Every player is already on the board. Pay when you want to know what to do about him.
Every projected player and every game on the slate is searchable from the first minute of your free month - what you buy is the arithmetic in the columns beside them. A new account gets full Active access free for 30 days, no card; choose a plan before it ends or access locks, never charges.
Two packages stay complete on every plan - win probability already ships free inside nflverse, and nobody sells an accuracy audit at any price. Charging for either would have been a guess.
Free month
NO CARD · NEVER AUTO-CHARGES
Everything Active opens, for 30 days, with nothing withheld to make the free month feel smaller than the plan.
- Every number we model, every row, every column.
- Modelled P(active) for every player on the live board, with confidence bands and source-conflict flags. The raw designations are free elsewhere; the calibrated probability, with its out-of-sample record published, is not.
- The rating and points impact of every dated, sourced injury situation.
- Read on the site. No export and no API - that is the next plan, and we would rather name the line than blur it.
No card, and nothing is ever charged automatically. Full Active access opens now, free, for 30 days. Choose a plan before it ends and access carries straight through; do nothing and access locks - your account and data stay exactly as they are, and purchasing CA$29/mo turns it back on the moment you're ready.
Accounts that existed before the trial replaced it keep Scout permanently - free, no card, no trial clock - because that is exactly how it was sold.
Active
One person making real roster decisions this week - drafters in August, lineup-setters in October.
- Every number we model, every row, every column.
- Modelled P(active) for every player on the live board, with confidence bands and source-conflict flags. The raw designations are free elsewhere; the calibrated probability, with its out-of-sample record published, is not.
- The rating and points impact of every dated, sourced injury situation.
- Read on the site. No export and no API - that is the next plan, and we would rather name the line than blur it.
- Seats
- 1
- Data leaves the site
- no - read on site
- API / day
- -
- Support
- Email, 1 business day
Feed
Anyone who needs the data inside their own spreadsheet, model, notebook or product rather than inside our web pages.
- The identical data, leaving our site: CSV and JSON, gated by the same engine as every page. XLSX and Parquet are not served yet - the reason is published rather than glossed.
- A metered API at 5,000 requests a day.
- Every release pinned - immutable version, SHA-256 of each upstream source, row counts, precision notes. Clone, rebuild, and the datasets come back byte-identical.
- Three seats. Internal use.
- Seats
- 3
- Data leaves the site
- csv, json
- API / day
- 5,000
- Support
- Email + Slack Connect, same business day
Your player is already on screen with an empty cell beside his name. Active is the only way to fill it - and it fills it on every row of every package at once.
Does the data leave the site? Active is read-on-site. Feed is files, an API, and version-pinned builds you can reproduce.
Need access without a subscription? A pass is one payment for a fixed window - no card kept on file, nothing renewing, nothing to remember to cancel. It ends when the window ends. Every pass grants the same Active tier, across every sport this catalog sells - the sport toggle picks which pass's copy and dates you're looking at, not a different level of access.
2026 Draft Kit
Not the cheapest way to buy 9 days - a month of Active is CA$29. It is the simplest: one payment, no card kept on file, nothing to remember to cancel before the season starts.
Full Active access to every NFL package - eleven of the twelve Active-tier packages, including The Weekly Board - from purchase through the final Week 1 kickoff. NHL Game Forecast is not included; it comes with the NHL Season Pass or any Active or Feed subscription. No renewal, no card on file.
Access ends Tue, 15 Sep 2026 00:15:00 GMT · non-refundable
2026 In-Season Pass
The same 162 days on Active monthly bills 6 times - CA$174. This is CA$37 less, with no card kept on file and nothing to cancel.
Full Active access to every NFL package from purchase through Super Bowl LXI on 14 February 2027 - the regular season and the whole postseason. NHL Game Forecast is not included; it comes with the NHL Season Pass or any Active or Feed subscription.
Access ends Mon, 15 Feb 2027 05:00:00 GMT · non-refundable
2026-27 NHL Season Pass
The same 217 days on Active monthly bills 8 times - CA$232. This is CA$152.01 less, with no card kept on file and nothing to cancel.
