A word of caution: Don’t get over-excited by the numbers below. Due to a small sample size and the diversity & difficulty of each qualified team’s different opponents in the first two rounds, we must take all the values with a grain of salt (always look for context and possible inflation of such values). All stats were collected before the start of the conference finals unless noted otherwise.

Scoring

We took a look at scoring trends for the four remaining teams and there are a few points worth mentioning. The Bucks scored a mind-boggling 48 points in the paint (44.5% of their total points scored), thanks to the dominant force that is Giannis Antetokounmpo, while the Clippers generated just 33.9% of their points close to the basket. This shows that inside scoring was, is, and will be an essential tool in all NBA offenses. 

Next, we checked out the distribution of assisted and unassisted made field goals. This part will give you a solid background in order to follow the play type numbers that come next. The main take away is that Phoenix includes lots of motion and ball movement, using their patented Spain PnR sets, resulting in assisting an outstanding 61.3% of their FGM. The other three teams do not assist at the same rate; in general, in the postseason, offenses are often more stagnant and the pace slows.

Four Factors

Although the Four Factors have become less and less popular through the years, they are still capable of providing a decent overview of a team’s offensive and defensive performance. The factors will be presented in order of significance.

The Clippers have vastly improved their shooting efficiency, mainly thanks to their 40.1 3P%. Players such as Reggie Jackson and Nicolas Batum have increased their playing time, as a result of being really effective jump shooters and spacing the floor for their two All-Star teammates. Phoenix is averaging just over 55 eFG%, whereas Atlanta and Milwaukee have struggled with their perimeter offense (35.2 & 31.2 3P%, respectively).

With the exception of Milwaukee, the other teams have been quite careful and precise in their decision making, on account of their prolific ball handlers (Trae-Kawhi-CP3). The lack of a traditional play-maker has not been a negative so far for the Bucks; however they must be vigilant from now on. Their big-3 (Giannis-Middleton-Jrue) commits almost two thirds of the team’s total turnovers (9.1 out of 13.6 TOV).

The ability to grab offensive rebounds and renew the team’s possession has obviously been affected by the soaring usage of smaller lineups. Indicatively, the Bucks took advantage of their size and crushed the offensive glass, especially in the series against Brooklyn. Indeed, they are the only qualified team to have improved on that element in the playoffs, compared to the regular season. The spread offenses of Atlanta and Phoenix naturally weren’t able to generate many second chance opportunities. As for the Clippers, they had a hard time grabbing offensive rebounds against bigger teams in Dallas and Utah.

As far as going to the free throw line is concerned, the Bucks have held opponents to the lowest playoff free throw rate, at least partially a result of their patented drop pick & roll coverage, which allows more pull up shots.  The Suns and Hawks visit the charity stripe at an almost identical frequency, while the Clippers have shown the biggest rise in free throw rate from the regular season to the playoffs (28.9, a 6.7 percentage point increase).  If we look at opponent free throw rate, Milwaukee also allows the least free throws per field goal attempt of the final four teams. Atlanta, on the other hand, has allowed more free throws in the playoffs; this comes as no surprise playing against Philadelphia and the offensive juggernaut of Embiid. (We also witnessed the hack-a-Simmons strategy.)

Play Type Data

As we have noted previously, it is important to remember that the playoff play type data is skewed from repeatedly playing the same two opponents in the first two playoff rounds. In addition, we are looking at a small sample of games and possessions. We should keep that in mind when comparing this data to its regular season counterpart compiled over a larger number of games against the entire league. Still, this data can help us see how the offenses are stylistically attacking and adjusting. To see what types of actions are typically classified under each play type, this explainer by Cranjis McBasketball for Nylon Calculus is very helpful.

Increased Isolation Play

Each of the four conference finalists have relied more on isolation play than they did in the regular season. The Clippers and Bucks have had the largest increase in the share of their isolation play; they have utilized isolation on 14.7% and 13.1% of their possessions, respectively. 

A rise in isolation play is common in the playoffs as defenses employ more switching schemes and possessions end up becoming late clock affairs more often. 

Isolation possessions are not typically super efficient, but the Phoenix Suns have bucked this trend in the playoffs so far. Led by efficient iso play from Paul and Booker, Phoenix scored 1.23 points per possession on this play type in the first two rounds. Devin Booker in particular has emerged as a greater threat here, scoring almost 1.2 points per iso possession in the playoffs compared to just 0.85 points per possession in the regular season. We should caution though that the sample is small- we are talking about only 63 total possessions for Phoenix in this play type.

In the conference finals, it will be interesting to see if the teams keep attacking 1-on-1. Milwaukee will probably utilize isolation less going against the Hawks rather than the switching Nets defense. The Suns will get more chances to attack perceived mismatches playing against a switch-heavy Clipper squad.

More Pick-and-Roll Ball Handling Finishing For Phoenix and Atlanta

The Hawks and Suns have each had pick-and-roll handlers finish more than 20% of their possessions, up from the regular season and a substantial part of their attack. Given the personnel on these squads and the amount of pick-and-roll they run, this makes sense. These are not the most efficient possessions typically, though Atlanta has managed to post 0.98 points per possession in the postseason. Phoenix is at 0.84 points per possession.

