The email keeps the headlines. This is the working: a graph and the full table for every section. Tap or hover any graph for the numbers.
Pace is how quickly you shoot plus how quickly they shoot. Our feed splits it into seconds per possession on offence and on defence. If a team's offence got faster and its defence didn't change, it chose to run. If only the defence number moved, opponents are running on it.
| Team | Pace rank | Pace | Off. secs | Def. secs | The story |
|---|---|---|---|---|---|
| Heat | 27th → 1st | 96.1 → 105.5 | 15.1 → 12.6 | 14.8 → 14.7 | Pure choice. The offence is shooting 2.5 seconds sooner, and the defence didn't change at all. Fastest offence in the league. |
| Magic | 30th → 13th | 95.5 → 100.4 | 15.6 → 14.2 | 14.6 → 14.5 | Same story, smaller: the slowest team in the league stopped walking it up. |
| Blazers | 16th → 3rd | 98.6 → 102.6 | 14.5 → 13.8 | 14.7 → 14.2 | Both: they push, and opponents run back at them. |
| Kings | 19th → 8th | 97.8 → 101.5 | 14.6 → 14.3 | 14.9 → 14.1 | Mostly the defence: opponents are getting quick shots on them. Not a good sign. |
| Nuggets | 9th → 23rd | 99.6 → 99.4 | 14.2 → 14.8 | 14.7 → 14.1 | Slower on offence (more Jokić half-court), while opponents shoot quicker on them. |
| Grizzlies | 2nd → 12th | 102.6 → 100.6 | 13.5 → 14.1 | 14.6 → 14.5 | The run-and-gun Grizzlies slowed right down on offence |
| Spurs | 11th → 24th | 99.2 → 99.4 | 14.4 → 14.2 | 14.7 → 14.8 | Their raw pace barely moved. The league sped up around them. |
J SWISH'S TAKEMiami going from 27th to 1st is the biggest style change in the league, and it's 100% on purpose. Nobody's talking about it.
Usage is the share of his team's plays a player finishes (a shot, free throws or a turnover) while he's on the floor.
| Player | Usage | True shooting | Minutes | The story |
|---|---|---|---|---|
| Dillon Brooks (PHX) | 18.0 → 28.4 | 55.4 → 56.7 | 31.8 → 31.3 | The biggest jump in the league, and he got more efficient doing it. |
| Michael Porter Jr. (BKN) | 20.2 → 28.6 | 61.7 → 61.2 | 33.7 → 32.6 | New team, number one option, same efficiency. |
| Nickeil Alexander-Walker (ATL) | 16.0 → 24.2 | 58.0 → 59.9 | 25.3 → 32.7 | Our Most Improved pick. More shots and better shots. |
| Jaylen Brown (BOS) | 28.7 → 35.9 | 55.6 → 57.9 | 34.3 → 33.1 | Carrying Boston. |
| Jalen Duren (DET) | 16.4 → 23.3 | 70.4 → 69.7 | 26.1 → 29.1 | 7 points more usage, and his efficiency didn't move. That's rare. |
| Kyshawn George (WAS) | 16.1 → 22.1 | 50.1 → 60.5 | 26.5 → 31.2 | Efficiency up 10 points: our NOT YET flag. See the hot list. |
| Ryan Rollins (MIL) | 18.4 → 24.4 | 60.0 → 58.6 | 14.6 → 32.5 | Minutes more than doubled. |
| Shaedon Sharpe (POR) | 25.5 → 31.4 | 55.3 → 53.9 | 31.3 → 27.7 | More shots in fewer minutes, and less efficient. The one on this list we'd worry about. |
| Jrue Holiday (POR) | 15.9 → 21.2 | 56.4 → 56.2 | 30.6 → 33.4 | New team, bigger role. |
| Deni Avdija (POR) | 23.2 → 28.2 | 60.7 → 62.0 | 30.0 → 34.1 | More of everything, and better. |
THE RULE WE USERole (minutes and usage) holds. Efficiency jumps give most of it back. So Duren, Alexander-Walker and Avdija are REAL. George's efficiency is NOT YET.