Full Active access from purchase through the end of the 2026-27 NHL regular season (Saturday, April 10, 2027) - the same grant the 2026 In-Season Pass sells for NFL. Regular season only, not the playoffs: the Stanley Cup Final has no fixed date announced this far ahead, unlike the Super Bowl.
Access ends Sun, 11 Apr 2027 04:00:00 GMT · non-refundable
Exactly what each tier opens
Generated from the same grant object the API enforces, so this page cannot advertise a paywall the engine does not apply. Pick a tier to see the page as that tier sees it.
- ◐Positional Stat Lines
- ◐Injury & Availability Intel
- ◐Market Divergence Board
- ○Licensed Data Access
- ◐Touchdown Equity
- ◐Draft Value Engine
- ◐Draft Availability Simulator
- ●Game ForecastHow should this game go, and how wide is the range around that?
- ●Verification LedgerIs this edge real, or is the model just reading a stale input?
- ◐Leg Correlation Matrix
- ◐Outcome Board
- ◐NHL Game Forecast
- ○The Weekly Board
- ●Positional Stat LinesWhat volume and production should I expect from this player?
- ●Injury & Availability IntelWill this player be on the field, and what is the injury worth in points?
- ●Market Divergence BoardWhere does my model disagree with the market, how much of that gap is my own error, and does the shipped projection already know what it is facing this week?
- ○Licensed Data AccessCU2 U 04Y%E $G NWGWY 2E% NYG E $8CGG2S%8 E%%% E NS62WS
- ●Touchdown EquityWho actually scores, and how much of a player's value is touchdown luck?
- ●Draft Value EngineWhat is this player worth relative to the replacement I could have instead?
- ●Draft Availability SimulatorCan I wait, or does this tier break before my next pick?
- ●Game ForecastHow should this game go, and how wide is the range around that?
- ●Verification LedgerIs this edge real, or is the model just reading a stale input?
- ●Leg Correlation MatrixIf I take both of these legs, how far from independent are they - and if I stack these two players in a lineup, how correlated are their actual fantasy points?
- ●Outcome BoardIf I add these specific players' projected points together, what range should I actually expect for the total - and how much does that range change once the measured correlation between them replaces the independence assumption every per-player floor/ceiling on this site (and everywhere else) implicitly makes when you combine two of them by hand?
- ●NHL Game ForecastHow should this game go, and where does our rating disagree with the market?
- ●The Weekly BoardIf the models had to commit for the weekend, what would they say - and what is their record for saying it?
- ●Positional Stat LinesWhat volume and production should I expect from this player?
- ●Injury & Availability IntelWill this player be on the field, and what is the injury worth in points?
- ●Market Divergence BoardWhere does my model disagree with the market, how much of that gap is my own error, and does the shipped projection already know what it is facing this week?
- ●Licensed Data AccessCan I build on this, and can I republish what I build?
- ●Touchdown EquityWho actually scores, and how much of a player's value is touchdown luck?
- ●Draft Value EngineWhat is this player worth relative to the replacement I could have instead?
- ●Draft Availability SimulatorCan I wait, or does this tier break before my next pick?
- ●Game ForecastHow should this game go, and how wide is the range around that?
- ●Verification LedgerIs this edge real, or is the model just reading a stale input?
- ●Leg Correlation MatrixIf I take both of these legs, how far from independent are they - and if I stack these two players in a lineup, how correlated are their actual fantasy points?
- ●Outcome BoardIf I add these specific players' projected points together, what range should I actually expect for the total - and how much does that range change once the measured correlation between them replaces the independence assumption every per-player floor/ceiling on this site (and everywhere else) implicitly makes when you combine two of them by hand?
- ●NHL Game ForecastHow should this game go, and where does our rating disagree with the market?
- ●The Weekly BoardIf the models had to commit for the weekend, what would they say - and what is their record for saying it?
Deliberately not withheld at any tier: floor and ceiling, because the product publishes a rule that no forecast ships without its uncertainty and a pricing page that broke that rule would be the exact failure the Verification Ledger exists to catch. Nor the ADP delta, which is board rank minus market rank - both already visible, so gating it would be theatre.