For the Hawks the pick-and-roll is of course the Trae Young show, and the young star has finished a preposterous 17.2 possessions per game as the ball handler. He has scored a respectable 1.02 points per possession, ok but not great overall offense but good for this play type and adequate in the half court. He is using about three more pick-and-roll ball handler possessions per game than in the regular season and scoring at a slightly higher rate. The Bucks’ defense is typically very concerned with stopping the ball handler from getting all the way to the rim and stopping the lob, so Young may get many floaters off the pick-and-roll in this series.

The Suns’ lower 0.84 points per possession on this play type surprised us as they posted almost 1 point per possession in the regular season and their offense has been very strong in the first two rounds. One possible explanation is that perhaps some offense generated from the pick-and-roll is being counted as an iso. There are grey areas in the classification of this type of data; this could also perhaps partially explain the Suns’ high isolation efficiency.

Spotups Important For All Four Conference Finalists

Spotup possessions are a big part of each of the remaining team’s arsenals. The Clippers, Suns, and Bucks have this as their most used play type in the playoffs and the Hawks have it as a close second. 

The high powered Clipper attack led all teams in regular season spotup efficiency at 1.24 points per possession, and in the playoffs they have become even more efficient. They posted 1.3 points per possession through the first two rounds. So many members of their roster are all hot at the same time; Reggie Jackson, Paul George, Terrance Mann, and Luke Kennard are over 1.4 points per possession (this is including the first two contests against Phoenix). If they want to keep their postseason run going with Leonard, they are going to have to keep setting up good looks and knocking down shots.

After a poor series of outside shooting against the Nets, the Bucks are down to 0.97 points per possession on spotups in the postseason, below their regular mark of 1.1 points per possession. Their supporting players are relied upon to make use of the defensive attention occupied by actions from Antetekounmpo and Middleton; they need to generate better offense on their opportunities off of the catch. 

The 2019 NBA Champs (doesn’t that feel like 10 year ago?) are only 17-19 at the All-Star break after starting the year a sluggish 2-8, but have actually climbed up to a positive +1.4 net rating, per basketball-reference. There are lots of intriguing developments with this team, so let’s dive in!

Stat 1: 54.2%. Pascal Siakam’s true shooting percentage so far this season.

Let me preface this section by saying that I am personally a big fan of Siakam and his game. A Cameroon native, the Raptor forward came to the NBA oozing raw talent despite having not played organized basketball until he was nearly 18, as described in this great piece by Jackie MacMullan of ESPN. Aesthetically, he’s a player I very much enjoy watching. His sweet spin moves on drives and his daring attacks of the basket in transition are appealing.

Siakam rose onto my radar in 2018-19 when he both expanded his offensive role and bumped up his true shooting percentage, a measure of scoring efficiency that takes into account shots from the floor and free throws, to a gaudy 62.8%. When Kawhi Leonard left the Raptors in the 2019 offseason, I was curious how Siakam might play in an even more expanded offensive role. The result was a drop-off in scoring efficiency, as shown in the graph below.

For context, Siakam’s 54.2% true shooting percentage is three percentage points below the league average of this season. His usage, a measure of his offensive workload, is up to 25% from about 21% in 2018-19.

I should note that this drop-off in efficiency that comes with an expanded offensive role is very understandable. In general, players tend to become less efficient as they are asked to take on a larger role and create more of their own shots. As evidence of his larger role as a solo creator, the Raptor wing’s isolations per game have doubled from 1.5 a game in 2018-19 to 3.2 a game in 2020-21, per Synergy. Moreover, the proportion of Siakam’s made field goals that are assisted has fallen; 49% of his 2-pointers and 98% of his 3-pointers were assisted in 2018-19 but just 37% of his 2-pointers and 83% of his 3-pointers have been assisted this season, per basketball-reference.

His 3-point stroke, a revelation in 2018-19 when he knocked down 36.9% of his shots beyond the arc, has deserted him so far this year and he sits at just 29.5%. Shooting can be fickle, so perhaps some improvement may be expected back up into the mid 30%’s. At the basket, he still makes almost 70% of his shots but these make up a smaller share of his total field goal attempts, per basketball-reference.

I should note that roughly average efficiency scoring on a high volume still has value to an offense. But part of the allure of Siakam came from his potential to combine both high volume and high efficiency. That has not happened yet, though at 26 and still relatively new to the game, the Toronto star still has time to continue to improve.

Stat 2: 2nd and 30th. The Raptors’ ranks in opponent turnover percentage and opponent FT/FGA, respectively.

When you watch Toronto play defensive, you notice a distinctive style. Defenders fly around the court, aggressively helping off their primary offensive assignment to attempt to cut off promising drives and snuff out openings. Often, the goal is to prevent juicy opponent opportunities at the rim. Not only do the Raptors’ defenders readily offer help, but they also actively double team ball handlers in the post or isolation more frequently than other teams.

