How to analyze basketball stats: a match analyst's method, tested on three Swish & Dunk games

A box score tells you everything and nothing. 112 points can be a great offense or just a fast game; 31 rebounds can be dominance or simply the other team missing a lot. This guide shows how a match analyst reads the numbers of a 48-minute basketball game: what to calculate first, which stats decide games and which ones lie. Then we put the method to work on three Swish & Dunk games played on October 4, 2026, to show how player skills and the choice of offensive and defensive tactics shaped what the match engine produced.

Step one: possessions and pace

A basketball game lasts 48 minutes, four quarters of 12, plus five-minute overtimes. In that time each team has the ball a certain number of times, and both teams have almost exactly the same number, because every possession ends by handing the ball to the other side. That number, the pace, is the starting point of any analysis.

The box score does not show possessions, but you can estimate them with a formula used by every analyst:

Possessions ≈ FGA − OREB + TOV + 0.44 × FTA

Shot attempts end a possession, unless an offensive rebound extends it; turnovers end it; and roughly 44 free throws in a hundred end a possession too (the rest are first free throws, and-ones or technical fouls). Calculate it for both teams and take the average.

Why does it matter? Because a team scoring 112 points in 100 possessions and one scoring 90 in 80 have exactly the same offense. Without pace you are comparing games of different length. Modern NBA teams play about 100 possessions a game; a slow team plays the low 90s.

Efficiency: offensive, defensive and net rating

Once you have possessions, divide: offensive rating (ORtg) is points scored per 100 possessions, defensive rating (DRtg) is points allowed per 100 possessions, and net rating is the difference. Today's NBA averages around 115 points per 100 possessions. These are the most honest numbers in basketball: whoever has the higher rating wins the game, regardless of pace.

A tip: compare a team's rating in one game with its season average. A team that usually scores 130 per 100 and stops at 103 did not just "have a bad night": something in the matchup took away what it normally does.

The Four Factors

Ratings tell you who was more efficient. To understand why, analysts use the Four Factors described by Dean Oliver: the four ways a possession can go well or badly. Every gap in rating can be traced back to one or more of them.

The factors work on both ends: your defense is measured by the same four numbers for the opponent. Two rules of thumb. First, raw rebounds deceive: a team that misses a lot has many more rebounds available to its opponent, so always use the percentage. Second, the factors trade against each other: an aggressive defense forces turnovers but often concedes better shots, more free throws and more offensive rebounds. The question is never "did we force turnovers?" but "did we win the sum?".

Shot profile: where the points come from

Field goal percentage mixes very different shots. Analysts split them into three zones: at the rim, mid-range and three. In the NBA a shot at the rim is worth about 1.3 points, a three at 36% about 1.08, a mid-range jumper at 40% only 0.8. This is why the modern game lives at the rim and behind the arc.

Three useful numbers: 3PA rate (threes / field goal attempts) describes a team's style; true shooting (TS% = points / (2 × (FGA + 0.44 × FTA))) measures a player's scoring efficiency including free throws; and assists per made basket tells you whether points come from ball movement or from individual creation.

When you look at three-point percentage, remember the sample. Over 25 attempts, a 36% shooter will land anywhere between 25% and 50% on most nights. Shot quality (how open, from where) depends on tactics and skills; making the shot carries a lot of randomness.

Reading individual stats

• Minutes first. 15 points in 20 minutes and 15 in 40 are two different games. To compare, scale to 36 or 48 minutes, but beware small samples.
• Efficiency, not volume. 20 points on 8-of-20 shooting can hurt a team; 14 points on 6-of-8 helps it. Look at TS% and turnovers next to points.
• Who has the ball when things go wrong. Turnovers concentrate on the players who handle most. A bench guard with poor ball security against a pressing defense is a classic weak link.
• Plus-minus with care. It depends on teammates, opponents and a few possessions. It is excellent for finding which lineup decided a game, poor for ranking players.
• Game Score (John Hollinger) sums everything into one number: points + 0.4 × FGM − 0.7 × FGA − 0.4 × missed FT + 0.7 × OREB + 0.3 × DREB + steals + 0.7 × assists + 0.7 × blocks − 0.4 × fouls − turnovers. It is what we use to pick the Players of the Month.

Game flow: quarters, runs and the finish

The final score is a sum of stretches. Quarter scores show when the game turned; lead changes and the largest lead tell you whether it was a fight or a procession. Check the last five minutes of a close game separately: who was on the floor, who took the shots, who turned it over. And remember garbage time: in a blowout, the last minutes distort the stats of players who barely play.