From our back-test: hot shooters keep about a quarter of the jump after 15 games, or about 40% if their free-throw percentage rose 3+ points too (the "FT tell"). Cold shooters keep only about an eighth of the drop, so most of them bounce back.
End-of-quarter heaves stopped counting from 2025–26, so this season's 3P% reads a touch high against last season's. It doesn't change any call here.
| Hot | 3PA | 3P% | FT tell | Projection |
|---|
| Cold | 3PA | 3P% | FT change | Projection |
|---|
J SWISH'S TAKEBuy KAT and Dort, sell Portis (his free throws went the other way). Dosunmu is the one hot hand we'd half-believe.
Lineups with 150+ possessions (29 qualify). Each bar runs from the points a five allows per 100 to the points it scores. Green means it outscores opponents; red means it gets outscored. The bar's length is the net rating. The dashed line is the league average, 115.9: a scoring dot right of it is an above-average offence, an allowed dot left of it is an above-average defence.
| Team | Five (possessions) | Off. | Def. | Net |
|---|
The Suns' five is a defence story: 96.8 points allowed per 100 is the best defence of any lineup on the list. The Kings' five is everything wrong at once: four scorers who all need the ball, plus Sabonis, and −18.4 to show for it.
PLAY THIS MORE · NOT YETBucks: swap Trent Jr. for Kuzma and it's +42.4 in 118 possessions. Small sample.
Who's guarding the other team's best player most, and how they're doing. This needs NBA.com's matchup data, which isn't reachable from our side yet. When it is, this section shows the top 10 by time guarding All-Stars, points allowed per 100 matchup possessions, and the biggest workload jumps from last season.
"Shot quality" is the effective field-goal % an average shooter would get on the exact shots a team took. Shot quality sticks. Shot-making mostly doesn't (our back-test: r ≈ 0.8 against 0.6). Teams above the line are hitting more than their shots deserve; teams below it are missing good looks.
| Players, 200+ shots | Shot quality | Actual eFG% | Gap |
|---|
J SWISH'S TAKEBuy the Blazers' offence, sell the Lakers'. The Bucks are hot on a poor team, so it's hiding a bigger problem. Jokić's +18.7 isn't luck: it's the model failing to understand him, and he breaks it every season. KAT is getting good looks and missing them: buy.
Net rating with every minute, against net rating with blowout 4th quarters taken out (any game 20+ points apart after three quarters). 20 of 30 teams shift by less than half a point, so early-season blowouts aren't inflating many numbers yet.
Better than they look: Thunder +15.1 → +16.8, Wizards −14.9 → −13.8, Cavs +3.3 → +4.4. Their benches gave points back. Worse than they look: Nets −11.7 → −13.8, Spurs +5.2 → +4.2, Magic +5.2 → +4.4. They padded in blowout minutes.
Pairs with 400+ possessions together, built from five-man lineup data. Each square is the team's net rating per 100 with that combination on the floor.
Jamal Murray without Aaron Gordon is a minus. With him, +22. Gordon makes Murray work. The Heat and Lakers pairs are ordinary apart and elite together: that's real fit. Grimes is +13 without Edwards and −22 with him. Sabonis is in three of the worst seven duos in the league.
One team per issue. Each arrow runs from passer to scorer, thicker for more assists; each circle is sized by assists made. Hawks, 21 games.
Johnson is the hub. He fed Okongwu 30 times, Risacher 29 and Alexander-Walker 24, and he's also the top finisher (Daniels → Johnson, 32). Dyson Daniels is the second engine: 32 to Johnson, 20 to Okongwu, 17 to Alexander-Walker. Trae Young barely shows up: 10 assists to his top target in this stretch .
Minutes a game in the first four weeks against the last two weeks.
Three Nuggets in the top eight risers (Nnaji, Cameron Johnson, Watson): someone's minutes are being shared out in Denver .
| Who closes | 4th-quarter minutes, close games | Games |
|---|
Close = within 8 after three quarters, out of 12 minutes. Camara closing for Portland ahead of bigger names is the coach-trust tell of the month.
