DAL
@NYG
WAS
BAL
@HOU
TB
@GB
@PHI
ARI
@IND
SF
TEN
@SEA
@LAR
JAXJavonte Williams — RB, DAL — 2026
Verdict
TARGET at 34.7 (FFC PPR mocks, 2026-07-07 — RB17, mid-3rd round). Williams brings confirmed near-bellcow usage — 65.2% backfield opportunity share, 66.7% of team inside-5 carries, 74.4% third-down snap share before his December shoulder/neck injury — in the NFL's highest-play-volume offense, with year-2 continuity of play-caller/QB/OL and zero backfield capital added all offseason. The market's case against him is fair: four mediocre years before 2025, a visible second-half fade, and a checkdown-only receiving profile — so it prices him a tier below usage-twin Kyren Williams (32.5) and behind two backs with weaker goal-line claims. Why the market is wrong: it is pricing the 2021–24 résumé and the injury-tainted December instead of the 2025 usage signal — his 11 rush TDs were xTD-backed (28 inside-10 carries, 20 of 30 team inside-5), not luck to be regressed, and the team told you the 2026 plan with a 3-yr/$24M ($16M gtd) re-signing in February plus a draft in which they added no RB. Not a MUST-HAVE: HVT sits a hair under the 5/g gate, the receiving profile is genuinely weak (2.7 yds/target), and the thesis rests on one good season.
Bull case
- Bellcow usage nobody outbid: 65% opportunity share, 62%/67% inside-10/inside-5 team shares, 74% third-down snaps (wks 1–14), and the team's answer all offseason was to re-sign him for $16M guaranteed and draft nobody — the claim is contract-confirmed and uncontested (nfl.com/dallascowboys.com 2/21/2026; espn.com draft tracker).
- The TDs are real, not regression bait: 10 of 11 rush TDs came from 28 inside-10 carries — xTD ≈ actual — attached to the league's #1 play-volume offense (Sharp Football, 2025) with a defense upgrade pointing at more positive scripts and more clock-kill carries.
- Price already discounts him: RB17 at 34.7 for last year's RB12 with elite YAC/att (3.6, 896 after-contact yards, 6th) and top-7 MTF — one slot behind Kyren Williams (32.5), the nearest archetype twin, and ahead of only Skattebo (37.9) and Etienne (38.9), both with materially weaker goal-line claims.
Bear case
- One good season in five: 2021–24 = 3.9 career YPC into 2025, −0.60 RYOE and 35.3% success in 2024; and even 2025 faded — 4.9 YPC through 11 games, under 4.0 in 3 of his last 4 (wks 13–16). If the first half was the Adams-scheme mirage rather than the post-ACL restoration, the yardage leg shortens (evidence hierarchy: one-season efficiency change deserves skepticism).
- The receiving profile can't bail him out: 2.69 yds/target, −67 air yards, 0.113 targets per on-field dropback — a PPR floor of receptions without yards, in the lowest-RB-target-share scheme context (13.5% team RB target share). If Jaydon Blue's maturation is real (first-team minicamp reps, coach praise — dallascowboys.com/atozsports/heavy.com, June 2026), passing-down work is the first thing to leak, and with it ~30% of the HVT base. Dallas openly wants to reduce his dependence (Schottenheimer, 6/19/2026).
- Body of work risk: ACL/LCL (2022) + a shoulder/neck stinger he played through for weeks before it ended his 2025 (IR 1/3/26), on a power-contact running style — a December repeat lands exactly in fantasy playoffs, and he turns 27 before the 2027 season, so this price window closes fast.
