clock menu more-arrow no yes mobile

Filed under:

Why Are the Computers So High on the Celtics’ Title Odds?

Boston has turned around its season, sure, but most projection systems now view it as the favorite in the East. Do the numbers know something that we don’t?

AP Images/Ringer illustration

The Boston Celtics are a force to be feared. They have the best point differential and defensive rating in the Eastern Conference, and they boast the conference’s best record since turning their season around on January 8. Within the past month, they’ve bludgeoned the 76ers by 48 points, the Heat by 30, and the Nets by 35 and 23.

The Celtics are also an enigma. For all their strengths, they also rank 18th in offensive rating, according to Cleaning the Glass, and they recently dropped games to the Pistons and Pacers. With less than a quarter of the regular season still to play, they’re in sixth place in the East, closer to the play-in bracket than the top of the standings.

And yet, despite that ambivalent résumé, projection systems believe in Boston’s potential. The Ringer’s NBA Odds Machine gives the Celtics an 11 percent chance to win the title, and we’re not alone: Basketball-Reference pegs the Celtics’ championship odds at 9 percent, ESPN at 13 percent, and FiveThirtyEight at an astounding, league-leading 17 percent.

Boston’s recent momentum, and the projections’ ensuing hopes, rests primarily on the team’s defense. Defense doesn’t actually win any more than offense in the playoffs, but Boston’s defense goes above and beyond the norm. During the Celtics’ run since January 8, their defense is a whopping 4.2 points per 100 possessions better than the next-best defense; the second-place Heat are as close to 14th place as they are to first, per CtG.

The league has come to value switching as a means of slowing screen-heavy offensive sets, and Boston offers the premiere example of this defensive style. The Celtics switch on 40 percent of opposing picks, the highest mark in the league, according to analysis of Second Spectrum data. (Miami’s at 39 percent; no other team is above 33.) And they allow just 0.86 points per possession when they switch a screen—the league’s stingiest mark.

Boston benefits from fitting elite individual defenders within that broader team concept. Seven Celtics have played heavy minutes since the trade deadline, and all are at least average on defense, with several rating among the league’s most impactful defenders this season, according to estimated plus-minus.

Celtics Defenders by Impact Per 100 Possessions

Player Defensive EPM Percentile
Player Defensive EPM Percentile
Derrick White +3.6 99th
Marcus Smart +2.8 98th
Al Horford +2.3 97th
Robert Williams III +1.5 90th
Jayson Tatum +1.2 87th
Jaylen Brown +0.8 80th
Grant Williams 0.0 62nd

They’ve even cleaned up their penchant for fouling: Before January 8, the Celtics were still a top-five defensive team but ranked only 18th in opposing free throw rate, per CtG. But the Celtics have improved to fourth in opposing free throw rate since that juncture.

The return of Marcus Smart from injury and the addition of Derrick White, via trade from the Spurs, help. But the most obvious shift that can help explain the Celtics’ current form comes from rookie coach Ime Udoka’s deployment of Robert Williams III, who is defending about half the number of screens that he was earlier in the season, according to analysis of Second Spectrum data.

Instead, Time Lord often plays a free safety role, nominally guarding an unthreatening perimeter player while Al Horford or another teammate takes the primary screen setter. Thus Williams—who ranks third in the league with 2.2 blocks per game—can roam to deter any attempt at the rim. Note that before he unleashes his 7-foot-6 wingspan in all the plays in this video, Williams guards a non-shooting wing like Torrey Craig or CJ Elleby rather than the opposing team’s best big.

Or check out the best example of all, from late in Boston’s win against Denver in mid-February. Instead of matching up against Nikola Jokic, Williams takes Aaron Gordon while Jaylen Brown guards the reigning MVP. After a switch forces the undersized White onto Jokic, Jayson Tatum doubles, which opens a passing lane to the hoop—only for Williams to burst over, erase the layup attempt, and help seal the Celtics’ victory.

(As an aside: Recent playoffs haven’t offered much in the way of stout defensive battles; across the past five postseasons, only six teams have won a game with fewer than 90 points scored. Yet between Boston and Cleveland, Miami and Toronto, this year’s Eastern Conference playoffs could bring defense back at the most important moments of the schedule, with the teams displaying stylistic diversity in how exactly they prevent teams from scoring, from Boston and Miami’s switches to Cleveland’s three towers to Toronto’s fleet of 6-foot-8 wings.)

But for all of their defensive prowess, the Celtics’ profile still has some blemishes. Boston’s record falls 5.6 wins short of where we’d expect based on its point differential, according to CtG, which is the largest gap for any playoff team, and the culprit is trouble in close games. Boston is just 11-18 in “clutch” games (defined by as games within five points in the last five minutes).

Typically, good teams in all games are also good in the clutch, and vice versa. The 11 teams with the worst records in clutch situations this season include Boston and 10 squads with overall losing records.

