Up: 2026-27 Fantasy Basketball Encyclopedia
Related: Using projection disagreement in fantasy basketball · Fantasy basketball decision calibration · How NBA injury reports should change fantasy decisions
Evidence ladder
| Grade | Evidence | Appropriate use |
|---|---|---|
| A | NBA/team transactions, rosters, injury reports, schedules, contracts, coach/player remarks, and official statistics | Identity, status, dates, observed performance, and direct statements |
| B | Stable statistical databases and transaction/depth-chart services with visible update dates | Cross-checks, historical rates, eligibility, and provisional roster structure |
| C | Specialist projections, rankings, podcasts, videos, and reported analysis with a named author and date | Attributed expectations, alternative assumptions, and expert disagreement |
| D | Identified social posts and substantive community analysis | Leads, hypotheses, market sentiment, and overlooked mechanisms |
| E | Unsourced screenshots, anonymous claims, copied aggregation, or pages whose season/update date cannot be established | Discovery only; never a factual foundation |
An official source can still be incomplete: a roster page does not declare a rotation, and a coach's July remark is not a binding minute allocation. Conversely, a careful specialist can make a strong forecast without becoming the source of record for a transaction. Evidence grade describes what a source can establish, not the prestige of its logo.
Four separate confidence labels
Every material player projection should distinguish:
- Roster confidence — confidence that the player will be on the team and eligible for the projected role.
- Minutes confidence — confidence in the active-game workload and rotation position.
- Rate confidence — confidence in per-minute or per-possession production and shooting efficiency.
- Availability confidence — confidence in games played, return timing, maintenance, or restrictions.
The overall label is the weakest high-leverage component, not an average that hides a fatal uncertainty. A stable star with unresolved surgery timing can have high rate confidence and low availability confidence. A healthy rookie can have high roster confidence and low minutes and rate confidence.
Freshness states
- Current — checked against the information set dated on the page, normally within seven days during the offseason transaction cycle.
- Usable but aging — still relevant, but one or more later roster events require an explicit adjustment.
- Historical baseline — intentionally describes 2025-26 performance or an earlier forecast.
- Stale or mislabelled — presents itself as current while retaining contradicted teams, roles, or season assumptions.
The access date and the source's own publication/update date are both recorded when available. A dynamically rendered page can change without preserving a public archive, so the encyclopedia records what was visible at extraction time.
Social, Reddit, and video rules
A social post or Reddit comment is attributed to its author or handle and labelled as analysis, sentiment, or an unverified claim. Engagement is not evidence of accuracy. A useful post supplies a mechanism that can be checked: a roster opening, coach quote, lineup pattern, injury update, or statistical comparison.
Videos and podcasts are indexed at the episode level. The note records title, creator, publication date, URL, relevant timestamps when recoverable, format assumptions, players discussed, and the episode's principal claims. A transcript summary is a paraphrase unless a short quotation is necessary and verified.
Projection versus rank
A projection estimates an outcome such as games, minutes, or a stat line. A ranking transforms projections through a declared format, population, and horizon. An ADP describes acquisition behavior. A take is an analyst's judgment. These fields are never substituted for one another.
Published ranks are retained with their native format. Cross-source comparisons follow Using projection disagreement in fantasy basketball: normalize dates, games, minutes, scoring rules, population, and per-game versus total value before interpreting disagreement.
Correction policy
Corrections preserve the earlier claim when it matters for understanding changing market belief, then add a dated correction. Silent replacement is reserved for spelling, broken links, and mechanical errors. A player changing teams requires updates to the player dossier, old and new team depth charts, the roster ledger, minute totals, and any affected rankings—not merely one line.
Open questions
- Which dynamic ranking pages expose stable snapshots that can be archived without violating their terms?
- How should source accuracy be scored when public preseason projections are revised repeatedly before opening night?
- Which community accounts have enough timestamped history to graduate from lead generation to a quantitatively weighted signal?