Twelve packages, one player search away
What each package answers. Expand a package for the exact fields inside it.
Positional Stat Lines
Full-season and weekly NFL player projections - points, floor, ceiling, and every underlying stat from carries to targets to receiving yards, broken out by position. Built from each stat's own year-over-year stability rather than one blanket regression rate, so volume and efficiency are never smoothed into the same number. Every weekly projection ships with a real range around it, not just a single guess.
What's included (40 fields)+
| Projected points | Full-season fantasy points, half-PPR baseline |
| PPG | Projected points per game played |
| Floor | Low-end season outcome after games-played risk |
| Ceiling | High-end season outcome |
| Attempts | Projected pass attempts for the week |
| Pass yards | Projected passing yards |
| Carries | Projected rush attempts |
| Rush yards | Projected rushing yards |
| Targets | Projected targets |
| Receptions | Projected catches |
| Receiving yards | Projected receiving yards |
| Weekly floor | 10th percentile of the modelled weekly half-PPR outcome |
| Weekly median | 50th percentile of the modelled weekly outcome |
| Weekly ceiling | 90th percentile of the modelled weekly outcome |
| Weekly SD | Standard deviation of the modelled weekly outcome |
| TD share | Share of the projection that is touchdown expectation rather than yardage - a high value means the ceiling is real and the floor is not |
| Volume | The opportunity half of a yardage projection - attempts, carries or targets, depending on the stat |
| Yards per touch | The efficiency half - the projected yards divided by the projected volume, so the two halves always multiply back to the published mean |
| Gate verdict | Whether this stat's fitted weekly distribution passed its pre-registered walk-forward gates - where it failed, the quantile columns are null on purpose |
| Pass TD | Projected passing touchdowns |
| Rush TD | Projected rushing touchdowns |
| Rec TD | Projected receiving touchdowns |
| Total TD | Projected total touchdowns across rushing, receiving and passing |
| TD dependency | Share of a player's projected points that comes from touchdowns |
| Volume hand-corrected | True when this week's volume projection was manually adjusted for a role change nflverse trailing data would not yet reflect |
| Yardage mean | Modelled mean yards for this stat, from the volume x efficiency decomposition - restates the same paywalled number the raw stat above hides |
| Yardage p5 | 5th percentile of the modelled weekly yardage distribution for this stat - null until this stat pair passes its walk-forward gates (see limits) |
| Yardage p25 | 25th percentile of the modelled weekly yardage distribution for this stat - null until this stat pair passes its walk-forward gates |
| Yardage p50 | 50th percentile (median) of the modelled weekly yardage distribution for this stat - null until this stat pair passes its walk-forward gates |
| Yardage p75 | 75th percentile of the modelled weekly yardage distribution for this stat - null until this stat pair passes its walk-forward gates |
| Yardage p95 | 95th percentile of the modelled weekly yardage distribution for this stat - null until this stat pair passes its walk-forward gates |
| Anytime TD % | P(at least one TD) = 1 - e^-lambda - Touchdown Equity's own field, riding along on the shared player_props row |
| Anytime TD % (reported) | The published, rounded anytime-TD probability shown alongside the raw computed value - Touchdown Equity's own field |
| Auction $ | Dollar value on a $200 budget - Draft Value Engine's own field, riding along on the shared season_projections row |
| Expected rec. TD | Receiving share of the player's expected-touchdown rate (lambda_td) - Touchdown Equity's own field |
| Expected rush TD | Rushing share of the player's expected-touchdown rate (lambda_td) - Touchdown Equity's own field |
| Flag | Draft Value Engine's value / reach tag, riding along on the shared season_projections row |
| Expected TDs | Poisson rate: expected touchdowns this game - Touchdown Equity's own field, riding along on the shared player_props row |
| 2+ TD % | P(two or more TD) = 1 - e^-lambda(1+lambda) - Touchdown Equity's own field, riding along on the shared player_props row |
| VBD | Points above the positional replacement baseline - Draft Value Engine's own field, riding along on the shared season_projections row |
Injury & Availability Intel
A calibrated probability that your player actually takes the field this week - not just a raw practice-report designation copied from a beat reporter. Every number carries a published, out-of-sample accuracy record, and rows flag a source conflict rather than silently picking a side when two feeds disagree.