Worst Teams in the Clutch

Team Clutch Record Clutch Win % Non-clutch Record Non-clutch Win %
Team Clutch Record Clutch Win % Non-clutch Record Non-clutch Win %
Celtics 11-18 38% 25-9 74%
Hornets 13-17 43% 17-15 53%
Hawks 10-14 42% 19-17 53%
Spurs 9-19 32% 15-18 45%
Pacers 9-26 26% 12-15 44%
Knicks 13-19 41% 12-17 41%
Pelicans 10-14 42% 15-22 41%
Kings 10-19 34% 12-21 36%
Thunder 14-19 42% 5-22 19%
Magic 8-15 35% 6-32 16%
Pistons 13-17 43% 2-29 6%

But this discrepancy shouldn’t be a concern going forward; with rare exceptions, past clutch performance is not predictive of future clutch performance. The correlation between regular-season clutch winning percentage and postseason clutch winning percentage is negative-0.05, among teams with at least five clutch playoff games since the 2010-11 season—or basically zero, signifying no meaningful correlation at all. (Looking at all playoff teams in that span, even those with just one clutch playoff game, the correlation is a still-tiny 0.14.)

The 2020-21 Bucks were 13-15 in clutch games in the regular season, then 7-2 in the playoffs en route to a title. The 2019-20 Heat were 18-18 in clutch games in the regular season, then 11-3 in the playoffs. The 2011-12 76ers had the worst clutch regular-season performance among all playoff teams in this span, with a 7-20 mark—then went 6-2 in the playoffs.

Even if their clutch struggles even out over the coming weeks and months, though, the Celtics’ offense remains a real concern. They still struggle to score at either of the two most lucrative areas on the court: They rank just 26th in proportion of shots at the rim, per CtG, and while they take a healthy amount of 3-pointers, they don’t actually make all that many. For context for this chart, league average 3-point accuracy this season is 34.9 percent. That means only one of the Celtics’ seven most-used players is even at the average. (Josh Richardson, sent to San Antonio in the White trade, was at 39.7 percent on 3.5 long-range attempts per game.)

Celtics’ Rotation From Distance

Player 3-Point Attempts Per Game 3-Point Accuracy
Player 3-Point Attempts Per Game 3-Point Accuracy
Grant Williams 3.4 43.7%
Jaylen Brown 7.4 34.8%
Jayson Tatum 8.4 32.9%
Marcus Smart 4.8 32.0%
Derrick White 5.3 30.6%
Al Horford 3.9 30.3%
Robert Williams III 0.0 0.0%
White’s figures here include his performance in both San Antonio and Boston. He is shooting 25.0 percent on 5.1 attempts per game since being traded.

Boston’s shooting problems are compounded by the lack of a central playmaker. The Celtics don’t have an offensive engine like most other contenders, which is why Tatum led the team but ranked only 52nd in the league in points created per 36 minutes when we ran the numbers in January. His assist rate hasn’t budged since last season. Similarly, Tatum leads the team in time of possession, but ranks in a tie for just 45th in the league, according to NBA Advanced Stats tracking (minimum 10 games played).

Rather, the Celtics have a number of good passers—including and especially their bigs—but no great creators in a half-court setting. They have to work hard to generate open shots and, when ping-ping-ping passing sequences don’t align just right, often end up settling for lower-quality isolation looks. They rank fourth in frequency of isolation plays but don’t score nearly as efficiently as teams like the Nets (who rank first in isolation frequency) and Bucks (third). Tatum, the team’s lone All-Star, ranks 12th out of 18 players with at least 150 isos this season—not terrible, but not the kind of go-to scoring this offense needs.

More zest in transition would help, and Boston has perked up in a small sample since trading for White. Pre-deadline, the Celtics added a piddling 1.4 transition points per 100 possessions, per CtG, which ranked 27th; since the deadline, they’re up to a seventh-place 4.3 transition points added.

If there’s one additional, under-the-radar warning sign for the team, it’s that the defense might not be quite as indomitable as it appears, thanks to a large dose of shooting luck. Boston’s opponents have underperformed their shooting expectation—based on factors like shot location and defender distance, from Second Spectrum—by the largest margin for any team. The gap is especially wide during Boston’s surge up the standings: Since January 8, Celtics opponents have an effective field goal percentage 3.3 percentage points worse than expected, while the next team on the list is just 0.8 percentage points below.

Thanks to its scheme and individual talent, Boston’s defense should still be one of the league’s best even if that run of fortune doesn’t quite continue—but given the Celtics’ tendency toward offensive stagnation, it might need to be the best, period, for the team to fulfill its promise. The 2003-04 Pistons are a historical anomaly for a reason; not many teams win the title solely because of defensive dominance.

The projection systems are fond of the Celtics because of both their high point differential and the talent on their roster; FiveThirtyEight, in particular, is enamored of the Celtics’ rotation. Of course, with such tough competition in the East, even the computers think there’s a much greater chance they lose in the first round (37 percent chance, per our Odds Machine) than win the title (11 percent chance). Yet because the field looks so wide open, Boston really might have one of the best chances among the group.

After all, the Celtics are capable of beating anyone: According to CtG, the Celtics have the league’s best net rating (plus-4.9 points per 100 possessions) against opponents that rank in the top 10 in point differential. For a team that was 18-21 not long ago, with growing concern over the fit between its two stars, even a 1-in-10 chance at the Finals marks a tremendous leap in the right direction.

Leaguewide stats through Sunday’s games.