This aggression manifests itself in lots of different ways, but one way in particular is forcing a bevy of turnovers at the expense of giving up a ton of free throws. The Raptors rank 2nd and 30th respectively in these two Four Factors, as seen in their placement in the top right corner of the plot below.

The turnovers generated are important and grant the Raptors opportunities that they relish to play in transition; Toronto is 2nd in transition frequency per Synergy. But on the other side of the coin, it can be hard to sustain a high level defense when you allow your opponents so many opportunities to generate efficient offense at the free throw line. Overall, Toronto’s defensive rating is a very middling 112.2, 0.2 points below league average per basketball-reference.

Beyond the free throws given up, the Raptors’ willingness to help off of shooters allows smart opponents the opportunity to generate a lot of open looks from behind the arc. Ball handlers driving to the basket can expect a swarm of help defenders to leave open shooters on the perimeter, including in the corners where 3-point shooters are generally the most deadly. The Raptors allow the 4th highest proportion of opponent field goal attempts from 3 and, to compound the issue, allow the 3rd most corner 3-point attempts per game, per Synergy.

Overall, I very much enjoy watching Toronto’s aggressive defense in action! It is a refreshing change of pace from many of the more passive schemes employed in the game today. In addition to their aggression, the Raptors employ a versatile mix of schemes as described here by Caitlin Cooper for Indy Cornrows and analyzed last season by J.A. Dubin and Krishna Narsu for FiveThirtyEight.

While their results this season have been just average on the defensive end, the Raptors were the 2nd best defense last season employing a scheme that was also very active, as measured by the Aggression+ metric that FiveThirtyEight unveiled last year. The Raptors ranked 2nd in Aggression+ this season as of February 16th, behind only the Hornets, per metric developer Krishna Narsu. So perhaps Toronto will right the ship defensively while sticking to their game plan.

One thing that does stand out when comparing this year’s squad to last year’s is that opponents shot a league low 33.7% on 3’s in 2019-20. That has regressed to 37.7% this year, per basketball-reference. So I suppose one could argue that Toronto was getting a bit fortunate last season (or unfortunate this season). In addition, last year’s Raptors were middle of the pack in allowing free throws, as opposed to their worst in the league mark this year. Regardless of the results, the Raptors’ defense certainly does things differently than many other NBA teams.

Stat 3: 4th. Chris Boucher’s rank in blocks in the entire NBA.

Speaking of defense, have you seen Chris Boucher play? The long, skinny 28-year-old former University of Oregon player has emerged as a shot blocking machine! He ranks 4th in the entire NBA in total blocks with 69, per basketball-reference, despite only playing 24 minutes a game in a bench role. He trails only rim-protecting stalwarts Myles Turner, Rudy Gobert, and Clint Capella. Check out these beauties:

You can see an example in the last clip of Boucher’s preternatural talent for extending to the perimeter and blocking 3-point shots. In fact, as of late February, Boucher was recording the most 3-point blocks per 100 possessions in league history according to pbpstats.com data pulled by Bryan Kalbrosky of HoopsHype. I think shooters sometimes don’t realize quite how much ground he can cover with his length.

On the offensive end, Boucher has blossomed into a dangerous pick-and-pop threat. Though his mechanics look a bit unconventional with the way he brings the ball over his head, the Raptor center has knocked down 44.5% of his 137 3-point attempts this season. While overall 43% of his field goal attempts have come from behind the arc, he demonstrates some ability as a pick-and-roll roll man as well.

Overall, Chris Boucher has emerged onto the scene as a formidable option off the bench in his first NBA season averaging over 20 minutes per game. I am excited to watch his development and am intrigued to see if his minutes will continue to increase. In the meantime, the block party will continue.

For our first week, we will focus on the three teams atop the Eastern Conference.

Stat 1: 1.16. The points per possession for the Brooklyn Nets on isolation possessions. (source: Synergy)

Sometimes a stat is counterintuitive. You look at it and it does not quite agree with your assessment from watching the games. But other times a number is very understandable, and this is one of those times.

When you have Kevin Durant, James Harden, and Kyrie Irving all on one basketball team together, it stands to reason that that team will be very good at isolation ball. The chart below puts the 1.16 points per possession figure into context. Not only are the Nets the most efficient isolation team of this season, but they are first in efficiency for all seasons since 2015-16!

We can see from this chart that the Nets don’t utilize isolation ball as much as the outlier 2018-2020 Houston Rockets (practitioners of what ESPN’s Zach Lowe eloquently described as James Ball), but it does consume 9.6% of their possessions. They are narrowly second in terms of frequency this year, behind only Damian Lillard’s Portland Trail Blazers.

The Nets’ ability to put the ball in the hands of any one of three elite individual scorers has not only powered Brooklyn to a tie for the #1 offense in the league through February; it has also made commentators consider them an especially dangerous playoff team. They sit as the second most likely team to win the NBA title, behind only the defending champion Los Angeles Lakers, on the Action Network’s consensus of NBA futures. When teams bear down on defense during the playoffs and force their opponents to rely on less transition and movement, the Nets are perhaps poised to not drop off as much as their rivals. (This is at least one plausible line of thinking. I haven’t studied playoff offense in general.)

