Connect your Sleeper league, bring your rankings, and plan the week. Draft tools stay in their own dedicated tab.
Find your Sleeper leagues by username. Selecting a league loads it automatically; choose your roster below if needed.
Once loaded, open Waivers to compare available players with your team. Draft configuration remains inside Draft.
Weekly rankings inform immediate lineup and waiver needs. Rest-of-season rankings inform longer-term value and redraft trades.
Works with your Sleeper NFL leagues. Compare the wire with your roster, FAAB, Late Round weekly/ROS ranks, projected lineup gains, and byes. Select Chopped mode only for a guillotine league.
Type your Sleeper username and find leagues, then select one to load it automatically. League and roster data are read from Sleeper; your imported rankings stay in this browser.
Open Weekly Rankings, ROS Rankings, and the Data Dump on Late Round, download their files, then upload them here. You can select multiple files per source to combine positional tables. GOING reads every workbook sheet, matches player names to Sleeper IDs, and keeps original columns as evidence. Files are stored in this browser only, not sent to the GOING data build or GitHub.
These expert rules apply only in Chopped mode. Dollar suggestions remain GOING heuristics, not published prices or competitor-bid forecasts.
Loading odds and historical projections…
Opportunity, role, scoring and model-quality charts that update with every weekly refresh. Each chart says whether it is predictive or only descriptive.
Loading charts…
One fast view of who stands out in each NFL market, built from GOING's projection, recent role, supported matchup evidence and game environment.
Any TD is an absolute evidence score led by a conservative touchdown estimate. GOING shrinks a player's short historical window toward a position baseline, then uses the cross-book market only as a small guardrail when at least three books reasonably agree. A disputed market is disclosed and excluded from the score. Scoring-area role, supported matchup evidence and team scoring environment can move the result only modestly. A player is not pushed toward 100 merely for ranking first.
Pass yards, pass TDs, rush yards and receptions remain slate-relative because their inputs are market-specific continuous measures: air/red-zone volume, carry or target role, opponent weakness and expected game script. Missing evidence is omitted and never filled with a zero. Evidence depth reflects sample size and input coverage, not validated accuracy. “Due” narratives and individual outcomes do not create a positive score bonus.
The eligible field requires either a recent unexpired sportsbook quote or recorded current-season participation, which keeps retired, inactive and stale roster profiles off the board without pretending a quote proves availability.
Freeze a score before kickoff to preserve what GOING knew then. Only the scores you choose are captured; this is not a representative product-wide backtest. Only the five existing NFL player Score markets are tracked; NFL/NCAA game markets have no defined GOING Score yet. No win-rate or profitability claim is made.
Choose players, inspect their projections, then build around your locks. Import your official salary slate before entry.
Projection-only mode needs no CSV. It creates ranked players and eight-player offensive roster ideas, but cannot verify salaries, contest eligibility or D/ST. Upload the official salary file before treating a lineup as contest-valid.
Build around how your Captain can succeed. These adjustable preferences follow ETR’s Showdown 101; they are not DraftKings requirements or proven winning rules. Player projections remain unchanged.
Leave salary unused when you choose; there is no universal optimal amount. Low ownership requires the private Showdown JSON, not the Classic JSON. Both teams stay represented. QB + kicker and Captain + opposing defense are disclosed conflicts, not automatic bans. Rules can make a slate or your locks infeasible; GOING will explain rather than relax them. Current-season score ranges still assume independence: these construction rules do not create a calibrated joint simulation.
Building the current slate from GOING projections.
Ranked by current-season observed scoring-spread proxy where supported, otherwise projection. Multiplier output is shown separately from the underlying projection; ownership leverage is not available.
DFS uses completed current-season games when at least two supported player games exist. For DraftKings Showdown only, a verified starting QB with insufficient current-season starting samples can use two or more verified starts from the immediately preceding season. This fallback is labeled separately and never mixed into a supported current-season projection. Classic and other positions retain current-season-only projections. Negative yardage is included in scoring. These small-sample means are unvalidated forecasting baselines.
DraftKings scoring uses full PPR and a $50,000 cap; FanDuel uses half PPR and a $60,000 cap. Both include supported yardage bonuses. Single-game mode uses six distinct players from one game; Captain/MVP earns 1.5 times points, and DraftKings Captain costs 1.5 times salary. Both teams are represented. Current-season observed scoring spread can change the research ranking; it is not a predicted ceiling.
Player scoring omits interceptions, lost fumbles, return touchdowns and two-point conversions. DraftKings D/ST uses partial current-season public team scoring, excluding ambiguous fumble-return TDs. DraftKings Showdown kickers use current-season field goals by distance and extra points; missing or stale kicking evidence is withheld rather than filled with imported platform averages. No historical cross-player correlation coefficients are applied; independence can miss real game dependence. Ownership estimates, when available, are separate from scoring projections. Suggested lineups are research drafts, never automatic entries.
Explore specialized touchdown questions. First TD researches the opening scorer; Slate Breaker researches the last touchdown; King of the Endzone projects longest-touchdown candidates. These are model scenarios, not guaranteed outcomes. Check the market rules and available prices.
Ohio · exactly 3 anytime-touchdown scorers · $5 minimum · +2000 minimum per slip. These settings were supplied by you; the full offer terms and entry limit are not verified.
