Field Yates leads a below-zero DEF field in 2026 Week 2 half-PPR accuracy

How to read a leaderboard below the random-order benchmark

The skill score measures how well a ranking matched actual scoring relative to a random order: 0 represents random order and 100 represents a perfect match. Most rankers land between about 5 and 35 because weekly fantasy scoring is noisy. These are not prediction-success percentages. With even the highest displayed result below 0, this week's ordering identifies relative performance within a negative-scoring field. Reading the placement without the scale would miss the report's central finding: the leading product still fell below the random-order benchmark. ^[2][1]

The stated 4.7-point lead over FantasyPros Consensus describes the separation at the top; the 29.4-point spread describes the full field. Those gaps answer different questions. The former shows how far Yates finished ahead of the next distinct displayed product, while the latter captures the distance between the highest and lowest results. Neither gap is a test of reliability or evidence of future superiority. Together, they show that a shared position below the benchmark did not mean the products produced interchangeable scores. ^[1][2]

Weekly ties conceal different observed histories

The middle of the board was not a sequence of uniquely separated results. Draft Sharks and Daniel Dopp shared a -29.4 score, while Liz Loza and FFToday shared -31.8. These ties matter because a displayed order should not imply an accuracy advantage where the measured scores are equal. The useful next question is how each tied result compares with its own product's record. That shifts the reading from placement alone to whether the same weekly score represented a similar departure from each product's observed average. ^[1]

For Loza and FFToday, the answer is no. Loza's through-Week 2 average was -8.8, compared with FFToday's -21.8, both over 2 eligible weeks out of 2 published weeks. Their identical weekly result therefore sat 22.9 points below Loza's average but 10.0 points below FFToday's. The tie describes their Week 2 accuracy, not matching records through the week. It also shows why distance below an average cannot substitute for the weekly leaderboard: a larger departure from a product's own history need not produce a lower weekly score than another product's. ^[1]

The weekly order differs from the through-week picture

Yates's leading result also fell below his own observed average. His -17.6 weekly score compared with a -4.7 average through Week 2, a stated shortfall of 12.9 points. FantasyPros Consensus recorded the same 12.9-point shortfall against its own average. Their shared direction of departure gives the top of the board additional context: neither leader was outperforming its through-week record. These comparisons describe the products' observed averages through the reporting week, rather than independent expectations for what either should have scored. ^[1]

Elsewhere, the weekly order and the through-week averages tell different comparative stories. Eric Moody finished above Daniel Dopp in Week 2, at -24.7 against -29.4. Yet Dopp's through-week average of -6.5 was above Moody's -7.6. Both averages covered 2 eligible weeks out of 2 published weeks, so the contrast is not an unequal-history comparison. It demonstrates why the weekly table should remain a weekly table: the ordering of these products in this result does not reproduce their ordering by observed average through the same week. ^[1]

The lowest score and the largest departure are different findings

At the lower end, ESPN app projections scored -45.9 and Footballguys (PPR board) scored -47.1. Their through-week averages were -17.1 and -22.4, respectively, each based on 2 eligible weeks out of 2 published weeks. Footballguys was lower on both measures, but the supplied weekly-to-average comparisons show a different ordering: ESPN app projections was 28.8 points below its average, while Footballguys was 24.7 points below. The lowest weekly result therefore was not the larger departure from its own observed record in this comparison. ^[1]

FF Toolbox (PPR board) requires a different interpretation of its average altogether. Its weekly score and through-week average both stood at -27.1, but the average included 1 eligible week out of 2 published weeks. That equality does not demonstrate consistency across eligible weeks; the average contains only the eligible weekly result. Its position above Draft Sharks and Dopp is a valid description of this week's displayed scores, while its average supplies less historical context than their averages based on 2 eligible weeks. ^[1]

Publication status and the takeaway for fantasy managers

None of the 14 displayed results is marked provisional; the supplied observations identify the Week 2 results as published. Here, provisional status is a publication-status distinction, not another accuracy score. The evidence does not specify the criteria for assigning that designation, so its absence should not be read as proof that revisions are impossible. Published results remain subject to the publication and correction process. The methodology also specifies weekly locks and permitted published observations, with corrections and permission changes able to alter the available evidence. ^[1][2]

For a fantasy manager, this sample establishes how the named ranking products compared with actual DEF scoring in 2026 Week 2 half-PPR. It identifies the leader, the stated gaps and cases where identical weekly scores mask different observed histories. It does not evaluate an entire service, explain why a product finished where it did or establish season-long superiority. The seasonal participation gate is separate, and future performance remains unresolved. The useful conclusion is a measured weekly comparison—not a promise about which ranking will be most accurate next. ^[1][2]

Sources

1. Fan50 evidence record · Checked October 29, 2026
2. Fan50 methodology · Checked November 5, 2026