I should note that it is a bit easier to put up record numbers for rate stats over smaller samples, so we should check back and see where Brooklyn stands at the end of the regular season. In the meantime, let’s keep watching KD, Kyrie and The Beard cook.

Stat 2: 39.4%. The Milwaukee Bucks 3-Point Percentage. (source: Basketball-Reference)

The Bucks are kind of a strange team for me this year. Over the previous two regular seasons, they built a juggernaut reliant on a suffocating defense that snuffed out opponent baskets at the rim while letting the opposition chuck away from behind the arc. Their offenses were good, but not as good as their defenses.

This season, Milwaukee has flipped the script by posting an offensive rating that is tied for best in the league while their defense has been more middle of the pack. Overall, it translates to the Bucks (maybe sneakily?) being really good again. Only the surging Jazz have a better net rating through February.

What jumps out when you look at their offensive stats is their sweet 39.4% 3-point percentage, fourth best in the association. The Bucks are not afraid to hoist them either; 41% of their field goal attempts have come from behind the arc per basketball-reference. So Milwaukee benefits immensely by being able to make their outside shots.

What is striking about how well Milwaukee is shooting the ball is that they were actually only middle of the pack last year. They have had the third largest jump in 3-point percentage over last year, as the chart below shows.

When you take a look at the individual shooting percentages in Milwaukee, you find career years all over the place. Khris Middleton, an All-Star snub (but that is for a different post), is shooting 43.4% on a team leading 182 attempts. Donte DiVencenzo is up to a sweet 39.4% after shooting only 33.6% last year on fewer attempts per game. Bryn Forbes, a deadeye former Spur, has made a gaudy 45.4% of his attempts, also a career high. Pat Connaughton and Jrue Holliday are in the high 30’s as well. Can everyone be hot at the same time?

A natural next question is are the Bucks getting especially good looks this year? The answer is not particularly, especially when compared to last season. The proportion of the Bucks’ 3-point attempts that have been “wide-open”, where the nearest defender is 6+ feet away, is actually down this season to 53.7% from 54.9% last season, per Synergy.

While it is of course quite the luxury to be shooting the ball this well, I think the Bucks’ strong 3-point shooting could be seen as a bit of negative indicator for their future performance. While we like to focus on the randomness of defensive 3-point percentage for good reason, offensive 3-point percentage is actually a stat that can fluctuate as well, although usually not to the same extent. In fact, the correlation between 1st half and 2nd half offensive 3-point percentage is actually only 0.33, lower than for other offensive statistics. If these first ~30 games prove to be a high watermark in terms of shooting, the Bucks’ offense could take its foot off of the gas a little in the second half.

Stat 3: 11.1. Joel Embiid Free Throws Made Per 36 Minutes. (source: Basketball-Reference)

Joel Embiid, the Philadelphia 76ers’ superstar center, dwarfs Trae Young in more ways than one. Not only does the 7-foot center tower over Young in stature, but he has a sizable lead in the competition for the free throw shooting title. On a per 36 minutes basis, Embiid leads the NBA with 11.1 made free throws; Young’s 8.9 ranks second. (As an aside, Trae being second here is an achievement unto itself.)

Embiid’s ability to get to the free throw line and make his freebies has powered a stellar offensive year that has him posting career highs in stuff like true shooting % and points per game. It should not be taken for granted too that the 76er center is shooting an impressive 85.6% from the line. If Embiid was shooting a more average 75% from the stripe, the 76ers would actually lose about 1.1 points per game given his high volume. (1.1 points per game is worth a little more than 2.5 games over a full 82.)

Moreover, the 76ers offense has floundered without their star center on the court. Per pbpstats.com, the 76ers’ offense falls by a gigantic 16.5 points per 100 possessions when Embiid sits in those games that he plays. While on-off stats are team stats and need to be treated with caution when applied to an individual, I think we can attribute meaning to this given Embiid’s large role in the offense. It is no wonder that Embiid has burst front and center into the MVP conversation!

It finally happened. On Wednesday, the Nets pulled the trigger on a blockbluster four team trade that brought James Harden to Brooklyn at the cost of Caris LeVert, Jarrett Allen, Taurean Prince, Rodions Kurucs, three unprotected first round draft picks and four unprotected first round pick swaps.

In this post, I want to analyze the implications of this trade from the Nets’ perspective only. A fair warning before I dive in: I will get a bit philosophical here and use this fascinating deal as a stepping stone to introduce a framework I like to use to think about team title potential. And I will fully admit up front that I find this trade exceedingly difficult to evaluate and assign a tidy grade to the Brooklyn front office.

A Framework For Thinking About Title Potential

Bear with me here, but let’s begin by zooming out and trying to put some numbers to the Nets’ title odds. At the end of the day Brooklyn is buying championship equity here, particularly championship equity in the 2020-21 and 2021-22 seasons (the two years where Durant, Irving, and Harden are all under contract).

How much championship equity? Well, we can look at betting markets to give us some rough idea. According to Forbes, BetOnline raised the Nets’ title odds for this season from +650 pre-trade to +275 in the aftermath. Translating these lines into percentages, this works out to a roughly 27% chance of title post trade vs. 13% pre-trade. We have went from needing three fair coin flips in a row to come up heads to two fair coin flips- not bad at all!

