Rest, load and outcomes

1 Question

This note asks two separate descriptive questions about a common theme: what an NBA team's schedule and its roster availability look like as conditioning variables known before tip-off. First, does the back-to-back load a team carries across a season show up as a measurable production or margin tax, and how is that tax spread across 90 published team-seasons. Second, how concentrated is a team's win outcome on its single most-indispensable player, measured two ways from two one-season windows, and how often do the two name the same player.

Both measures are pre-tip-off conditioning variables. Rest days are fixed once the league schedule is released; which player is or is not available is known, with injury reports, before a game starts. Neither measure is presented here as a forecast, and neither is tested here as a live input to a probability model.

2 Data and definitions

The Schedule Fatigue Tax module (novel_schedule_fatigue_tax.json) reports 90 team-seasons across three NBA seasons, 2023-24 through 2025-26, one row per team-season with its back-to-back game count, back-to-back frequency, and two tax variants. The site manifest's entry for this module carries no as_of date -- date not published -- so the module's own schedule_inequality and per_season_b2b_freq_range fields are the only timing context available beyond the pooled season range.

That module's source_artifacts field points to schedule_density.json, which supplies the underlying per-36 box-composite means by schedule-density state: 12000 back-to-back player-games, 17531 in a 3-in-4 window, 20120 in a 4-in-6 window, and 40157 rested player-games (2 or more days rest, no compression), pooled from 77744 total rows and 7222 team-games across the same three seasons. schedule_density.json's manifest as_of is also not published. Its own methodology states the three density classes overlap -- a back-to-back that closes a cluster is also a 3-in-4 -- so their player-game counts are not additive.

The Schedule Fatigue Tax module's mechanism_receipt field cites a separately tested effect, -1.73 points per 100 team offensive rating (rested minus back-to-back, n=4732, p 0.005645490426098632), and names the validation ledger as its source. The showcase export, mechanism_ledger_export.json, records two b2b_rest_penalty runs rather than one, both verdict CONFIRMED_LOCAL and both carrying the corpus tag player_boxscores_2024_25_2025_26: the n=4732 run just cited, and a second at n=7192 with effect -1.955 and p 0.0001174325580081608. Each run stores its own rested-versus-back-to-back split in an evidence string, and both carry a null as_of, unstamped by the ledger's own run_ts convention.

The Load-Bearing Index module (novel_load_bearing_index.json) reports 30 teams, each with two independently sourced estimators: estimator_a is an Elo on/off win-probability comparison for the 2024-25 season, and estimator_b is a raw win-rate comparison between games a player was active and games that player was missed, for the 2025-26 regular season through 2026-04-12. These two as_of windows (as_of.estimator_a, as_of.estimator_b) are printed inside the module itself and are not the same population; that distinction is kept separate throughout this note.

nba_form_curves.json carries no back-to-back, rest, or schedule field in either its insights or showcase build, so it contributes nothing here. nba_consistency_profiles.json (as_of 2026-04-12) does not condition its coefficient-of-variation numbers on rest either, but its own caveat and not_this fields explicitly flag rest as an uncontrolled variable in that module, consistent with the treatment below.

3 Method

SFT is defined two ways in the source module: SFT_credible = b2b_freq * (-1.73 pts/100 ORtg), using the tested mechanism_receipt anchor, and SFT_descriptive = b2b_freq * (-0.315 composite/36), using the raw pooled per-36 delta from schedule_density.json. Both variants use the same team-season back-to-back frequency as the only team-specific input; the multiplier itself is a single leaguewide constant, not something re-estimated per team. A team's SFT ranking is therefore a rescaling of its own back-to-back frequency, not an independently measured team-specific fatigue effect.

The rest-state margin test behind the -1.73 anchor compares average scoring margin on zero rest days (a back-to-back) against average margin on one or more rest days, within one corpus of player boxscores. It is not a controlled experiment: the schedule-maker chooses which games fall on a back-to-back, and travel, opponent quality and home-away patterns are not held fixed between the two rest states.

LBI is defined as LBI_a = the largest, over a team's roster, of win probability with a given player minus win probability without that player (Elo on/off), and LBI_b = the largest, over a team's roster, of win rate while a given player was active minus win rate while that player was missed (raw record). Agreement is flagged true only when the two estimators name the same player for a team. Because estimator_a draws on 2024-25 and estimator_b draws on 2025-26, a match or mismatch between them is a directional cross-check across two different one-season samples, not a paired measurement of the same games.

