NFL Analytics Dashboard2026 · Week 4
Oct 1, 7:05 PM EDT
Slate

Zach Ertz

TE
39% to play
Game
2026 wk 4
Total
43.5
Implied team total
20.3
Depth
#2
Recent snaps
66%

35 yrs · 6′5″, 250 lb · Stanford · 2013 rd 2 pk 35

Projections

MarketProjectionLikely rangeFair oddsEvidence
Receiving
Receiving yards29.50.0 – 62.5–Solid
Receptions3.41.0 – 5.4–Solid
Longest reception15.74.6 – 29.9–Solid
Rush + rec yards29.94.7 – 62.3–Solid
Touchdowns
Anytime TD17.5%no range+472Solid
First TD scorer2.9%no range+3293Solid

Form

WindowGPTgtRecRec ydRec TDTgt shareWOPRRZ tgtSnap%PPR
Last 335.74.042.00.0016%0.331.0058%8.2
Last 556.04.044.40.0019%0.441.0066%8.4
Last 10105.43.736.40.2019%0.421.0064%8.5
2026 season12.01.09.00.008%0.141.0038%1.9

Per-game averages over completed games, newest first. Windows match the rolling features the models are trained on.

Career by season · totals

WindowGPTgtRecRec ydRec TDTgt shareWOPRRZ tgtSnap%PPR
2026121908%0138%2
2025137250504419%6966%126
2024179166654718%72367%177
202374327187119%3865%52
2022106947406418%41779%116
20211711274763519%72174%181
2020117236335119%4881%77
20191513588916624%82085%216
2018161561161,163826%92792%280
20171411074824823%71780%202
20161410678816418%61786%184
20151511275853218%7971%170
2014168958702314%51350%144
2013155636469411%4640%107

Per-game averages over completed games, newest first. Windows match the rolling features the models are trained on.

Injury history

20 injury entries across 12 seasons · 8 weeks ruled out or doubtful · most often concussion

2025Shoulder
2024Concussion
2023Knee
2022Knee×2Calf
2021HamstringNot Injury Related
2020Ankle
2019Ribs×2Hamstring
2017ConcussionHamstring
2016Rib×2Hamstring
2015Concussion×2GroinHip
2014KneeOblique
2013Shoulder

Model track record

out-of-sample, 2019 onwards
MarketGradedIn bandAvg missvs form
Receptions9583%2.0+0.1
Receiving yards9580%21.2+0.4
Rush + rec yards9583%20.9+0.8
Anytime TD9528% vs 27%0.4-0.0
First TD scorer956% vs 5%0.1-0.0
Longest reception9486%7.0+0.1

Out-of-sample: each prediction came from a model trained only on earlier seasons. In band should sit near 80% where a band exists — that is the calibration claim, checkable here. It is tinted only when it strays further than sampling noise explains, which for one player is roughly ±7 points. vs form is how much closer the model landed than guessing his recent average. This grades the model against RESULTS, not against a betting line.

Last 10 games

GameOppSnap%TgtRecRecYdCarRushYdPassYdTD
2026 wk 3CHI38%2190000
2025 wk 14MIN69%21110000
2025 wk 13DEN67%13101060000
2025 wk 11MIA75%84420000
2025 wk 10DET81%54540000
2025 wk 9SEA63%44460000
2025 wk 8KC62%64160000
2025 wk 7DAL67%73370001
2025 wk 6CHI60%66430001
2025 wk 5LAC53%1000000