At this point, you may be rolling your eyes and saying betting markets are flawed and Brooklyn is over-valued after a splashy trade, so what are you doing here?, but indulge me for a moment. Let’s say our estimate for Brooklyn’s 2021 title odds jumped from 13% to 27%. And now let’s continue this thought experiment and think about the Nets’ title potential over the next 10 years, while also considering the counterfactual where Brooklyn does not make the Harden trade.

There are a couple of pieces to this potential projection. The first essential piece to evaluating this trade is to ballpark how much title equity the Nets bought for themselves in the 2020-21 and 2021-22 seasons. We already discussed what a plausible estimate for this year might look like, but how about next year? Well, Harden and Durant would be in their age 32 and 33 seasons respectively so decline (or injury) is certainly plausible. But we have to still consider that the pair could each be top 10 players next season, a recipe for title relevance. So an 18% title chance in 2021-22, down from 27% for this season to account for a bit more uncertainty a year out, seems reasonable.

Now, to fully calculate the championship equity Brooklyn gained over the next two seasons we need to estimate their title potential if they had not made the trade. For this season we can take the 13% chance at a title from the pre-trade betting market. How about next season? This is not an easy question, but let’s use the general principle of regressing the Nets a bit more towards league average in the face of future uncertainty and say a Harden-less Brooklyn would have had a 10% chance at a title in 2021-22.

If we take these estimates at face value than the Nets have just bought themselves 0.22 expected championships over the next two seasons. Is that good or bad? On the surface, in a league with 30 teams and one ultimate prize, that number seems pretty good.

But we need to consider the costs of the trade. The Nets have dealt away three promising young players and over a half decade worth of first round picks. There is certainly pain to be felt.

Let’s try to quantify that pain by looking at Brooklyn’s title potential from 2023 to 2030 (as the graph shows). We are looking out far into the future and I don’t have a magic crystal ball so this may be a fool’s errand. We do have one guiding principle though- if you go far enough into the future then, all else being equal, my default championship odds for a team are 1 out of 30 (or 3.3%). This is because a team’s record in any given season has little relationship to its record 4 or 5 years down the line. In other words, a decent estimate for the champion five years from now is to pick a team at random out of a hat.

With that bit of explanation, we can see why I give the Nets a 3.3% chance of winning a title in each of the 7 seasons from 2023-24 to 2029-30 if they did not make the Harden trade. But now, without control of their first round draft pick until 2028, let’s give them a 1% chance of the title from 2023-24 to 2026-28 and a 3.3% chance again in the two years after that.

If you add up the championship equity lost in the 7 seasons from 2023-24 to 2029-30, we are estimating that the Nets are losing roughly 0.11 expected championships. So in this plausible scenario (that is still certainly just a rough estimate) we find that Brooklyn has gained about 0.11 expected titles over the next 10 years. So is the risk worth the reward??

Well, perhaps we should be more of a skeptic about the upside of this trade. The Nets lost two key pieces in Caris LeVert and Jarrett Allen that will hurt their title pursuit over the next two years. Depth suddenly becomes an issue. The defense could simply not be good enough (a very real concern). The big rotation certainly looks ultra thin absent any other moves. And do the duplication of strengths of Brooklyn’s big three (they all thrive as on ball scorers) minimize the team’s upside?

A skeptic could formalize their skepticism by saying that this year’s team has an 18% chance of winning it all, not 27%. And next year they could peg Brooklyn at 13%, not 18%. Then the math no longer adds up- the Nets would only be gaining 0.08 expected championships over the next two seasons at the cost of 0.11 over the rest of the decade. The graph below shows the more pessimistic estimate.

This thought exercise of assigning numbers to championship probabilities I think at least puts us in the right mindset for evaluating the trade. One thing it made me think about more deeply is that winning a championship is just plain hard. Our generic 3.3% estimate of title odds for a future normal team that actually has its draft picks is not all that much better than the 1% odds I saddled the draft pick-less Nets with. In other words, the future title potential that Brooklyn loses from losing its draft picks is not as large as one might expect. Winning a championship is just plain hard to begin with.

What Is The Larger Takeaway Here?

The point of this analysis was not to offer precise, model-driven projections of the Nets after the Harden trade. The point was rather to introduce a framework for how to think about blockbuster trades aimed at turbocharging a team’s immediate title pursuit. And this type of framework is becoming increasingly relevant given the rise in the number of “all-in” type trades where contenders sacrifice most of their eligible future draft picks to land a mega-star.

I think it is informative to think about these type of trades from a cost-benefit perspective. The benefits for a team trying to win now come back to increased championship equity for this season and hopefully the next few seasons as well. The costs can be thought of in terms of the championship equity sacrificed farther out in the future from the lack of draft picks, valuable young players, etc.. The tricky part is quantifying the costs and benefits in a reasonable way.