Projection & comps
| Floor (p20) | Median (p50) | Ceiling (p80) | |
|---|---|---|---|
| Games | 13 | 16 | 17 |
| Carries — yards | 176 — 740 | 240 — 1,070 | 277 — 1,300 |
| Targets — rec — yards | 29 — 20 — 105 | 48 — 33 — 180 | 60 — 41 — 235 |
| Total TD | 6.5 | 10.5 | 14 |
| PPR points | ~145 | ~220 | ~275 |
Build (bottom-up): team ~65 plays/g and ~26 rush att/g with ~22.5 backfield carries/g (team profile + nflverse 2025 actuals, 2026-07-07); median = 15.0 carries/g (down from 15.75 actual — honors Schottenheimer's stated intent to lighten the load) × 4.45 YPC (regressed from 4.77; RBWR 12th, scheme intact) + 3.0 targets/g at 2025's 68.6% catch rate. TDs anchored to xTD: 2025's 28 inside-10 / 20 inside-5 carries at a 60–67% team share on 79 DAL red-zone drives supports ~10–12 expected TDs; he scored 11 (+2 receiving) — median carries the role forward at slight regression (10.5 total), it does not extrapolate a spike. Floor = shoulder/soft-tissue recurrence + Jaydon Blue maturing into a real passing-down role. Ceiling = weeks 1–14 usage (16.5 carries/g, ~76% snaps) held for 17 games. Median ≈ RB10–14 finish vs an RB17 price.
Games risk: medium — 2022 ACL/LCL, shoulder/neck stinger battled for weeks before season-ending IR 1/3/2026 (Todd Archer/ESPN via X, Dec 2025; FantasyPros/RotoWire, 1/3/2026); fully healthy at June 2026 minicamp, held out of one day strategically to evaluate backups (heavy.com, 6/19/2026). Age 26 all season (turns 27 on 4/25/2027, Sleeper DOB 2000-04-25); career touches ~1,051 (858 carries + 193 rec — NFL.com/PFR career page, retrieved 2026-07-07; 2024–25 legs verified vs nflverse cache) — well under the 1,800 cliff line, and 2025's 287 touches is far below the 370 red-flag threshold.
Comps (volume + goal-line lock, thin receiving, good offense): Kyren Williams 2024 LAR (316-1,299-14, 34 rec) · Josh Jacobs 2024 GB (301-1,337-15) ≈ ceiling · Joe Mixon 2024 HOU (245-1,016-11, 36 rec) ≈ median · James Conner 2024 ARI (236-1,094-8, 47 rec) · David Montgomery 2023 DET (219-1,015-13) ≈ floor-median seam.
No external projections in data/projections/ (directory absent) — sanity check limited to market rankings: RotoBaller has him RB21, other outlets "locked-in top-10 with top-5 upside" (retrieved 2026-07-07); this eval's median sits between, above his RB17 price.
Usage profile (2025, DAL, 16 games — nflverse cache + pbp, pulled/computed 2026-07-07 unless noted)
| Metric | 2025 value | Band | Read |
|---|---|---|---|
| Snap share | 68.0% season (weighted; 69.0% simple avg); 75.9% wks 1–14 (snap_counts.csv) | Elite | Late dip (29/57/27% wks 15–17) was injury (shoulder/neck stinger, exited Christmas game, IR 1/3/26), not role loss. Weeks 1–14 is the true signal per the late-split rule. |
| Opportunity share | 65.2% (303 of 465 backfield carries+targets; computed from cache) | Good, top of band | 252 of 383 backfield carries (65.8%); next back: Malik Davis 52 |
| Weighted opp /g | 23.7 ((252 + 2.5×51)/16) | Good, near elite | |
| High-value touches /g | 4.94 ((51 tgt + 28 inside-10 carries)/16) | Good, at elite doorstep | The scoring engine is real, but it's TD-access-driven, not target-driven |
| Inside-5 carry share (team) | 66.7% (20 of 30; computed from nflverse pbp) | Elite | Next: Dak sneaks 4, Davis 2. Goal-line role is uncontested |
| Inside-10 carry share (team) | 62.2% (28 of 45); 10 of his 11 rush TD came inside the 10 | Elite | TD total is usage-backed — xTD ≈ actual, no regression flag |
| Third-down snap share | 62.0% season; 74.4% wks 1–14 (participation.csv × pbp join, run/pass plays) | Good→Elite | He does not leave the field on money downs |
| On-field share by score state | trailing 7+: 63.5% · one-score: 71.9% · leading 7+: 52.5% (same join) | Script-robust | He stays on when trailing; dips only in clock-kill blowouts — the *opposite* of the fragile-grinder pattern |
