There's a tendency for some in the sports world to hear the word "analytics" and tune out. They might use the old expression, "There's lies, damn lies and statistics" to support their stance. Or, they might just say analytics come from a bunch of nerds that don't understand the game.
Here's an example to help illustrate why that stance is wrong, and why you want to utilize NBA analytics in your life:
On Friday, Kevin Durant joined Kyrie Irving on the sidelines for the Nets due to a shoulder injury. With their MVP sidelined, could the Nets still beat the Magic on Friday night? And by how much? According to Caesars Sportsbook, the Nets were still favored over the Magic by 9.5 points. Should they be?
These type of questions are perfect for the family of Basketball Player Index (BPI) analytics to weigh in on. Unlike many other team- and game-evaluation models, BPI is built from the player up. So, when MVP-caliber players are missing and the team looks entirely different, BPI still has a mechanism to estimate what the team might look like. So, what did BPI say?
According to the BPI Game Predictor, the Nets should've still been favored to win, but by a smaller margin than their 9.5-point spread. Thus, per BPI, with Durant out the Magic might've made good Best Bets candidates to win with the points against the favored but injured Nets.
Fast forward to the game itself. The Magic zoomed out to a 16-point lead after one quarter, but the Nets stormed back to lead by seven after three quarters. After a seesaw game, the final tally: Nets 115, Magic 113. Those that noted the BPI estimates and took the Magic with the points were happy that night.
This was just one example, but the BPI family of stats and projections are available for every game in the NBA all season. BPI estimates the caliber of every NBA team at any given time, for both the present and the future, and helps quantify what we might expect for the team over the course of the season.
What is BPI? Building BPI from RPM
"Allow me to re-introduce myself..." -- Jay-Z
ESPN has provided BPI as a method of evaluating NBA teams for years, but this offseason we tinkered with the recipe to optimize performance. Today, we re-introduce a new-and-improved BPI that generates an entire library of predictive and evaluation tools.
But, how does the new BPI work?
I'm glad you asked.
Many team evaluation and predictive metrics rely on a macroscopic view, factoring in how teams have performed as a unit with measures that track team offensive scoring efficiency, team defensive scoring efficiency, team net scoring margin, etc. While BPI does account for these elements, it goes a step further by also factoring in the individual player contributions that make up each team. The BPI calculation utilizes Real Plus Minus (RPM) to quantify the impact of every player on every team, essentially as a measure of the amount of talent on a given team.
This gives BPI a granularity, a way to measure incremental changes in a team's outlook based on personnel, injuries, trades and player availability. For example, consider the situations of the Brooklyn Nets, Golden State Warriors and the Philadelphia 76ers:
The Nets would have begun the season with the top BPI score in the league of 5.66 if they would have their entire team playing. With Irving not playing due to New York City's COVID-19 vaccine mandate, the Nets' BPI immediately dropped to fourth-best and they currently have the seventh-highest BPI in the league at 2.6.
The Warriors, on the other hand, have ascended to the top BPI in the league based on a combination of their excellent start to the season as well as an offense led by Stephen Curry, by-far the highest rated player in the NBA according to RPM. While regular season BPI reflects the current make-up of the team, Playoffs BPI is calculated using the players expected to be on the team when the playoffs arrive. The Warriors are still missing injured star Klay Thompson, who is expected back in the near future. Thompson is expected to have a positive impact on the team, with a positive RPM. Thus, the Warriors' Playoffs BPI is even higher than their top-rated regular season BPI by 0.2 points.
The 76ers are currently missing both of their All Stars, with Joel Embiid sidelined with COVID-19 while Ben Simmons continues his season-long absence. Their BPI reflects their current lineup, missing both players, and has fallen to eighth in the league. But, they are projected to have Embiid and potentially Simmons back for the playoffs, so the 76ers' Playoffs BPI is significantly higher than their current BPI, a full 0.6 points higher that's the difference between the eighth-highest regular season BPI and the fifth-highest Playoffs BPI.
Always learning and improving
"You must learn!" -- KRS-1, Boogie Down Productions
The new BPI calculation also includes a progressive factor to keep track of how team performance differs from the expectations of RPM. This Talent Adjustment coefficient is measured iteratively, after every game of the season, allowing BPI to learn from experience and generate more accurate predictions over time. Effectively, BPI is a sum of Talent (as measured by RPM) and Talent Adjustments that gets better as the season progresses.
The Talent Adjustment coefficient also contains information that helps explain some of the narratives that we observe on the court. Take, for example, the Suns' offense and the Warriors' defense. Based on their offensive talent as measured by RPM, the Suns would be expected to play like the 23rd-best offense in the league. However, so far this season, they have played like the 9th-best offense. That difference (23rd vs 9th) is what makes the Suns' Offensive Talent Adjustment the highest offensive adjustment in the league. Why might this be? Well, of course, the Suns' offense is led by Chris Paul, the maestro of his generation. Ever since Paul arrived with the Suns, we've worn out the narrative that his leadership and floor generalship allow him to maximize the production of the Suns' offensive unit. It turns out that there is tangible data to support the notion that the Suns' Paul-led offense is better than the sum of its parts.