Finally, it’s important to note that the “championship equity” framework is not the only relevant way to think about a trade. Plenty of trades and signings essentially don’t move the title needle at all; they perform other functions such as increasing the chances that a team will make the second round of the playoffs or make the playoffs at all. It can be quite reasonable to just make an ok or good team a bit better at a fraction of the cost of the Harden trade without really increasing expected titles. And another downside of throwing all of one’s draft picks into a trade is that a team is increasing their chances of being bad or really bad in the future, not just out of the title picture.

With all that being said, this Nets fan is personally hoping that the benefits of the Harden trade will outweigh the costs for Brooklyn. At the very least it will be an exciting next few years at Barclay’s Center!

Photo Credit: The photo at the beginning of this post was taken by Derral Chen on November 18, 2012 and is posted to flickr here. It has been made available under this license.

But I have the annoying “How real is this?” rattling around in the back of my head tempering my reaction to any small sample size. I know that weird things can happen over a short time period in the NBA. Outside shooting can ebb and flow as a prime example. The Basketball Gods fan be fickle.

Faced with this uncertainty, I like to try to quantify things. We know the Nets were projected to be a good, but not spectacular regular season team this year; sportsoddshistory.com had their win total over/under set at 45.5 games which translates into about 52 wins over a standard 82 game slate. Using this information (what I like to think of as a preseason prior) plus what we’ve seen on the court so far, what are our reasonable expectations going forward?

One simple way to combine the preseason prior with the on-court play is a linear regression. Let’s use preseason win total over/under and current season point differential as the independent variables. Here the dependent variable is the win pace for the team over the rest of the season. As an example of how to interpret this, if the model output is “45 win pace” that means that we expect our team to win the equivalent of 45 out of 82 games over their remaining schedule.

With this preamble out of the way, take a look at these projection lines:

Here I have plotted projections for our preseason 52 win team in four different time increments: after 10 games, 20 games, 30 games and 40 games. We see that the slope of the lines gets progressively steeper and this should make sense. A team’s point differential after 20 games has more predictive power than their point differential after 10 games, 30 games has more weight than 20, and so on.

Ok, now for the fun part! Let’s say we make it through 10 games of the season and the Nets are still dominating with a cool +10 point differential. We then consult the blue line in the graph and find that Brooklyn would be projected to play like a 55.5 out 82 win team going forward. That’s really impressive obviously, but not quite 60 wins level of impressive. We would want to have a bit of caution.

What about if they are still outscoring opponents by 10 points a game after 20 games? Then we can see from the blue line that we would update our view of the Nets higher still to a 57.5 win team going forward. If we make it through 40 games and the Nets are still +10 then this model sees them as a 61 win juggernaut.

I find it fun to look at a graph like this (which is probably a strange thing to find fun), but I should caveat this a bit. I projected what a hypothetical +10 point differential Nets team might look like, but this is on the high side of possible outcomes. Even though Brooklyn has looked like world beaters in the early going, we are much likelier to come back in a couple weeks and be talking about a +6.0 point differential team (which is still highly respectable).

Ok, two more important caveats. First, this type of analysis would be better if it took into account the strength of opponents played. In other words, something like basketball-reference’s Simple Rating System (SRS) would be better than point differential. The second caveat is that linear regression (and models in general) are better at making projections for more commonly seen input values. That is, the model knows how to handle a Nets team with a +5.0 point differential through 10 games better than if that team was at a gaudy +15.0, for example.

The reader may have noticed that I did projections for a team that has played at least 10 games already, not 2 games like the current Nets. I may be a coward, but I’m admittedly scared to use point differential to project after 2 games. Let’s sit back, enjoy the basketball at Barclays Center, and see where things stand in a few weeks.

To answer these questions, I looked at national polls from all presidential elections since 1980 (a total of 10 elections). These national polls were helpfully compiled by G. Eilliot Morris (who is currently employed by The Economist) in an R package called ‘politicaldata’.

I wanted to compare the accuracy of polls 1 month out to those taken right before the election. To do this, I found the average absolute error of all national polls taken about 1 month before the election (25-35 days before to be precise) and compared this error to the same calculation for polls taken within a week of the election (1-7 days). The table below shows these calculations:

We can quickly see that the polls were better some years than others. In 1992 and 1996 polls were not very accurate with 1 month to go; 2000, 2004, and 2008 conversely were within 1 point of the final result. We also see that polls taken within a week of the election were more accurate in some years, such as 2012 and 2016. For other cycles, like 1980, the picture a month before the election was actually less blurry.

The most important row in the table is the last row which shows the average absolute error over all 10 of the elections. We can see that the average absolute error 1 month before the election is 3 points, which is actually not too far off the 2.5 point average absolute error for polls one week before the election. So even though 1 month may feel like plenty of time for the race to change, polling history suggests that our picture of the race now is close to as good as it will get.

This is not to say that the polls on election day are still going to show an 8 point Biden lead! And that average absolute error is just that, an average. If we instead have a large polling miss like in 1992 or 1996, (and that miss underestimates Mr. Tump) the final popular vote margin could be Biden +2.4. Trump would then probably be a narrow favorite in the Electoral College due to his relative strength in the battleground states. That could happen of course, but is not likely. It is important to remember too that a polling miss could also benefit Mr. Biden.