| Routes /g · route participation | Routes UNVERIFIED; on-field for 66.2% of team dropbacks (452/683, computed) | — | On-field ≠ route (includes pass-pro); treat as upper bound on route participation |
| TPRR | UNVERIFIED; targets per on-field dropback 0.113 (computed) → TPRR est. ~0.14–0.16 | Below Good | The one soft spot: 2.69 yds/target, −67 receiving air yards — pure checkdown diet |
| Targets /g | 3.19 (51 in 16 g; receiving.csv) | Good (low end) | 17 of 51 targets came trailing 7+ — the receiving floor shows up in losses |
| xFP | UNVERIFIED (no provider xFP available); actual 242.8 PPR, RB12 total and 15.2 PPG (RB12, min 8 g) — computed from nflverse cache | RB1-range PPG | Usage supports the scoring level; not a TD-fluke rank |
Efficiency (tiebreaker, not thesis): YAC/att 3.6, 896 yds after contact (6th) (Elite; PFF via web, 2025 season, retrieved 2026-07-07) · MTF 54 rushing, 7th among RBs → 0.21/carry (Good–Elite; PFF, retrieved 2026-07-07) · RYOE +0.62/att (Good, near elite; ngs_rushing.csv) · breakaway rate 5.2% (13/252, computed from pbp) · success rate 45.6% EPA-positive (computed; fringe-Good) · 8+ box rate 22.6% (ngs_rushing.csv). Two-season caveat: 2024 DEN was −0.60 RYOE (ngs cache), 2.9% breakaway, 35.3% success (computed from pbp) — the 2025 efficiency *change* is a one-season sample and gets no projection credit (median uses regressed 4.45 YPC). Within 2025 there was also a fade: 855 yds at 4.9 YPC through 11 games, then under 4.0 YPC in 3 of 4 games wks 13–16 while battling the stinger (thelandryhat.com, Dec 2025). The verdict leans on usage, not the efficiency spike — consistent with the evidence hierarchy.
2024 baseline (DEN, 17 g, cache): 30.2% carry share, 52.2% snap share, 139-513-4 (3.69 YPC), 70 targets/12.75% TS — a three-way committee. The 2025 role change was team/scheme-driven (new team, new deal, Klayton Adams' downhill run game), so it projects forward; the 2024 raw totals are void for projection.
Context (data/team-profiles/DAL.md, built 2026-07-07)
- Continuity is the asset: Schottenheimer year 2 calling, Dak year 4 with him, all 5 OL starters return, league-most 65.9 plays/g in 2025 (Sharp Football), −1.0 PROE (nfelo). Gap/duo-leaning run scheme under OC Klayton Adams was built around exactly Williams' between-tackles/contact-balance profile.
- OL: RBWR 72% (12th), PBWR 66% (12th) — mid-tier; interior (Smith/Beebe/Booker) is the strength, which suits a duo/gap diet. One live variable: Guyton's health at LT.
- Backfield competition: none added — no draft pick, no FA of consequence (vacated: Miles Sanders' 20 carries/8 targets). RB2 is an open camp battle among Jaydon Blue (2025 R5 — team wants him to win it; "completely different player" per Schottenheimer, but 2025 work-ethic/trust issues limited him to 38 carries), Phil Mafah (2025 R7), Malik Davis (UDFA). All are day-3/UDFA claims — per rb.md §9, not a threat to an incumbent on $8M/yr featured-role money.
- Game script: win total 8.5 (FD/DK; BetMGM 9.5, Over money) = neutral-to-lean-positive — clears the ≤7 grinder-landmine line comfortably. The defense added Gary/Lawrence/Downs/Q.Williams after a 511-points-allowed disaster; fewer shootout deficits should shift snaps toward the clock-kill scripts that feed carry volume. And per the score-state split above he *keeps* the field when trailing (63.5% on-field) — his weekly range is unusually script-robust for a low-target back.
- Committee 2×2: high standalone / modest contingent-threat — lead back whose backups are all day-3 unknowns: the "lead back with a weak backup" quadrant, priced like the capped one.
Tripwires (void this eval and re-run)
- Dallas adds any veteran RB with a real claim (> ~$3M/yr or trade) or any drafted RB — capital added beats camp wins.