Similarly, while the Warriors are known for their offense, they also sport one of the two best defensive BPI scores in the league and their Defensive Talent Adjustment is by-far the highest in the league, almost twice as high as the second-place Bucks. Again, this suggests that their team defense is better than the talent on the team supports. And, again, this unit is led by one of the greatest leaders and anchors in the NBA. Draymond Green has one Defensive Player of the Year award on the mantle, but his brand of aggressive, vocal defense that's heavy on helping, switching and disrupting all action from the 3-point line-in makes him arguably the best defensive player of his generation. Like Paul on offense, Draymond's defensive impact extends beyond the box scores and has helped turn average defensive talent into one of the best units in the league. A notion, again, supported by the Talent Adjustment coefficient.
Making BPI work for you
"Speeches, only reaches those That already know about it This is how we go about it, bout it" -- Outkast, Humble Mumble
While those that are already into NBA analytics might be excited to explore the new BPI family of stats, and their outcomes, we hope that many of you that may be just discovering BPI can also get excellent use out of it. There are, after all, so many ways that BPI can help improve your basketball experience.
I've already pointed out some sports gambling examples, where BPI can help with everything from daily wagers against the spread to season-long futures picks. I utilize BPI heavily when writing NBA Best Bets, and last season my Best Bets picks were correct more than 59% of the time in the regular season and 56.5% of the time in the playoffs. This success rate is, shall we say, potentially very rewarding. And with the new BPI calculation, this season's Best Bets may be even better.
For futures, BPI can help estimate playoffs and title chances for each team. For example, according to Caesars Sportsbook, the Nets are still heavy favorites to win the championship this season at +260. However, according to BPI, the Nets only have about a 4% chance to win this season. BPI also helps estimate the likeliest play-in matchups for both conferences, and even how a team's chances to makes the playoffs changes if they win/lose specific games down the stretch
You can use BPI to help win your barbershop arguments over what team is the best, even if you didn't wager on it. Winning the argument is its own reward, and supporting your arguments with analytics in addition to narrative can be enough to earn the victory in the court of public appeal.
You can use BPI in fantasy basketball and Daily Fantasy Sports (DFS) to evaluate the match-up difficulty and expected team scoring outputs to help determine what lineups you want to start on your teams. This type of next level analysis can be what puts you over the top to fantasy championships.
You can utilize BPI and RPM to evaluate trades, either hypothetical ones with the trade machine or real ones that shape team destinies. You can see, immediately, how that new free agent or draft pick might change your team's fortunes.
Bottom line and Best Bets for Wednesday
NBA action has always been faaaaantastic. But these days, there are more ways for all of us to participate in the game and take our experience to the next level. And that is the moral to the BPI story: however you enjoy basketball, and in whatever ways you enhance your NBA experience, utilizing BPI can help make that experience even better.
In fact... wanna have some fun?
Let's put BPI to the test with a mini Best Bets for tonight's big games that will be televised on ESPN:
Brooklyn Nets (-1.0, 216) at Boston Celtics
This shapes up as a close game between the Atlantic Division-leading Nets and the Celtics, who are in a three-way tie for second. Both teams have won three straight games; the Nets have won 11 of their last 13 with a 7.8 point average scoring margin while the Celtics have won eight of their last 11 with an 8.3 point scoring margin. The Celtics and their opponents have combined to score at least 216 points in five of their last seven home games, while the Nets and their opponents have combined for at least 216 points in four of their last six road games.
According to the BPI Game Predictor, the Celtics are slightly favored with a 52.9% chance to win outright and a projected score of 109.0-108.2. With a margin this close, if I made a pick at all I'd lean toward the team getting the points. The current Caesars Sportsbook line only has the Nets giving one point, which is essentially a pick'em, but I shopped around and found that FanDuel had the Nets giving 1.5 points. In addition, the BPI-projected total is just over the 216 line. While I might not normally bet with such close margins, today we're giving BPI a big challenge.
Pick: Celtics +1.5 (if I can get it), over 216
Philadelphia 76ers at Golden State Warriors (-10.5, 216)
The Warriors have been nigh unbeatable this season. After going undefeated even in the preseason, they've shot out to the best record in the NBA (15-2) and by-far the best scoring margin in the league at +13.5 points. They've won eight straight home games by an average of 20.6 points, and they've combined with their opponents to score more than 216 points in four of their last five home games with an average combined score of 227 points during that span.
The 76ers are still expected to be without Embiid, but the question is whether they will have any of Tobias Harris (hip), Seth Curry (back) or Danny Green (hamstring). All three are listed as game-time decisions, and are thus their potential of playing is included in the 76ers' current BPI.
According to the BPI Game Predictor, the Warriors have a 72.7% chance to win with an estimated scoring margin of 7.1 points (117.4 - 110.3), but that's assuming those three 76ers players are all available on Wednesday. In that case, the Best Bets prediction based on BPI would suggest taking the 76ers and the points. However, all three of those players missed at least the last game with their various injuries, and I anticipate that some subset of them will sit again on Wednesday. And even if they do play, there could be minutes restrictions that could affect their availability. This would reduce the game to the top-rated team according to BPI, at home, against an injured team whose game-time BPI would be significantly lower without potentially all five preseason starters.
Pick: Warriors -10.5, over 216