What's included (25 fields)+
| P(active) | Modelled probability the player takes a snap this week |
| Confidence | high / medium / low / low_out_of_domain, from corroboration and data coverage |
| Source conflict | True when two independent feeds disagree on the designation - published, not resolved |
| Sources | Which feeds contributed this row |
| Situation | The dated, sourced fact - never a rumour |
| Rating impact | Rating points debited from the team |
| Points impact | The same adjustment expressed in game points |
| Source | Publication and date the fact came from |
| Availability factor | Multiplier applied to the player's stat line and TD rate |
| Attempts | Projected pass attempts for the week - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared player_props row |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared player_props row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared player_props row |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared player_props row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Volume hand-corrected | True when this week's volume projection was manually adjusted for a role change nflverse trailing data would not yet reflect - Positional Stat Lines' own field |
| TD share | Share of the projection that is touchdown expectation rather than yardage - Positional Stat Lines' own field, riding along on the shared player_props row |
| Anytime TD % | P(at least one TD) = 1 - e^-lambda - Touchdown Equity's own field, riding along on the shared player_props row |
| Anytime TD % (reported) | The published, rounded anytime-TD probability - Touchdown Equity's own field |
| 2+ TD % | P(two or more TD) = 1 - e^-lambda(1+lambda) - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected TDs | Poisson rate: expected touchdowns this game - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected rush TD | Rushing share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected rec. TD | Receiving share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
Market Divergence Board
See exactly where our model and the betting market disagree on a game, ranked by the size of the gap, with a plain-language read on which side is more likely right. Includes a matchup-strength layer, so you know whether a player's projection already accounts for a tough or soft defense. No picks, no EV, no stake sizing - the numbers, side by side, nothing else.
What's included (30 fields)+
| Divergence | Elo-only minus market, in percentage points - the ranking key |
| Model implied | Unblended rating-based probability |
| Market implied | Vig-free market probability from the two-way moneyline |
| Blended | The 75/25 blend actually used downstream |
| Market spread | Published spread, used as input and calibration check |
| ADP divergence | Season-long equivalent: board rank vs market rank |
| Matchup index | How much MORE (>1) or LESS (<1) the opponent has allowed to this position than league average, trailing and shrunk to its own sample size |
| Matchup rank | 1 = toughest defense against this position, 32 = easiest, among teams with enough trailing games to rank |
| Environment gap | Matchup-implied points minus the FROZEN shipped projection - never a rewrite of the projection, a second read beside it |
| Distortion flag | True only when the gap is large, the trailing sample is thick enough, AND the position's matchup effect beat a naive baseline out of sample - see limits |
| Matchup validated | Whether this position's matchup index actually predicted future points allowed better than assuming league average, on held-out data |
| DvP league average | League-average value allowed to this position - the baseline dvp_index is measured against |
| DvP sample size | Trailing games behind this position's DvP index - larger samples get less shrinkage toward league average |
| DvP confidence | Confidence in this position's matchup index, from trailing sample size and consistency |
| Matchup-implied points | Points implied by the matchup-only read, independent of the frozen shipped projection |
| Distortion direction | Which side the flagged distortion leans - under or over the frozen shipped projection |
| Auction $ | Dollar value on a $200 budget - Draft Value Engine's own field, riding along on the shared season_projections row |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Flag | Value / reach tag where board and market diverge materially - Draft Value Engine's own field, riding along on the shared season_projections row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rec TD | Projected receiving touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rush TD | Projected rushing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared season_projections row |
| TD dependency | Share of a player's projected points that comes from touchdowns - Touchdown Equity's own field, riding along on the shared season_projections row |
| Total TD | Projected total touchdowns across rushing, receiving and passing - Positional Stat Lines' own field, riding along on the shared season_projections row |
| VBD | Points above the positional replacement baseline - Draft Value Engine's own field, riding along on the shared season_projections row |
Licensed Data Access
Every dataset behind ApexDatum, versioned and hash-verified: pin a release, rebuild it yourself, and get back byte-identical files. Built for anyone who wants the numbers inside their own spreadsheet, model, or product instead of on our web pages.
