* Rounded experimental estimates. 100.0% is not an official elimination flag. Negative points are modeled through projected fumbles / interceptions, not a broad negative-score tail. “Live” means the NFL game is in progress; on-field snap status is unavailable.
Rank
Team
Points
Proj*
Perf(σ)
P5*
P95*
Left
Remain*
FAAB
Projected chop 92.44
1
x
23.1%
0.00
115.10
—
79.3
154.5
8
115.10
$733
2
x
14.6%
0.00
123.41
—
84.1
167.4
8
123.41
$587
3
x
13.4%
0.00
123.78
—
86.2
165.6
8
123.78
$331
4
x
9.8%
0.00
130.68
—
89.3
176.4
8
130.68
$81
5
x
9.4%
0.00
131.35
—
90.0
176.7
8
131.35
$299
6
x
8.9%
0.00
131.80
—
90.5
177.4
8
131.80
$4
7
x
7.9%
0.00
134.21
—
92.1
181.0
8
134.21
$200
8
x
6.6%
0.00
135.28
—
93.9
180.7
8
135.28
$221
9
x
6.4%
0.00
138.75
—
94.1
188.2
8
138.75
$262
Current low (scored) 0.00Scored Remaining*
Pregame 92.44
Week 10 · Projected chop 92.44 pts. The median of the lowest final team score in each simulation, using the same live projections, SUBs and Stacks setting as chop odds.
Pregame 92.44 pts · middle 80% 74.72–109.00. Last recorded before the week’s first kickoff: . Model participation-weibull-copula-v9-empty-slots, Stacks ON.
Above this line does not guarantee safety. Use each team’s chop odds to assess its risk.
rsk3200.00 scored115.10 final*
CHOP RISK: HIGH → LOW · SELECT BAR / LOGO
CURRENT LOW SCORE0.00rsk320
GAP TO NEXT TEAM0.00Les Misthrôwables is next above
STARTERS TO PLAY729 teams with players remaining
TEAMS IN FIELD9 teamsLowest weekly score is at risk
* Our model: full-game baseline = Σ(projected stat × league scoring weight). Remaining = full baseline before kickoff, zero after completion, or baseline × fraction of regulation time left during a live game. Projected finish = official current score + remaining projection. Overtime and unresolved end-of-regulation clocks use a flagged allowance of up to ten minutes at pregame rate; missing live clocks use a flagged full-game allowance. Missing individual forecasts assume zero points and zero SD. Injury tags are display-only and do not override the selected-week forecast. Projections preserve the baseline mean; possession and usage changes are not modeled.
Experimental survival estimates: 20,000 simulations; one lowest team is chopped. For projected positive scoring A, the full-game net spread uses the player’s NFL position: QB √(6² + (0.21 × A)²); RB max(5, 0.6 × A); WR max(5.5, 0.6 × A); TE max(4.5, 0.7 × A); points. DNP, early-exit and normal-participation scenarios share the same fitted mean and variance. Scenario variance is deducted before drawing normal scoring; injury risk is not added twice. Penalty counts remain Poisson, and positive scoring follows Weibull conditional on time played. Turnover deductions can still produce negative scores. Penalty and scenario means impose a minimum variance when the fitted variance is too small. Completed games add zero remaining points. Out/IR/PUP/SUS badges do not change the simulation.
Historical participation scenarios*
Fit: 2022–2023; checked on 2024. DNP means zero points; early exits stop scoring at a sampled historical exit time; normal participation follows the scoring distribution.
Pos
DNP*
Early exit*
Sample
2024 in 5/95
QB
0.05%
2.82%
1,082
87.3%
RB
0.34%
1.40%
2,178
88.8%
WR
0.27%
1.91%
3,544
82.8%
TE
0.32%
1.25%
1,402
87.4%
* Retrospective positive-forecast sample: most known inactives are excluded, so these are not general pre-kickoff injury probabilities. Historical selection requires ≥8 projected points for QB and ≥3 for RB/WR/TE; this is not an NFL starter-only sample. Rates use half-count smoothing; early-exit rates are conditional on playing. An exit requires a recorded injury and no subsequent participation in a game with at least 95% play coverage. Unrecorded injuries may be missed. Live draws condition on the player still participating, remove DNP risk and past exit times, and keep earned points fixed. Pregame injury tags do not override Sleeper projections. A confirmed new OUT during a game stops remaining scoring and uncertainty; earned points and the original forecast remain intact. Missing status fields cannot clear a confirmed OUT. Otherwise remaining scoring uses the forecast scaled by game time.
2024 coverage is the fraction of actual scores inside the nominal 90% range, among records with complete scoring-event inputs. Coverage falls short, especially for WRs; these tails are experimental and not validated 90% guarantees.
The published forecast is treated as the overall mean. Conditional normal-game production is adjusted to preserve that mean and the fitted variance. This changes tail shape without a second injury haircut; full-game SD remains fixed for performance comparisons. Source archives are not frozen kickoff forecasts.