- Camp/preseason reports show Jaydon Blue (or Mafah) taking first-team third-down/two-minute reps in consecutive weeks — the satellite-role leak starts there.
- Any shoulder/neck recurrence or missed camp/preseason time for injury (his June absences were coach-confirmed strategic, not medical — that distinction flipping is the tripwire).
- ADP rises inside ~pick 24 (round 2, 12-team) — at that price the market demands the receiving production he doesn't produce; verdict degrades toward HOLD/FADE.
- Consensus win total drops below 7.5 (e.g., Dak injury) — grinder-lean profile on a negative-script team caps the median.
Sources
data/stats/2025/+data/stats/2024/: rushing.csv, receiving.csv, snap_counts.csv, ngs_rushing.csv, participation.csv, pbp_summary.csv — nflverse via nflreadpy, pulled 2026-07-07 (REG only). Verified this run: 2025 252-1201-11 / 51-35-137-2, 54.2% carry share; 2024 139-513-4 / 70-52-346-0; RYOE +0.62 (2025) / −0.60 (2024); 8+ box 22.6%- nflverse play-by-play 2024–25 via nflreadpy, loaded and computed 2026-07-07 — inside-10 28/45 (62.2%) and inside-5 20/30 (66.7%) team shares, 10 of 11 rush TD inside the 10, breakaway 5.2%, success 45.6% (2025) / 2.9% & 35.3% (2024), 17 of 51 targets trailing 7+, DAL 79 RZ drives; third-down (62.0% season / 74.4% wks 1–14) and score-state on-field splits (63.5/71.9/52.5%) via participation.csv join on run/pass plays
data/adp/adp_ffc_2026-07-07.csv— FFC PPR mocks, 2026-07-07: Javonte 34.7 (RB17); neighborhood: Kyren Williams 32.5 (RB16), Skattebo 37.9, Etienne 38.9data/raw/sleeper/players_2026-07-07.json— age 26, DOB 2000-04-25, North Carolina, 5 yrs exp, depth chart RB1, no injury statusdata/team-profiles/DAL.md(built 2026-07-07) — coaching/scheme/OL/win total/vacated math/committee context- NFL.com/PFR career page (retrieved 2026-07-07) — career 858 carries, 193 receptions (≈1,051 touches); 2024–25 components verified against nflverse cache
- PFF via web search (retrieved 2026-07-07) — 2025: 54 MTF as a runner (7th among RBs), 3.6 YAC/att, 896 yards after contact (6th)
- nbcsports.com (ProFootballTalk) / dallascowboys.com / nfl.com / cbssports.com (2/21/2026, retrieved 2026-07-07) — 3-yr/$24M, $16M gtd re-signing; $8M AAV tied ~16th among RBs
- FantasyPros / RotoWire (1/3/2026) — shoulder/neck IR placement ending season; Todd Archer (ESPN, via X, Dec 2025) — "battling a stinger the last few weeks"; Adam Schefter (X, Dec 2025) — Christmas-game shoulder exit; si.com/foxsports.com injury timeline (retrieved 2026-07-07)
- thelandryhat.com (Dec 2025, retrieved 2026-07-07) — 855 yds/4.9 YPC through 11 games, <4.0 YPC in 3 of 4 wks 13–16, called for a complementary back (none was added)
- heavy.com (6/19/2026) — minicamp hold-out was strategic (RB2 evaluation), Schottenheimer on reducing Williams' dependence; dallascowboys.com minicamp observations + atozsports.com OTA storylines (June 2026) — Blue/Mafah first-team RB2 reps, Schottenheimer "completely different player" on Blue; heavy.com "one-hit wonder" (2026) — market bear narrative
- rotoballer/yahoo/foxsports (retrieved 2026-07-07) — 2026 market sentiment (RB21 rank vs top-10 takes)
- UNVERIFIED / estimates: routes run & true TPRR (est. ~0.14–0.16 from computed 0.113 targets-per-on-field-dropback); provider xFP; PFF pass-block grade (proxy: he held 74% third-down snaps wks 1–14 and 66% dropback participation as a 5th-year vet — the §9 gate is functionally passed on usage evidence)