The method in seven steps

1. Pace: estimate possessions and compare them to both teams' usual pace.
2. Ratings: points per 100 possessions on both ends, against season averages.
3. Four Factors: find the one or two that opened the gap.
4. Shot profile: rim, mid-range, three, transition. Does the shot diet match the tactics?
5. Matchups: which player skills explain the numbers on each side.
6. Flow and rotations: quarters, runs, the finish, who played which minutes.
7. Luck: what is repeatable (shot quality, turnovers, rebounding) and what is not (a 59% night from three, 9-of-20 free throws).

Reading a game in Swish & Dunk

Every Swish & Dunk game is simulated possession by possession (the whole process is described in how the match engine works). After the final buzzer the match report shows the box score, the full play-by-play and the offensive and defensive tactics used by both teams. That is all you need for the method above.

The key to connecting stats and tactics: in the engine a tactic is a set of multipliers on player decisions and duels, not a script. Each one leaves a recognizable fingerprint in the box score:

And the most important point: the multiplier works through your players, not instead of them. A defense that strengthens rim protection makes your center a better shot blocker; if you do not have a center, there is nothing to multiply. Matchups are by position: your point guard defends theirs, your center defends theirs. Who each tactic beats is covered in the complete tactics guide; what each skill does is in the guide to all 49 skills.

We chose three games played on October 4, 2026 in which both teams had set an offensive and a defensive tactic. Ratings and Four Factors are calculated with the formulas above, using the average possessions of the two teams; data on play types and shot zones come from the play-by-play; the "rotation Overall" is each team's average Overall weighted by minutes played.

Game 1: Khartoum Hammerhead 94, Hong Kong Monarchs 112

Division 9, Group A. Khartoum: Princeton Offense + Drop Coverage. Hong Kong: Pace and Space + Drop Coverage. Rotation Overall 70.8 vs 70.5: two teams of the same level, which makes it a good tactical lab.

What to expect. On the tactics matrix, Pace and Space against Drop Coverage is a clear edge for the offense: Drop drops the big near the rim and concedes the perimeter, exactly where Pace and Space wants to shoot. Princeton against Drop is even. So before tip-off the edge belonged to Hong Kong on one end and to nobody on the other.

Khartoum's offense did its job. 118.5 points per 100 possessions, ten above its season average, with only 6 turnovers (7.6%) and 25 assists on 36 made baskets. Those are the Princeton fingerprints: the lowest turnover risk in the game and the most ball movement. Drop Coverage also forces few turnovers, so the two effects added up. 13-of-26 from three: Drop leaves shooters open on this side too.

Hong Kong's offense went off the charts: 141.2 per 100, 31 points above its average, a 74.6% eFG. 13-of-22 from three, 8-of-13 on catch-and-shoot plays. Even Sebastien Lacroix, a power forward with a three-point rating of 43, went 2-of-4 from deep: Drop Coverage lowers the contest on every perimeter shot, including the ones you would happily give up. An honest analyst has to add that 59% over 22 attempts goes beyond what the tactic explains: Drop gave them good shots, a hot night did the rest.

The detail that turned a loss into a rout: Drop Coverage without a center. Khartoum used no center at all. Its tallest rotation player was power forward Andre Nguyen (206 cm), and three point guards (Francisco Gonzalez, Xenophon Georgiou and Delfin Carmona) played 96 of the 240 minutes. Drop Coverage multiplies rim protection, but the shot is contested by the defender in the shooter's position: multiplying a guard's interior defense does not build a wall. Hong Kong went 15-of-21 at the rim, with 218 cm center Efstathios Dimitriou at 5-of-8. Khartoum paid the price of Drop (open threes) without collecting the reward (a protected rim).

Size cost them on the glass and at the line too. Khartoum recovered only 11.4% of its misses, Hong Kong 27.3%, even though Drop strengthens the defensive glass. Hong Kong got to the line 27 times on 61 shots (0.44): Lacroix went 6-of-6, Ilias Chrysostomou 5-of-9, both strong at drawing fouls and attacking the basket. Khartoum, meanwhile, went 9-of-20 from the line, which is not a tactical problem but costs about five points against an average night. Andre Nguyen and Delfin Gimenez both played all 48 minutes, and in Swish & Dunk ratings fall as stamina drops: Hong Kong scored 35 in the fourth quarter.