Lineup Proj shows the pregame baseline before kickoff and after final; live rows show projected total. The baseline is Sleeper’s available weekly forecast, not an archived kickoff snapshot. Row shading compares actual points (final) or projected total (live) with that baseline, divided by the full-game model SD, including the minimum variance required by penalty and participation scenarios. Tints increase at 0.5σ, 1σ and 2σ; unstarted players and missing or zero-SD inputs stay neutral. These are retrospective position error scales, not individually calibrated player deviations. Missing player projections display 0.00; an asterisk identifies missing forecasts assumed to be zero points and zero SD. A supplied forecast supersedes the assumption.
Position scales were fitted to 2022–2024 Sleeper projected/actual scoring errors in 2022–2023 regular seasons, checked on 2024 regular season. Historical forecast timestamps are postgame, so this is not verified pregame calibration. The sample requires positive archived forecasts and verified participation; general pre-inactive-list DNP risk and projections below the training range remain unvalidated. NFL position is separate from FX/SFX. Draws are shared for the same player across rosters. Experimental same-game correlation links normal scoring when Stacks is enabled; negative yardage, injury-specific severity and usage changes are not modeled. Missing required scoring-event inputs suppress all probabilities. Percentages round to one decimal: 100.0% is a model estimate, not mathematical certainty or an official elimination. Current scores are fixed; stat corrections are excluded. Simulated final scores round to 0.01 points before ranking. Ties split simulated elimination equally, unlike Sleeper’s season-points/draft-position rule. The field assumes one weekly elimination and uses selected-week matchup IDs, excluding confirmed earlier-week losers.
Team Perf(σ) = (projected final − original starter projection) / full-game simulated team SD. Both SD and RmSD use the selected Stacks setting, ON by default. Before any starters play, they match. Finished teams use their actual score. Bench players are excluded. Full-game SD stays fixed as games finish unless lineup or forecast inputs change. Final-score ranges instead use remaining simulation uncertainty: 1/99, 5/95 and 10/90 span the middle 98%, 90% and 80% of simulated scores. These are experimental prediction intervals, not confidence intervals for a mean. Completed rosters collapse to their actual score; stat corrections are excluded.
Tied scores share ranks. CHOPPED marks the unique lowest score only after every NFL game that week is final; tied or incomplete results remain unconfirmed. Final earlier-week losers are excluded from later previews. Stat corrections can update the result. Remaining players include live and unstarted starters. Points display two decimals, rounded to nearest; full projection precision is retained when summing.
Discord tools
Sleeper omitted projected fumbles lost for Corey Kiner; using the supplied projection without adding an unsupplied penalty forecast.
ⓘ Metrics
REC: receptions / targets (3/6 = 3 catches on 6 targets). A dash means that count is unavailable. F: final.
Points: live Sleeper stats scored with this league’s rules, reconciled with official matchup points. Finished games are saved; corrections refresh hourly for 24 hours, then freeze. Commissioner score overrides remain authoritative.
Proj: pregame forecast for finished players; scored + estimated remaining for live players. 0* means a missing forecast assumed 0 points and 0 SD, not a supplied zero.
[SUB]: projected totals and chop odds assume an eligible bench replacement for a zero-projection starter before kickoff. Existing starters can rotate between eligible slots to make room; [MOVE] shows these slot changes. Every healthy starter is retained, and each bench player is used once. Actual lineups stay unchanged. Players who have kicked off cannot move or be replaced in the estimate.
SD / Perf(σ): full-game team standard deviation and performance against the original projection. Full-game team simulations use the selected Stacks setting, ON by default, just like RmSD and chop odds. SD stays fixed as games progress unless lineup or forecast inputs change. +1σ is one SD above baseline.
RmSD: uncertainty in points still to score. Final-score ranges simulate only remaining production plus points already earned; finished players add no remaining uncertainty. SD and RmSD match before any starters play.
Final %ile: 5/95 spans the model’s 5th to 95th percentiles (middle 90% of outcomes); 1/99 and 10/90 span 98% and 80%. These are experimental prediction intervals, not guaranteed bounds.
Completed standings: OPROJ sums that week’s original starter forecasts; OP5/OP95 (or the selected bounds) show the full-game range. Actual %ile places the final score in that original modeled distribution: low means an unusually weak result, high means unusually strong. These use available Sleeper weekly forecasts, which are not guaranteed frozen kickoff forecasts. Bench players are excluded.
Rank / Chop: ranks compare selected-week team scores; tied current scores use projected scores for ordering. Chop probability is the chance of the lowest score among all eligible teams, with ties split equally. Unknown projections are excluded from displayed ranks/margins; probabilities require the full field.
Projected chop: the dotted line is the median of the lowest team score in each joint simulation. Hover or tap its label for the middle 80% range. It updates with live scores, remaining projections, SUBs and the Stacks setting. The pregame comparison preserves the last recorded forecast before the week’s first kickoff. A score above the line does not guarantee survival.
Clock assumptions: overtime uses up to ten minutes at pregame scoring rate. Missing live clocks use a flagged full-game remaining allowance until the feed recovers.
(30)23.20 pts*
(13)15.15 pts*
(20)11.06 pts*
(2)15.29 pts*
(3)9.37 pts*
(7)12.50 pts*
(25)14.29 pts*
(14)14.25 pts*
