can charge users for enhanced versions. With so much focus on statistics, the NBA introduced a weightlifting rule this year. Too many players don’t take 50-foot shots because they’re afraid of, yes, box scoring. And don’t you dare change the way it looks or we’ll riot.

Here’s my method for perusing box scores: Whenever I open a random box score, especially one from a completed game, I look to see who scored first. For a quick and dirty efficiency test, measure their field goal attempts against total points scored. Next, check the 3-point pen. This is an easy way to spot randomness; sometimes, a team may only hit its own jump shot and the opponent’s miss. For example, this season, when teams hit at least five more three-pointers than their opponents, their record is 95-39; when their three-point shooting percentage improves by at least 10%, their record is 98-26. Especially for surprising results, any extreme differences are worth noting. The only column more predictive of outcome than three-pointers is points scored.

advertise

After that, the most indicative statistics are those based on possession: Who makes more turnovers? Who has more offensive rebounds? ;Who gets more free throw opportunities? But to do this quickly, just look at which team attempted more shots. It’s not quite like a projection — teams with at least 10+ attempts are just 79-66 this year — but it generally sums up the numbers in one go. It might also just show which team drew more whistles, but thankfully there’s an easy way to check.

There are other important things that a score can’t convey, most obviously pace, and of course, it’s no substitute for watching the actual game. But we need to go through this to understand how I plan on producing the worst scoring in the league. I need you to understand what parts have value and what parts don’t, and how I plan to completely strip any order or usefulness from our beloved spreadsheets.

Let’s create the worst score ever

This has nothing to do with glitches, but they help inspire people. When ESPN reverted its score data to an old format, I suspect it interfered with a browser plug-in I was using that created this horrible image. Even though I only watched it one night, it still bothered me.

Then, last week, there was another example of box scores not scaling correctly to skinny browser windows. It makes me laugh, but I did it myself.

No, we want to make a perfectly functional box score most Typical statistics. We just wanted to make it really scary.

advertise

Our scoring father, Pollack, is credited with being the first to keep separate records of offensive and defensive rebounds. A common statistical mistake I see fans making all the time is blaming blowout losses on poor rebounding, forgetting that defensive rebounds are a statistic that only accrues when the defense forces missed shots. it yes An indicative stat: Teams with a rebounding lead of at least 10 are 80-25 this season. But the statistic of total rebounds tends to steal the nerve from a good offensive or defensive performance to gain those advantages. This may indicate that better rebounding was a factor in the results, but it’s not apparent from the totals. If you must, comparing offensive rebounds is a better metric.

So, with apologies to Pollack, let’s stop splitting the backboard again. Or, better yet, we just remove the offensive rebounding column. Statistics have chosen not to separate 2-point attempts from total field goals made; you can still count it, but only with some mental arithmetic. To get more valuable offensive rebounds, I’ll let you do the same.

I’ve taken the total score from the Denver Nuggets’ 115-106 win over the Charlotte Hornets on Sunday. This will be my ruined thing. So far, looking good.

Another overrated scoring check is assist-to-turnover ratio, at least in the sense that it provides little context about a player’s performance that night. Assists can be spotty and erratic, turnovers aren’t necessarily bad passes, and a heliocentric player tallying five turnovers doesn’t mean his team has a turnover problem, no matter how gaudy that number looks.

advertise

I don’t care. We’re trying to create something evil here. We’ll use assist-to-turnover ratio in place of assists and turnovers. If a player has a ratio of 2, it’s fun to figure out whether it’s 2 assists and 1 turnover, or 8 assists and 4 turnovers.

Also, we change the minutes to minutes every 48 minutes. I considered adding other minute-adjusted statistics; minutes per 48 minutes is the least intrusive. It’s worth understanding how any given player’s playing time affects the overall game.

Unfortunately, we have some players with infinite assist-to-turnover ratios, but that’s just how the math works. In case you’re wondering, Louis King’s 2021-22 season with the Sacramento Kings was the longest season played by a player with an unlimited assist-to-turnover ratio: He dished out nine assists and zero turnovers in 104 minutes. What’s even more shocking is that Louis King wasn’t a player from the 1980s, but that’s not the point. He even attempted 45 shots that season, but never failed.

advertise

In the analytics world, more than ever, they’re combining steals and blocks and calling them stocks. It provides a better understanding of a player’s overall defensive disruption and eliminates those annoying situations where a block is recorded despite swiping down Feel It’s like stealing. I’m all for combining them. But wait: A player might poke away a loose ball, get around that player, dive to the floor, and then happen to touch out of bounds as he swallows the ball. This doesn’t even register in any meaningful way in the underlying scoring or play-by-play data. Why shouldn’t we graph deflection as well? We just released another column. Of course, the fact that deflections correlate almost exclusively with steals—a player gets one steal for every two deflections—shouldn’t mean we’ve chosen not to include this statistic in its more confusing form.

Since we can get the biases from an obscure tab on the NBA Stats page, we’ll add those as well.

Also, remember the rough efficiency test? Is a player scoring more points than he is shooting? This makes sense because getting free throws is a form of shot control; star players often make up for poor shooting nights by shooting free throws. But flagrant fouls and technical free throws sometimes give stars extra points that have nothing to do with their scoring. Let’s separate the number of free throws in these situations from the total. This is the next rendition of our box score.

While the Nuggets didn’t attempt any technical or foul shots in Sunday’s game, it’s nice to know their scoring wasn’t inflated. This is probably the most important addition to the score yet.

advertise

Fouls can be misleading. Someone might commit two intentional fouls in the final minute of a game, get ejected, and never sit a minute because of foul trouble. The same could be said for a star who played lower minutes in the first half, who scored three straight quick shots but was unscathed the rest of the game. To better understand whether the foul issue affected a player’s game, we’d better change it to a personal foul at halftime.

Now let’s change the 3-point shot to a 2-point shot. This is a very useful and predictive statistic for the Frankenstein we are creating. We’re just giving it the same 2 treatment it’s been given around the world for decades.

I would love to add the number of jumps per player per game. In case you were wondering, Peyton Watson led Denver with 44 points while Bruce Brown only scored 12 points. However, I could not find any indicative correlation between the number of jumps and the results, so we excluded it. The same goes for falling. Yeah, like the number of times a player falls to the ground. Tim Hardaway Jr. had three points, but Brandon Miller had a game-high four points. But we’ll leave them out.

advertise

I think we do a good job with those statistics and columns. The only thing we have to do is disrupt it in the most unpleasant way possible, and I’ll happily do that. This is a mutation of our beloved; these mental arithmetic columns cannot exist next to each other.

Every stat available in traditional box scoring, assuming you flip the ball at least once to avoid infinite ratios, is technically still available in this revamped version. (And, I guess, second-half fouls, but when do those matter.) We even added deflections as a more realistic way to track damage, never mind that the current steals stat does a pretty good job of that. Now you’ll never wonder if Nikola Jokic and his 4-for-4 shooting from the free throw line benefited from shots he didn’t get from a shooting foul. I just checked the technical free throw column. He did not score in Sunday’s game.

I would like to sincerely apologize to the Denver Nuggets, Tim Legler, and Harvey Pollack for this experiment. ESPN, please don’t read this.