← Analysis
SR-2026.4August 18, 2026

The sit edge held. Start calls show a small edge too.

In July we published a sit-skill finding based on 1,875 resolved 2025 claims. Several readers pushed on it. Two questions came up repeatedly: will the edge hold with more data, and what does the position mix do to the numbers? Here is the answer.

All figures below come from scripts/base-rate-study.ts, run August 15 against the full graded corpus (SR-2026.4 bar, ruleset start_sit_v3). The script is committed and reproducible.

The baseline

53.1% of all discussed players across the full corpus finished below the startable threshold, top-12 for QB and TE, top-24 for RB and WR. That is the blind-sit rate: what a strategy of calling "sit" on every player an analyst mentions would score.

By season:

SeasonBlind-sitn (sits)n (starts)
202351.2%7103,395
202453.2%3,02714,048
202553.3%3,09815,804
All53.1%6,83533,247

The baseline is stable across seasons. A bit more than half the players analysts discuss do not have startable weeks. Calling sit on all of them gets you to 53%.

The sit edge survived and narrowed

Real sit accuracy across all seasons is 62.9%. Against the 53.1% baseline, that is a +9.9-point edge.

By season:

SeasonSit accuracyBaselineEdge
202365.5%51.2%+14.3
202463.2%53.2%+10.0
202562.1%53.3%+8.8

The edge was widest in 2023 (small sample, 710 calls). It narrowed as the corpus grew. With three full seasons it sits at roughly 9 to 10 points, not the 12.3 the July article reported for the early 2025 cohort. The July figure came from the smaller early-2025 window, and we said then the number would move as the corpus grew.

The sit edge survives the baseline test: analysts outperform the blind-sit baseline by 9.9 points, while start calls outperform the corresponding blind-start baseline by 2.0 points.

Start calls: a correction

The July article called start-call accuracy "indistinguishable from noise." That was correct for the data at the time. With a larger corpus the picture clarifies.

Real start accuracy across all seasons is 49.0%. The blind-start baseline (the fraction of discussed players who do have startable weeks) is 46.9%. That is a +2.0-point edge.

By season:

SeasonStart accuracyBlind-startEdge
202351.8%48.8%+3.0
202448.9%46.8%+2.2
202548.4%46.7%+1.7

Start calls are not random. They are also not the skill center of this corpus. A 2-point edge on 33,000 calls is real but small. The sit edge at roughly 10 points is the story.

Position mix matters

The blind-sit rate is not the same at every position. Players analysts discuss have different baseline rates of finishing below the startable threshold:

PositionBlind-sitSit accuracySit edgen (sits)
WR58.1%67.8%+9.72,756
TE55.0%66.6%+11.6814
QB53.0%61.6%+8.71,421
RB45.1%55.0%+9.91,844

WR discussions have the highest sit-failure rate (58.1% of discussed WRs do not have startable weeks). RB discussions have the lowest (45.1%). A show that mostly discusses RB start/sit decisions faces an easier blind-start environment and a harder blind-sit environment than one heavy on WR discussion.

The edge metric accounts for this. It compares each show's real accuracy against its own discussed-population baseline.

Per-show sit edges

Shows with at least 30 resolvable sit calls in the corpus, sorted by name. The baseline column is each show's own discussed-population sit rate, not the corpus-wide figure.

ShowSit baselineSit accuracySit edgen (sits)
4for4 Football Podcast53.3%58.8%+5.5102
ESPN Fantasy Focus Football48.3%59.4%+11.169
ETR Podcast (Establish The Run)43.3%59.0%+15.761
Fantasy Alarm Radio52.5%62.4%+9.9356 †
Fantasy Football Counselor49.0%63.2%+14.2133
Fantasy Football Today53.4%62.5%+9.02,974
Fantasy Points Podcast53.4%59.9%+6.4675
FantasyPros Podcast55.2%62.4%+7.2622
Good Old Boys Fantasy Football49.3%54.9%+5.682
PFF Fantasy Football Podcast58.3%64.5%+6.2214
Razzball Fantasy Football56.3%62.1%+5.8219
RotoBaller Fantasy Football50.3%51.6%+1.295
RotoWire Fantasy Football54.0%68.2%+14.1176
The Fantasy Footballers52.2%69.1%+16.8446 ††
The Ringer Fantasy Football Show50.5%70.9%+20.479
Underdog Fantasy Football48.0%71.0%+23.069
Yahoo Fantasy Football Podcast50.8%67.3%+16.4382

† Fantasy Alarm Radio has 30 past-season sit calls that have not yet resolved. Those calls are not reflected in the accuracy or edge figures above and could shift the numbers on resolution.

†† The Fantasy Footballers has 21 past-season sit calls that have not yet resolved.

What comes next

The 2026 season kicks off September 9, and the first full Sunday slate is September 13. The 72 per-week thresholds are locked under SR-2026.4 through the season. No threshold changes, no retroactive regrading. When the season ends we will evaluate whether the startable-week bar warrants adjustment for SR-2027, and publish the proposed changes before any regrading happens.

The sit-edge finding is now on three seasons of data. We will report where it lands at the end of the 2026 season.

Subscribe to The Sunday Evidence Email for in-season updates as each week's claims resolve.

Verify it yourself

Every number in this article traces back to a receipt. See the live board, or how the bar is actually computed.