What I am saying is that we should pay attention to the horse race polls now and not simply through our hands up and say it’s still too far away from the election to look at polls. We will have a better picture of the race on election day than we do now, but not a much better picture.

I was interested in where the shifts were from the 2012 to the 2016 election. Where did Democrats lose (and gain) ground? Of course, when you lose about 2 million votes in the popular vote margin, as the Democratic candidate did from 2012 to 2016, you are doing more losing of ground than gaining of ground.

The first thing I saw in the data was that smaller counties, already a strong area for Republicans, became even redder in 2016. Take all counties with less than 250,000 people. In 2012, Romney won a large 58% of the two-party vote from these counties. In 2016, Trump won an even larger 64% of the two-party vote from these counties. Just looking at raw vote margin, we saw a 3.6 million popular vote swing overall towards the Republican candidate. That is more than the overall popular vote swing of 2.1 million votes!

Remember that margins are important, even in places that are already pretty red or blue. At least in modern elections, if the Democrat can not lose too badly in rural America, they can make up the deficit in more urban areas and win the state overall (of course, this depends on the relative urban/rural makeup of a state). If they give up more ground than usual outside of large cities, they have an even tougher hill to climb.

Take the very important 2020 swing state of Pennsylvania as an example. Let’s divide the state into two parts, the very urban counties of Philadelphia and Allegheny (where Philadelphia and Pittsburgh reside, respectively), and the entire rest of the state.

In 2012, Obama won Philadelphia County and Allegheny County by a combined 583 thousand votes. He lost the rest of the state by a wide 273 thousand votes, but that means he still maintained a 310 thousand vote winning margin.

Now, let’s look at 2016. Clinton actually won Philadelphia County and Allegeny County by roughly the same 583 thousand votes. So far, so good. But she lost the rest of the state by an even larger 627 thousand votes, meaning she lost the state overall by 44 thousand votes! The areas outside Philadelphia and Pittsburgh were her undoing.

Ok, I should mention that this is a bit of an oversimplification. Dividing the state into just urban centers and everything else glosses over the suburbs, where Democrats made real progress in 2016. We can see on the image above shades of blue and purple in Montgomery, Chester, and Delaware counties, indicating a shift toward the Democrat.

Still, Pennsylvania echoes some of the trends we saw overall in the country in 2016. Clinton actually improved a bit on Obama’s 2012 overall performance in counties classified by the NCHS study as “Large Central Metros”, but was undone by losses in the smaller counties.

The fidelity of this type of analysis can be improved if you go to a smaller level than county (such as precincts), but I think the larger point still stands. Democrats lost real ground in rural America in 2016, and it remains to be seen if they will improve on their performance there in 2020.

For a moment, let’s focus our attention on raw vote margins. In the 2016 race, Democratic candidate Hillary Clinton received roughly 2.9 million more votes than Republican candidate Donald Trump. This was a smaller margin than the approximate 5 million vote lead Barack Obama achieved over Mitt Romney in 2012. So we have a 2.1 million shift in the popular vote towards the Republicans that we have to account for between 2012 and 2016. Where did these votes come from?

Looking at the map above, we see lots of light shades of red. Mr. Trump was able to improve on Mr. Romney’s 2012 vote margin in many states across the country, but made particular headway in the Midwest. The three darkest shades of red on the map are Ohio, Michigan, and Pennsylvania. The Republicans improved their vote margins by 613,000, 450,000, and 354,000 respectively in these three states. Crucially for the Electoral College, they flipped from Democratic victories in 2012 to Republican victories in 2016.

Other states also shifted to the right in 2016, but received less attention because the winning party did not change and thus their movement had no effect on the Electoral College. Going down the list to the next three largest shifts towards the Republicans, we have Missouri, Indiana, and New York. Each of these states swung roughly 250,000 votes towards the Republican candidate for a combined approximate 750,000 vote swing. We don’t really think about states like New York at the presidential election level, but safe blue or red states can still have a large effect on the popular vote.

Another solidly blue state in the 2016 election that had a huge effect on the popular vote was California. As we can see by its dark blue shade on the map above, Mrs. Clinton actually improved on Mr. Obama’s 2012 margin in California by roughly 1.3 million votes. In fact, she won the state over Mr. Tump by about 4.5 million votes, more than her 2.9 million popular vote victory overall! Unfortunately for Democrats, this additional vote margin in California did not have any effect on the Electoral College vote.

Two other large swings towards the Democrats in 2016 were in Texas and Utah, where Mrs. Clinton improved on Mr. Obama’s 2012 margin by 455,000 and 284,000 votes respectively. Neither of these shifts changed either state from red to blue though.

For reference, here are tables of the top 10 states that shifted to the right and the top 10 states that shifted to the left, in terms of raw vote margin, from 2012 to 2016. Data for these charts, and the map above, came from the MIT Elections Lab.

With the eighth seeded Trail Blazers upsetting the Lakers last night (less surprising) and the eighth seeded Magic upsetting the Bucks earlier in the day (more surprising!), the natural question is what can we expect going forward. Should we have higher expectations for our underdogs in future games of the series?

