Pelle Larsson Is Available; the Feed's Timeline Is Not: A Data Forensics Read of Miami's Point-Guard Problem

The Rotoworld item carried a headline of eleven words: "Pelle Larsson (finger) available vs. Charlotte." Under two hundred words in total. A routine fantasy-feed blurb of the kind a streamer reads with morning tea before locking a lineup. I read it and inside thirty seconds three separate dates collided.

The headline says Saturday, against Charlotte. One point inside says the regular-season opener is Wednesday against Minnesota, and that Larsson did not play the final two preseason games. A second point describes a Saturday loss to the Timberwolves in which Larsson started. A third gives last season's average as 11.4 points per game; a fourth gives the same season as 11.5.

Same page. Same player. Same season. Three timelines and two averages.

When a feed contradicts itself three times inside two hundred words, the correct output is not a projection — it is a provenance audit. Every metric needs provenance, or it becomes a story with a spreadsheet, and a story can be carried anywhere by anyone.

A fantasy news feed is not entertainment copy. It is a decision instrument. Somebody burned a roster spot on that blurb, somebody moved a season-long wager, somebody updated a depth chart note. If three structural faults sit under two hundred words, those faults travel straight into the decision. So I coded the feed the way I code a box score: all 42 information points tagged IP-1 through IP-42, each given a confidence mark, cross-checked against the others, and discounted wherever they collided.

The discipline is not new. In August 2026, after India lost 70-84 to Kazakhstan in Beirut, I hand-coded forty minutes of FIBA play-by-play into a spreadsheet. India took 22 mid-range jumpers at 35 percent and only 12 three-point attempts at 25 percent; the model said swapping ten mid-range looks for threes added six points. I did not write that story until I had rebuilt the timeline — I rebuilt the play-by-play before I trusted the mid-range story. In the 2026 lockdown the same habit ran backwards: 47 national-team box scores from 2026 to 2026 into a SQL database, which showed the 2026 Wuhan win was an outlier at 52.1 percent eFG against a 42.3 percent average everywhere else, with no structural improvement following it. An outlier called a trend is the easiest lie a data man can tell.

Run this feed through that sieve.

The first fault is the timeline. The most plausible reconstruction is that this page is a composite of two or more blurbs written in different weeks; IP-16 explicitly notes a truncated duplicate of the Larsson item sitting at the end of the page. Any per-game tactical conclusion drawn from this text therefore carries low confidence. That is not pedantry. If you read the availability sentence as current and it belongs to a different week, your decision is correct for the wrong Saturday.

The second fault is the statistic. 11.4 against 11.5 points per game. A tenth of a point is trivial in magnitude and I will not spend two hundred words on it. Its importance is directional: it proves two separately written blurbs were stitched together rather than one report edited. When a page carries two averages for one season, every other number on it — 54 starts, 26.4 minutes, 0.9 threes — takes a reliability discount.

The third fault is the largest, and it is architectural rather than basketball. In this feed's world Giannis Antetokounmpo plays for Miami, Klay Thompson was bought out by Dallas and signed in Miami, LeBron James sits on a favourites list with Philadelphia and Kawhi Leonard with Toronto. This is either a constructed 2026-27 timeline or a badly corrupted feed. No external cross-verification is possible, so everything below is internally consistent reasoning about the scenario as presented — not a factual claim about the real NBA. Confidence that the inconsistency exists: high. Confidence about which version is correct: low.

Even after those three warnings, one genuine, diagnosable basketball fact survives, and it is IP-13: Giannis and Bam Adebayo can initiate the offense and probably will have to, because Miami lacks depth at point guard. Point-forward initiation as a point-guard workaround is an established archetype with established failure modes — live-ball turnover inflation, late-clock stagnation when the initiator is a non-shooter and the defence goes under, and playoff blitzing, because a switching defence will happily force Adebayo or Giannis into a two-man game with no third handler to punish the rotation.

That is where Larsson's real value sits. Davion Mitchell played 33 minutes in the opener, handed out a game-high 10 assists and committed six turnovers. Dru Smith is out with a hamstring. There is no third pure ball handler. Larsson logged four assists in the opener, five assists in 25 minutes, and the feed itself says he took on more playmaking responsibility. On this specific roster that is a non-replaceable function, and it is a stronger basketball argument for his minutes than his scoring. Mitchell's ten assists are the pretty number; the six turnovers are the actionable one.

On the other side stands the spacing constraint. Giannis and Adebayo both operate inside the arc, so Miami's fix is shooting around them rather than structural innovation — and that shooting is Klay Thompson, described as still an elite shooter and exactly what the Heat need. Larsson is not a spacing solution. At 0.9 threes per game in 26.4 minutes he is a low-volume, low-gravity perimeter profile, and the feed states plainly that he has not found his shot this season. In a Giannis-centric offence the wing who cannot shoot is the wing who gets squeezed, however good his passing.