4 Results

Two tables below describe rest state at the player-game level; a third ranks team-seasons by SFT; two more rank teams by LBI, one per estimator.

The tested margin effect and the raw per-36 deltas point the same direction: less rest associates with lower measured production or margin. The SFT module converts the tested anchor into a team-season index by multiplying it through each team's own back-to-back frequency.

Novel Schedule Fatigue Tax

All 90 published team-season Schedule Fatigue Tax values, 2023-24 through 2025-26.

The Load-Bearing Index results below rank the same 30 teams under each estimator separately. Only 1 of 30 teams is named by both estimators (IND, Pascal Siakam), an agreement rate of 0.0333 (agreement_summary.n_agree over agreement_summary.n_teams_both_estimators).

CHA, LAC, MIL and DET each appear in the estimator B top five while sitting outside the estimator A top five for the same team, which illustrates that the two estimators frequently disagree on magnitude as well as on which player is named.

5 Robustness and what would falsify this

The tested rest effect is recorded twice, at n=4732 and at n=7192, both landing on a negative margin difference of the same sign and similar size (-1.73 and -1.955 points per 100 offensive rating, p 0.005645490426098632 and 0.0001174325580081608). What that does and does not establish is worth stating exactly. It establishes that a second, larger recorded run reproduced the sign and rough magnitude of the first. It does not establish an out-of-sample replication: both runs carry the same corpus tag, and the ledger records no relationship between the two row sets, so whether the 4,732 rows are a subset of the 7,192 is simply not recorded and cannot be assumed either way. The project's standing discipline treats a single-corpus result as unconfirmed until it clears a second, independently sourced corpus; this note does not claim that second corpus exists for b2b_rest_penalty.

A falsifying result for reading SFT as team-specific, rather than schedule-inequality-driven, would be a team-by-schedule interaction test showing that some teams absorb a back-to-back worse or better than the leaguewide -1.73 anchor predicts. No such interaction test is published here, so the current ranking should be read as which teams were exposed to more back-to-backs, not which teams are more fatigue-prone.

For LBI, a robustness check would run both estimators on the same season for the same 30 teams; the agreement rate would be expected to rise if the underlying signal were estimator-independent. At 1 of 30 (3.33 percent) across two different seasons, the current data does not support treating either estimator's top-named player as validated. Running estimator B on the 2024-25 season, matching estimator A's window, would be the natural next check.

6 Limitations

This note does not establish that any specific team-season carries a fatigue effect different from the leaguewide average, because SFT is that average effect multiplied through a team's own schedule exposure, not a team-specific estimate. It does not establish that a team's win outcome causally depends on the named load-bearing player, because both LBI estimators are raw associational comparisons across games that player did or did not play, with roster, opponent, rest and blowout-substitution differences uncontrolled. And it does not establish one shared population for the load-bearing comparison, because the two estimators are drawn from different NBA seasons.

• SFT ranks team-seasons by a constant leaguewide effect times each team's own back-to-back frequency, so the ranking is mathematically the same as ranking by back-to-back frequency alone.
• The two b2b_rest_penalty runs (n=4732 and n=7192) share one corpus tag and the overlap between their rows is not recorded, so they are not two independent samples; rest days are chosen by the schedule-maker rather than assigned at random, so either run is an association, not a randomized comparison.
• The two LBI estimators cover different seasons (2024-25 versus 2025-26 through 2026-04-12), so their 1-of-30 agreement is cross-season corroboration, not a paired estimate.
• Both measures aggregate to the team level and do not adjust for roster changes, opponent strength, minutes distribution, or blowout-driven lineup substitutions.
• Estimator A publishes no game-count denominator, so its ranking cannot be weighted by sample support the way estimator B's active and missed game counts allow.

7 How to read this on the site

The observation-dependence inspector explains how sample size and observation windows change what a descriptive number can support; the same caution applies to every n and every as_of window cited above. The published figure for the Schedule Fatigue Tax module sits alongside its underlying rows in the analytics library. The nba-schedule-compression-profile analysis exposes the same schedule_density.json rows behind an equal-weight compression index, a different summary of the same schedule shape, not a duplicate of SFT. The star-removal-team-win-probability and lineup-proxy-active-missed-record analyses expose the scenarios and player tenure windows feeding LBI's two estimators. Read a team's SFT or LBI number next to its row there before treating either ranking as a settled profile.