The verdict. Khartoum's offense was fine; its defensive choice was wrong twice. Against Pace and Space you need a defense that contests the perimeter (Switch-Everything or Match-Up Zone), and Drop Coverage only makes sense with a real rim protector. Without one, the tactic multiplies a strength the team does not have.

Game 2: Sacramento Crusaders 85, Bilbao Destroyers 92

Division 2, Group A. Sacramento: Motion Offense + Full-Court Press. Bilbao: Triangle Offense + Man-to-Man. Rotation Overall 86.7 vs 85.1, and Sacramento at home with a 13-4 record: on paper, the favorite. Bilbao never trailed and led by as many as 15.

The headline stat says the press worked. Bilbao turned it over 17 times (20.9% of its possessions, against 12.8% on the season), and Norbert Zahn, steal rating 99, came up with 7 steals. The press also found the weak link: next to Lothar Voss (6 turnovers), the other big offender was Howard Jefferson, a 68 Overall shooting guard in a team of 87-91 rated players, with ball security 68 and decision making 55, on the floor for 31 minutes: 5 turnovers. Pressure always lands on the weakest ball handler on the floor.

The Four Factors say otherwise. First, the turnover battle ended even: Sacramento gave it away 16 times too, so the net gain of the press was about one possession. Second, every possession that beat the pressure turned into a great shot. Full-Court Press weakens the contest both on the perimeter and at the rim, and the Triangle is the system that attacks the paint hardest: Bilbao made 25 of 37 twos (67.6%), 14-of-21 at the rim and an incredible 11-of-16 from mid-range. The press also costs fouls (17, i.e. 21 Bilbao free throws) and defensive rebounds (Bilbao recovered 25% of its misses). A 12-point gap in eFG decided the game; the turnovers did not.

Sacramento's offense ran into Bilbao's defenders, not into a system. Motion against Man-to-Man is even on the matrix: Man-to-Man strengthens nothing and opens nothing, so everything comes down to individual duels. And those duels were brutal: Roberto Ferrer (perimeter defense 97, interior 100, shot contest 100), Johnathan Albee (interior defense, contest and blocks all at 100) and Roberto Felder (interior and contest 100). Sacramento, which averages 132 points per 100 possessions, stopped at 104: 8-of-22 in isolation, 4-of-13 on catch-and-shoot plays. Momcilo Simic (21 points on 8-of-11, all 48 minutes) was the one who found a way through.

Two more details. Sacramento's 20-of-20 at the line was not enough to make up for the gap from the field. And the 16 minutes given to Jaxon Washington, a 66 Overall point guard, ended at −9, the worst plus-minus on the team: in a seven-point loss, those are minutes that weigh.

The verdict. The Triangle is one of the offenses that beats the press: it protects the ball, attacks the rim and crashes the offensive glass, exactly where the press is weakest. Against a Triangle the best choices are Pack Line or Drop Coverage. This game is the perfect example of a misleading stat: 17 turnovers forced and 12 steals look like a defensive masterclass, and instead they were the cost of 37 easy twos.

Game 3: Philadelphia Isis 74, Denver Zombies 75

Division 1, Group A, the top league in the game. Philadelphia: Princeton Offense + Switch-Everything. Denver: Pace and Space + No-Middle. The two starting fives are equal (average Overall 89.4 vs 89.6), but Denver's rotation is deeper (88.0 vs 83.8). Denver came in at 13-4 with the best offense in its group. Eight lead changes, eleven ties, Philadelphia up 57-51 after three quarters.

A slow game, and that is a tactical choice. 72.8 possessions, about nine fewer than both teams' season average. Princeton is the slowest system in the game: Philadelphia's possessions were long and patient. That is why 74 and 75 points do not mean bad offenses, and why you have to look at the ratings: fewer possessions also mean more variance, which helps the underdog stay close.

Switch-Everything against Pace and Space: the matrix in action. Denver averaged 131.7 points per 100 possessions and 41.2% from three; it finished at 103.0 and 7-of-26 from deep, 2-of-13 on catch-and-shoot threes. Not because of bad shooters: Nolan Forstie and Pavle Nesic both have a three-point rating of 100. Switch-Everything raises the contest on every perimeter shot, and Philadelphia has the defenders to make the most of it: Parks Randle and Ipolitas Simkus at 100 perimeter defense, Miguel Carrasco and Pearly Laboone at 95, Nicolas Flores at 90. A multiplier on elite defenders is the toughest perimeter defense you can build.