On playoff transferability I will be blunt about the limit. Committee initiation is regular-season survivable and playoff-fragile. Larsson's defensive switchability as a 6-foot-5 wing is the trait that usually survives playoff downsizing, but Stage-1 offers one steal and one block on a single-game line. I will not assess his defensive value; I am flagging it as unknown. I will not manufacture what is not there.

The player data is clear and unromantic. Second season: 11.5 points, 3.5 rebounds, 3.4 assists in 26.4 minutes, with 54 starts in 70 games. That is not a breakout profile and not a fringe one — it is a nominal starter with solid volume. And that is exactly where Klay's arrival cuts. Klay taking the starting wing spot is a demotion signal for starting-lineup value but not necessarily for total minutes, because Miami needs a bench creator and Larsson is one.

A pattern emerges that the feed does not state: a four-year extension signed last month, immediately followed by a bench re-slot. Call it paid-then-reslotted. The team bought role flexibility, not a starting job.

The age curve supports it. The feed does not give his age, but a completed second season and an entry into a third places him around 24 or 25, and his game is skill and connective rather than explosiveness-dependent, which ages well. A four-year extension at that point is the classic buy-out-the-ascending-years structure.

The most underrated data point is durability. He played through illness and a finger sprain. For a team whose core is injury-adjacent — Adebayo's groin tightness and his absence from the final two preseason games is the latest evidence — that availability is a genuine asset.

And then the number every fantasy blurb waves around: 15.3 points on 55.6 percent shooting across four career games against Charlotte. Four games is a noise-level sample and I am excluding it from any projection model, because opponent personnel, minutes and context are all unrecorded. Sample size, error bar and the alternative explanation belong in the same paragraph as the finding.

The shooting slump, by contrast, is a current problem rather than a narrative: 4-of-15 in the opener, with six of seven free throws only partly offsetting it.

Here is the honest gap. Stage-1 provides no TS%, no eFG%, no PER, no USG%, no plus-minus, no on/off, no EPM, and no team OffRtg, DefRtg or Pace. So I will not call Larsson efficient or inefficient, because I have no basis for the call. I will not claim Miami's offence is working or failing either. A man who builds a 47-game database to prove a win was an outlier does not call a wing efficient off four games.

No salary figures are given, so I stop at structural inference: with Giannis, Adebayo, Wiggins and Mitchell on the books, Miami is almost certainly over the cap and operating in apron territory. Nick Richards' non-guaranteed deal, guaranteed only by an Opening Night roster spot, is a contract-driven incentive — his performance may not represent his true level, and any model needs an incentive adjustment before using his numbers.

Now the turn where I have to argue with my own frame. The blurb's entire structure assumes availability is the story. My read is the opposite: availability is the least informative variable here. He was available, he played, he will be available — a stable fact already priced in, because he has played through illness and a sprain. What is unstable, and what is actually moving, is his role. He had 54 starts, signed a four-year extension, and a month later an elite shooter landed on his spot. The question is not whether he plays. It is whether his minutes hold above 24 off the bench.

A second counter-signal, filed as a question rather than a verdict. His 0.9 threes say he is not a shooter, but six of seven free throws in a 4-of-15 game, on low three-point volume, suggests a rim-pressure and transition scoring profile rather than a stationary one. A rim-pressure profile fits better beside two non-shooting bigs than a stationary shooter does, because it pulls the defence inward instead of demanding the space Giannis and Adebayo already occupy. That is a mild positive counter-signal to the conventional read. Confidence is low to medium, so it stays a question. The model started as a question about inefficiency, not a verdict about a loss.

The third inversion hides in the blurb's vocabulary: it conflates two value curves. Larsson's real-basketball value is as a secondary creator and connector — the thing that works even when the shot does not. His fantasy value is as a steals-and-assists streamer — the thing that works only when minutes arrive. "Available, so take him" merges the two, and that is where fantasy advice does its damage.

Three numbers to watch. Mitchell's turnover rate over the next five games: if it stays near six, Larsson's secondary-creator role grows, and that is the most reliable minutes guarantee he has. Adebayo's groin: if it forces a stagger, the split minutes between Giannis and Adebayo open space for Larsson automatically. And the decisive one — whether Larsson's minutes drop below 24 once Klay starts. If they do, the streamer case dies; if they hold, the fantasy value survives.

One last thing, larger than the exercise. A two-hundred-word blurb cut itself three times: in its timeline, in its statistics, and in the shape of its world. No prediction built on that feed can be more reliable than the feed. Next time a fantasy feed tells you a player is available, ask which week wrote that sentence. In a transfer window where every team, every agent and every feed is selling a story, the most valuable skill is not scouting. It is provenance.