28 Nov, in Denver. The Nuggets led by 15 at half-time. By the model, that's a Nuggets win about 89% of the time . The Spurs won the second half by 18.
Also this week: the Grizzlies came back from 15 down at the half in New Orleans (won in overtime), and from 13 down in LA two nights later. The live version will plot it play by play, with the exact play where it flipped; that needs the play-by-play feed.
League-wide, teams were −1.8 a game on the second night of a back-to-back, against +0.4 otherwise (106 second nights). Each bar is a team's margin on night two minus its margin otherwise. Teams have 2 to 5 second nights each, so every team call here is NOT YET.
Fall apart: Kings −29.0 on second nights (−8.2 otherwise), Raptors −9.8 (+8.1), Suns −6.8 (+4.9). Don't care: Hornets +8.0 (−8.4), Blazers +8.3 (−4.9). Must Watch flags any game where one team is on night two.
| Call | Made | Result |
|---|---|---|
| Hornets offence stays above average | Week 1 | ✓ 6th in offensive rating |
| Bulls' 3P% defence is fake | Week 1 | ✓ opponents shot 36.8% the rest of the way |
| Celtics: slump, not a slide | Week 1 | ✓ went 56-23 |
| Spurs are real | Week 1 | ✓ 4th in net rating |
| Reaves's role is real but cools | Week 1 | ✓ 24.1 per 75 |
| Kyshawn George: efficiency fake, role real | Week 1 | ✓ |
| Curry's hot start holds | Week 1 | ✗ 38.8%, needed 40% |
| LaMelo's heater cools | Week 1 | ✓ |
| Cold starts bounce back (White, LaVine, Brunson) | Week 1 | ✗ only Brunson did |
| Lakers' shot-making shrinks | Week 1 | ✓ |
| Chet's blocks come back | Week 1 | ✓ |
In this sample these are shadow-season calls, graded with hindsight. Live, each call is graded at its deadline and this page updates.
We played the rest of the season and the whole playoffs 10,000 times. The Thunder take the title in 55.7% of runs. The most likely Finals is Thunder v Knicks (16.6% of runs), and the Thunder win that in 5 .
How it works: team ratings from point differential adjusted for opponents, blended with last season until about 15 games; the real remaining schedule; play-in; 2-2-1-1-1 home court; 10,000 runs. A projection, tagged SIMULATION RESULT, not a Lab result.
J Swish's ladders after month one. Each list is sorted by average rank across the stats shown, not by a single score. The 65-game flag says whether a player is on pace to qualify .
High-leverage possessions only, season to date. Bars are clutch points; the label is true shooting %. Blue = above 58% TS, grey = below.
Where each player's shots come from, last season against this one. The biggest change is called out on the right.
Under the new lottery the bottom three get fewer balls than teams four to ten (and can't fall below pick 12). So the Wizards, the worst team, have a 5.0% shot at #1, while the Kings, seventh-worst, have 8.6% .
Each player's per-75 profile this season against every season since 2000-01 (1,500+ minutes). The bar shows how close the match is: longer = closer. Style, not quality: a comp says he plays like that season, not that he's as good.
The plain win chance, then what it drops to once the schedule bites (second night of a back-to-back, third game in four nights).
J Swish's five picks this week with his confidence. Make yours; they're saved on this device only. The season leaderboard comes when the newsletter poll is live.
Price = Lab Net (Lab Offence + Lab Defence, per 100). BUY = shooting well below his own normal on similar shots, or a role that's growing. SELL = shooting well above his normal, which rarely lasts.
In this sample the calls are replay-graded on the rest of the 2025-26 season: did the BUY's shooting rise and the SELL's fall? Live, each call is graded at its deadline, right or wrong, and the hit rate stays on this board.
LAB IT · J Swish is a synthetic, AI-generated host. Not affiliated with the NBA or any team.