What's included (6 fields)+
| Dataset version | Immutable version string pinned per release |
| Source hash | SHA-256 of every upstream file that produced the release |
| Generated at | UTC build timestamp |
| Row counts | Per-dataset row counts for integrity checking |
| Precision notes | Which fields carry full precision and which inherit a rounded export |
| Provenance ledger | Per-signal verification status, delivered as machine-readable metadata alongside the data |
Touchdown Equity
Anytime-touchdown and multi-touchdown probability for every player, built from each team's own scoring budget so the numbers never overstate the stars at the expense of the rest of the depth chart. The share of the budget we can't attribute to a named player is published, not quietly folded into the leaders.
What's included (42 fields)+
| Anytime TD % | P(at least one TD) = 1 - e^-lambda |
| 2+ TD % | P(two or more TD) = 1 - e^-lambda(1+lambda) |
| Next-4 positional SOS | Mean opponent DvP index at the player's position over the coming four weeks, joined from the positional_sos dataset; above 1.00 = softer than average. Held ex-ante at the build week |
| Expected TDs | Poisson rate: expected touchdowns this game |
| Team expected TDs | Team offensive TDs from the fitted points-to-TD relationship |
| Modelled share | Fraction of the team TD budget attributable to named players |
| Unmodelled TDs | TD budget the model cannot attribute - reported, never redistributed |
| TD dependency | Share of a player's projected points that comes from touchdowns |
| TD over expected | Prior-season touchdowns above/below what field position predicted |
| Matchup index | How much MORE (>1) or LESS (<1) the opponent has allowed to this position than league average, trailing and shrunk to its own sample size |
| Matchup rank | 1 = toughest defense against this position, 32 = easiest, among teams with enough trailing games to rank |
| DvP sample size | Trailing games behind this position's DvP index - larger samples get less shrinkage toward league average |
| Season positional SOS | Mean opponent DvP index at this position across the full season; above 1.00 = softer than league average |
| Playoff positional SOS | Mean opponent DvP index at this position across the fantasy playoff weeks 15-17 |
| Anytime TD % (reported) | The published, rounded anytime-TD probability shown alongside the raw computed value |
| Expected rush TD | Rushing share of the player's expected-touchdown rate (lambda_td) |
| Expected rec. TD | Receiving share of the player's expected-touchdown rate (lambda_td) |
| Opponent DvP (QB) | This game's opponent defense-vs-position index against QBs - above 1.00 means the opponent allows more than league average |
| Opponent DvP rank (QB) | Opponent's DvP rank against QBs this game, 1 = toughest, 32 = easiest |
| Opponent DvP rank (avg) | Average of the four positional DvP ranks against this game's opponent |
| Opponent DvP (RB) | This game's opponent defense-vs-position index against RBs |
| Opponent DvP rank (RB) | Opponent's DvP rank against RBs this game, 1 = toughest, 32 = easiest |
| Opponent DvP (TE) | This game's opponent defense-vs-position index against TEs |
| Opponent DvP rank (TE) | Opponent's DvP rank against TEs this game, 1 = toughest, 32 = easiest |
| Opponent DvP (WR) | This game's opponent defense-vs-position index against WRs |
| Opponent DvP rank (WR) | Opponent's DvP rank against WRs this game, 1 = toughest, 32 = easiest |
| Auction $ | Dollar value on a $200 budget - Draft Value Engine's own field, riding along on the shared season_projections row |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| Flag | Value / reach tag where board and market diverge materially - Draft Value Engine's own field, riding along on the shared season_projections row |
| Attempts | Projected pass attempts for the week - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| Rec TD | Projected receiving touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| Rush TD | Projected rushing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared player_props/season_projections rows |
| TD share | Share of the projection that is touchdown expectation rather than yardage - Positional Stat Lines' own field, riding along on the shared player_props row |
| Total TD | Projected total touchdowns across rushing, receiving and passing - Positional Stat Lines' own field, riding along on the shared season_projections row |
| VBD | Points above the positional replacement baseline - Draft Value Engine's own field, riding along on the shared season_projections row |
| Volume hand-corrected | True when this week's volume projection was manually adjusted for a role change nflverse trailing data would not yet reflect - Positional Stat Lines' own field, riding along on the shared player_props row |
Draft Value Engine
Draft rankings built from what drafters actually do, not what one analyst thinks they should do: real consensus ADP across nine platforms, a value-above-replacement score, and auction dollars included at every paid tier so you know exactly how big a reach or a steal really is.
