# Division Picks Land at 27% and Andrew Erickson Owns Both Ends of the Week | MissedTakes Daily

*Published 2026-08-30 — MissedTakes*

The grading run that landed this week was the biggest the site has processed: **2,534 takes settled in seven days**, against 461 new takes logged. And for five prediction types, the week's grading *is* the whole record. Division Winners (351 graded), Trade and Signing Calls (349), Sleeper Calls (308), Conference Champions (187) and Fantasy Bust Calls (84) all show the same graded totals for the last seven days as they do for all time. Nothing in those buckets had been settled before this week.

That makes the aggregate numbers on those five types the cleanest read we have on how the field did against results, so start there.

## Season-shape picks were a bloodbath

Division Winners graded out at **27.4% average accuracy** across 351 calls. Conference Champions were worse: **25.1%** across 187. Those are the two lowest-accuracy types of the five, and neither is a coin flip — picking one of four teams and one of two conferences finalists respectively means the naive baseline is low too, which is why the skill deltas stay positive (+19.64 and +6.66). Pundits beat a dumb baseline. They still got roughly three of four division calls wrong.

The team-level splits explain where it went. Grouped by the team being picked, calls on the Eagles (22 graded) and Broncos (15) hit at 100%, the Jaguars at 85.7% and the Seahawks at 84.6%. Calls on the Ravens (35 graded) hit at **0%**. Chargers: 0%. Commanders: 0%, with a skill delta of -156.6, the worst mark on the board. Bills backers went 6.7% on 30 calls. The field wasn't uniformly bad — it was badly wrong about a specific handful of teams and right about another.

Individually, on 8+ graded division calls, [Marcus Mosher](/pundits/marcus-mosher) led on accuracy at 60% (10 graded), with [Jamie Erdahl](/pundits/jamie-erdahl) at 50% on 14 and Kyle Brandt at 45% on 20 — Brandt's +65.6 skill delta is the best of any pundit with 20 or more. At the other end, Nick Wright went 7.1% on 14 graded division calls with a -42.43 skill delta, and Colin Cowherd 13.2% on the largest sample of anyone, 53 calls. On conference champions, Rich Eisen graded at 7.7% across 26.

![Division-winner calls by team picked](/posts/visuals/5013410c-8b5e-4246-827e-c6e381c75796.png)

## Transaction calls were the one place the field was worse than naive

Trade and Signing Calls came in at 51.3% accuracy — respectable-looking — but with an average skill delta of **-15.07**, the only negative among the five newly settled types. Coin-flip-looking accuracy on claims a naive guess handles better is exactly what the skill metric exists to catch. Chris Rose graded 0% on 12 transaction calls (-100 skill), [Tom Strachan](/pundits/tom-strachan) 19% on 21 (-52.38), Jake Ciely 26.3% on 19. Rich Eisen was the best of the group at 77.3% on 22.

Sleeper Calls landed at 52.6% (+12.32) and Fantasy Bust Calls at 58.3% (+27.93) — the fantasy buckets held up better than the season-outcome ones.

Mock Draft Picks (501) and Game Outcome Picks (462) were the two biggest slices of the week's grading, but both are partial cuts of much larger pools (2,913 and 1,070 graded all time), so we can't cleanly isolate this week's performance in those types. We're not going to pretend otherwise.

## One opinion, best call and worst call

[Andrew Erickson](/pundits/andrew-erickson) sits atop both the week's best-calls list and its worst-calls list — with the [same take](/opinions/3809e587-6fa2-464b-ab98-fd96d353c1bd). His grab-bag of Jordan Mason, Broncos-Titans, Garrett Wilson and depth-chart calls graded 100 with a +200 skill delta on its Fantasy Bust component and 0 with -200 on its Sleeper component. That's not a scoring bug; it's what happens when one segment contains several distinct claims and each gets settled on its own merits. His career splits say the same thing: 88.89% accuracy on 11 graded bust calls, 44.67% on 37 sleeper calls.

[Mitchell Eisenstein](/pundits/mitchell-eisenstein) pulled the same double, 100 and 0 on [one opinion](/opinions/7fedb68a-c73b-4823-9b46-158905a559ef). [Jourdan Rodrigue](/pundits/jourdan-rodrigue) and Tom Strachan posted clean 100s; [Cynthia Frelund](/pundits/cynthia-frelund) and [Ryan Wormeli](/pundits/ryan-wormeli) posted clean zeros.

## Nobody changed their mind

The reversal tracker is empty this week. Zero logged mind-changes against 461 new takes. Worth stating plainly rather than dressing up.

## Consensus: the field still can't price Kansas City

The 2025 win-total consensus board is unchanged and still ugly at the top. The field was **6.3 wins too high on Kansas City** across 13 calls — every single one of those calls sat between 10 and 15 against an actual 6. Rich Eisen alone had them at 10, 12, 10, 15 and 15. The field was 5 wins *low* on New England (11 calls, actual 14), Seattle (7 calls, actual 14) and Denver (6 calls, actual 14), and 4.8 high on Washington.

On awards, the 2025 MVP ballot was the field's best work: 26 of 100 picks went to Matthew Stafford, who won. Offensive Rookie of the Year split 5-5 between Ashton Jeanty and Tetairoa McMillan out of 29 picks, with McMillan winning.

Standings, sample sizes and confidence flags are on the [leaderboard](/leaderboard) — note that nine of the top ten are still tagged "limited" confidence.