Five alternatives, ranked by estimated chance of all three scoring. Different games only in this first view. The estimate assumes those games behave independently; it is not a proven hit rate. The shared prize has no estimated dollar value here.
Entry count, excluded games/players, opt-in requirements, and submission deadline still need checking in your offer. Options are not labeled promo-qualified. Same-game tickets need Fanatics’s actual combined price and a separately supported joint estimate.
Options update automatically from fresh Fanatics prices and Going Long touchdown estimates. Players must appear on the current active roster without an injury designation. This is an early research screen, not official game-day clearance; confirm official inactive reports separately.
Checking Fanatics prices and player availability automatically…
Players ranked within each upcoming game by the existing first-touchdown model. It weighs early possessions and scoring-area opportunities across both teams, including other scorers and no touchdown. Prior-season data can support the estimate; current roles and availability may differ. These estimates are not validated results.
Choose your slate and games, then build and edit a research draft at one sportsbook. GOING searches for high estimated win chance, not guaranteed winners or the highest payout. Prices and model estimates need review.
Only checked games in the active date/slate can supply legs. Different-game drafts use one leg per game; same-game research uses one checked matchup.
Settings refresh suggestions but do not replace an edited draft. Use Build / replace draft when ready. Price limits apply to different-game tickets only. A profit boost increases qualifying winnings, not win chance; actual bonus caps and book rules still apply. Search is bounded, not an exhaustive optimum.
Select a game. See the football case, the risk and the price.
Probability caution: in audits the model's prop probabilities were overconfident (real hit rates tracked the book price, not high model percentages). Receptions are recalibrated, passing yards use a refit model and touchdown averages are corrected; receiving and rushing yards are still raw, and game lines equal the posted price. Even so, live tests found the model adds nothing detectable beyond the book price, so read any value estimate as a price comparison, not an edge, and treat any model probability above about 80% as unreliable.
Loading matched football evidence…
0–100 is a transparent football-case rating, not win probability or betting value. Directional ratings combine model direction, current production, usage and verified context. TD role research and market-anchored game leans use descriptive labels rather than a misleading numeric floor. Spread and total leans compare current-season scoring context with the posted line; they are not validated forecasting advantages. Unknown is neutral. Evidence strength sets the rating band; projection distance relative to model spread separates bets within that band. Open a card for the recipe and sources. Related markets can be correlated; no profitable sportsbook edge is claimed.
Full research methodCheck the evidence first, then browse by game. Estimated chance and betting value are different.
NFL / NCAA · evidence first
Football evidence readiness. Likelihood and price are shown separately.
Loading the current slate…
Current support means the football evidence agrees. Developing needs more evidence. Check first names a blocker. Price checks remain separate; no label promises a winner.
Read the Confidence methodCandidates clear the existing checks. Watchlists still need a better price or stronger evidence.
Fresh Pinnacle-led comparisons remove the books’ built-in charge and reduce the estimate for uncertainty. These are research candidates, not proven edges.
Reasons to investigate, not proven advantages. More categories do not mean a measured increase in win chance.
Estimated return means how far ahead or behind the model expects $1 bets to finish over many repeats. It can be wrong. Book prices and historical averages are different kinds of evidence.
Checking current prices and available evidence…
Price comparison and historical-model disagreement are available with fresh matched quotes. Role, pace and pressure need observations within 45 days plus enough measured plays. Older observations stay visible in model details but do not establish a current role.
True routes, current injuries, receiver alignment, drops/luck, career milestones and weekly advanced-stat trends are not supported here yet. Opportunity research includes dated position-based defensive comparisons. Pass-snap charting is credited to FTN Data via nflverse; it is not route tracking. Missing charting never counts as zero.
Counts describe available looks, not how much to bet. Multiple looks can depend on the same game outcome.
Full-game selections at the same book, from different games. First TD, period baselines and flagged model checks are excluded. The combined chance assumes the legs do not affect each other; shared model errors can still affect both. Payouts below multiply single prices and are illustrations, not bookmaker quotes. Check the actual ticket. Same-game and pick’em builds need separate payout and joint-outcome models, so they are not priced here.
Looking for a particular bet type? Compare available markets here. For research ordered by estimated chance, start with Today; for one matchup, use Games. Availability alone is not a recommendation.
Research estimates use the daily roster/injury snapshot: confirmed outs are removed, uncertain players are reduced, and only a verified next player on the depth chart receives a capped role boost. Final inactive news can still change after the snapshot. No stake is suggested until the model has been validated.
One per line, tab or comma separated:
player, team, opp, market, line, overOdds, underOdds, projMean, projSd, kickoffISO
Leave mean and SD blank to match a historical profile. Markets: pass_yds, rush_yds, rec_yds, receptions, pass_tds, rush_tds, rec_tds, atd.
For ATD rows: player, team, opp, atd, , atdOdds, , eTD, , kickoffISO — leave line/underOdds/projSd blank, put λ (expected TDs) in the projMean slot.
Compare spreads, totals and moneylines for a game. Open a matchup for book-by-book prices; Today groups model-screened research angles by game.
One per line: away, home, spread(home), total, mlAway, mlHome, kickoffISO
Spread is the home team's number (negative = home favored). Team projections match the exact teams and kickoff from the historical pipeline.