Switch-Everything has a weakness, the rim, and the numbers show it: Denver went 14-of-20 there (70%). But Pace and Space sends few shots inside, only 20 attempts at the rim against 26 threes. That is why the matrix gives the defense the edge: the hole is exactly where that offense does not go.

Turnovers are the other mark of the two systems. Denver gave it away 14 times (19.2%), 7 of them out of the pick-and-roll; point guard Erasmo Sáez, despite ball security 100, finished with 5 turnovers and 2-of-12 shooting against a defense with Simkus and Carrasco, both rated 100 in steals. Philadelphia, with Princeton's low risk, gave it away only 9 times.

Philadelphia's offense, on the other hand, stalled. Princeton pushes toward the paint, No-Middle protects it: the two effects cancel out and Philadelphia went 12-of-21 at the rim. The best weapon was the post, 7-of-8 for 14 points with Carrasco (post offense 100) and Laboone. The pick-and-roll, which Princeton uses rarely, was a disaster when it came up (2-of-11), and the threes did not fall (8-of-26). Denver stayed alive with what Pace and Space produces even on a bad shooting night: transition (10 shots, 13 points against Philadelphia's 2) and offensive rebounds (28.6%, four from Charilaos Oikonomou, offensive rebounding 100).

Where the game was decided: the rotation. Read the plus-minus by unit. Philadelphia's starters won their 29 minutes by 4 points; the second unit broke even; the third, with players between 64 and 76 Overall, lost its 5 minutes by 5. And those minutes landed in the finish: with 5:17 left Philadelphia led 68-67 with that unit on the floor. Forstie (93 Overall, 48 minutes) scored an and-one, Fabrice Joly hit a jumper, and Denver went up 72-68. Stephane Jacquet answered from three, then with the starters back Laboone fouled Oikonomou on a three: 3-of-3 at the line, 75-71. Simkus hit a three with 30 seconds left, Nesic missed at the other end, and the clock ran out with the ball in Carrasco's hands.

The verdict. Philadelphia won the tactical battle: it took 29 points per 100 possessions off one of the best offenses in the league and lost by one point to a deeper team. The lesson is in the numbers: the minutes you give to the end of your bench have to be played sometime, and in a one-possession game they can cost more than any tactical choice.

What the three games teach

• Season averages are the best context. Denver −29 per 100 possessions, Sacramento −28, Hong Kong +31: without the comparison those games look like "low-scoring" or "high-scoring", with it they tell a tactical story.
• The flashiest stat is not always the one that decided the game. Sacramento forced 17 turnovers and lost on shooting; Hong Kong had more turnovers than Khartoum and won by 18.
• A tactic multiplies what your players have. Drop Coverage without a center, Switch-Everything with five elite perimeter defenders: same idea, opposite results.
• The matrix sets the odds, not the result. Two of three teams with the tactical edge won; Philadelphia, which had it, lost on rotation and one possession.
• Always separate quality from luck. 59% from three and 9-of-20 from the line are not repeatable; turnovers, rebounding and shot selection are.

Frequently asked questions

What is the most important stat in a basketball game?

Points per possession, usually written as offensive rating (points per 100 possessions). Totals like points, rebounds or assists depend on how many possessions a game has; efficiency per possession does not. The team with the higher rating wins, and the Four Factors explain why.

How do you calculate possessions from a box score?

Possessions ≈ field goal attempts − offensive rebounds + turnovers + 0.44 × free throw attempts. Calculate it for both teams and take the average, because the two sides of a game have almost the same number of possessions.

What are the Four Factors in basketball?

Dean Oliver's four keys to winning a game: shooting (effective field goal percentage), turnovers (turnovers per possession), offensive rebounding (share of your own misses you recover) and free throws (free throw attempts per field goal attempt). Shooting matters most, roughly 40% of the weight, followed by turnovers, rebounding and free throws.

Is plus-minus a reliable stat?

Only in context. A player's plus-minus depends on who shared the floor with him, on the opponents he faced and on a handful of possessions. Use it to spot which lineups or stretches decided a game, not to rank individual players.

How can I tell if my tactic worked in Swish & Dunk?

Look for its fingerprint in the stats. A tactic that protects the rim should lower the opponent's efficiency at the basket, a pressing defense should raise their turnover rate, a three-point offense should raise your share of threes. If the fingerprint is there but you still lost, the problem was the players or the matchup, not the system.

Want to see how these numbers have changed in real basketball? Our analysis of 30 NBA seasons follows pace, threes and efficiency from 1996 to today.

Last updated: October 4, 2026