As it turns out, yep, a bit higher expectations. But we should be cautious.

With the data I had handy, all playoff game results from 2003-2018 for a total of 240 series, I built a simple linear regression. The inputs: the home team’s margin of victory in game 1 (INPUT 1) and the difference in strengths of the two teams in the series, as measured by the difference in their regular season point differentials (INPUT 2). The output: the average margin of victory for the team with home court advantage in the rest of the series.

(To be clear, we are measuring the remaining series margin of victory for the “favorite”, the team with home court advantage in a theoretical game 7. So we would be projecting the Lakers’ margin of victory in the remaining series games against the Blazers.)

In other words, we really want to know if the score in game 1 (INPUT 1) adds something meaningful to our prediction of the scores in future game of the series, after controlling for the strengths of the two teams (INPUT 2). Because it’s a quick morning post, here is the grizzly R output of the linear regression.

The result? Well, game 1 adds something to our prediction for future games. The coefficient for this term is about 0.09. This means that if the away team wins the first game in the series by 10 points, we would expect them to do about 1 point per game better in future games of the series than if they had lost game 1 by a point. Also, we see that the coefficient for the game 1 score is statistically significant at the 0.05 level (the p-value is the “Pr(>=|t|)” column). In other words, that 0.0869 value does seem to be different than 0, though it’s not significant at the 0.01 level for what it’s worth.

The (statistical) significance of the game 1 result actually kind of surprised me. I mean, we are often cautioned not to read too much into small sample sizes for good reason, and one game is indeed a small sample. Fluky things can happen. Our favorite can have an off shooting night, like the Lakers last night when they shot 15.6% from beyond the arc.

I still think we should not go overboard here. An underdog winning a close game 1 portends them being only fractions of a point better on average for the rest of the series than if they had lost game 1. Also, I think this result somewhat reflects the shortcomings of regular season margin of victory as a measure of team strength. Sometimes we are dealing with clearly different playoff squads than regular season ones and the game 1 score simply reflects this (think Cavaliers vs. Raptors in 2018).

And of course, it goes without saying that banking a win is good for the underdog’s odds of pulling the upset beyond any predictive power it might have on future scores in the series. They now only have to win 3 out of 6 rather than 4 out of 7.

The moral of the story for game 1’s? Adjust your priors for the series, but not too much.

If you want to replicate this analysis, the code is posted to Github.

Using pre-season betting data from sportsoddshistory.com and pre-playoffs betting data from vegasinsider.com, we can see roughly how the NBA title picture has changed:

We see the LA teams and Milwaukee reside solidly above the dashed line meaning that they are seen as stronger favorites today than they were in October. None of these teams by themselves are viewed as Warrior-like juggernauts of the past few seasons, but they are each in the neighborhood of 20-25% likely to win it all, according to the Vegas Insider odds. Collectively, they consume nearly 66% of the championship equity for this playoffs.

We can also see that the squads viewed as their biggest competition before the season are not judged to be as dangerous now. The Rockets and 76ers are each solidly below the dashed line, indicating a drop-off in title odds. The Sixers have suffered through somewhat of a disappointing year, posting only the 10th best net rating in the NBA. Now, they have to battle in the Eastern Conference playoffs without their second best player, Ben Simmons, who suffered a partial dislocation of his left patella in the seeding games. Houston attempted to shake up their squad with the addition of Russell Westbrook in the offseason and the jettison of Clint Capela in favor of Robert Covington and small-ball at the trade deadline. They have been an interesting team to watch, but haven’t quite been as dominant in the regular season as they may have hoped heading into the season.

Two other teams, the Raptors and Celtics, have had strong regular seasons and have driven their title fortunes up a bit. But Toronto and Boston are not seen as truly strong contenders; their title odds hover at 8% and 5% respectively.

Just how much of a 3-horse race are we witnessing? Well, going back to 2005, this year’s Lakers are seen as the strongest “third” contender. They have the best title odds of the team with the third highest chance of winning the title at roughly 19%.

How does this year’s title race compare to the previous years in terms of stability? To measure this, I found the difference between each team’s pre-season title odds and pre-playoffs odds for each year from 2005 to the present. Then, I added the absolute values of those differences together for each year. A larger sum implies a more different title picture from the start of the season to the start of the playoffs.

From this graph, we can see that 2020 has been more volatile than two recent seasons, 2019 and 2017, but not as volatile as some of the title races in the mid to late 2000’s. Initially, I thought this season would rate as one of the least volatile due to the stability of the three favorites, but the drop-off in fortunes of the next tier down (the Rockets, 76ers, and the (gulp) Warriors) has driven this metric to about average.

This type of analysis needs the caveat that I am only looking at title odds from two sources (sportsoddshistory.com and vegasinsider.com), so take it with a grain of salt. In particular, the order of the three favorites I think can certainly be debated. But I think the basketball community would agree that the LA teams and Milwaukee have separated themselves at least a bit from the pack. Heading into the weird Bubble Playoffs of 2020,  if someone other than the Bucks, Clippers, or Lakers takes the crown, that would be at least a mild surprise.