What's included (35 fields)+
| VBD | Points above the positional replacement baseline |
| Tier | Projection tier - the unit that actually breaks in a draft |
| Week difficulty | Opponent Elo above league mean for each remaining week, rest-adjusted with a coefficient that earned its place on a held-out season |
| Season SOS | Mean remaining-schedule difficulty; rank 1 = hardest road. Preseason predictive power is weak and stated as such - the validated use is in-season |
| Next-4 SOS | Mean difficulty of the coming four weeks - the window that beat the naive last-season-record baseline in every season backtested (etl/sos_backtest.json) |
| Playoff SOS | Mean difficulty across fantasy playoff weeks 15-17 |
| Positional next-4 | Mean opponent DvP index at this position over the coming four weeks; above 1.00 = softer than league average |
| Reach index | This player's VBD minus the expected best VBD still available at your NEXT pick, at your slot - the cost of waiting, in points |
| Survives to next pick | Probability the player is still on the board at your next pick, from the availability model recovered from this board's own curves |
| Verdict | PICK_NOW / VALUE_WAITS (undervalued and likely to survive) / TRAP (a discount you will not get to use) / WAIT / TOSS_UP |
| Auction $ | Dollar value on a $200 budget |
| Market ADP | Consensus ADP across nine platforms |
| ADP delta | Board rank minus market rank; positive means the market is late |
| Flag | Value / reach tag where board and market diverge materially |
| Board rank | Rank on this board, all positions |
| Opponent Elo | Opponent's Elo rating for this week's matchup, the raw input behind the difficulty score |
| Season SOS rank | Rank of this player's season-long strength of schedule, 1 = hardest road |
| Matchup index | How much MORE (>1) or LESS (<1) the opponent has allowed to this position than league average, trailing and shrunk to its own sample size |
| Matchup rank | 1 = toughest defense against this position, 32 = easiest, among teams with enough trailing games to rank |
| DvP sample size | Trailing games behind this position's DvP index - larger samples get less shrinkage toward league average |
| Season positional SOS | Mean opponent DvP index at this position across the full season; above 1.00 = softer than league average |
| Playoff positional SOS | Mean opponent DvP index at this position across the fantasy playoff weeks 15-17 |
| Survives to this pick | Probability the player is still on the board at THIS pick, from the availability model |
| Best available at next pick | Expected best VBD still available at your NEXT pick, at your slot |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rec TD | Projected receiving touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rush TD | Projected rushing touchdowns - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared season_projections row |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared season_projections row |
| TD dependency | Share of a player's projected points that comes from touchdowns - Touchdown Equity's own field, riding along on the shared season_projections row |
| Total TD | Projected total touchdowns across rushing, receiving and passing - Positional Stat Lines' own field, riding along on the shared season_projections row |
Draft Availability Simulator
Stop guessing whether your guy makes it back to you. A live simulation of the picks between now and your next turn, tuned so late-round uncertainty behaves like a real draft room instead of a flat coin flip.
What's included (8 fields)+
| Availability curve | P(available) at each of the 14 picks from your slot |
| Gone by | Expected pick number at which this player is taken |
| ADP | Consensus ADP driving the simulation |
| WR tiers left | Expected WR tier 1-4 survivors at that pick |
| RB tier 3 left | Expected RB tier 3 survivors at that pick |
| TE tiers left | Expected TE tier 1-2 survivors at that pick |
| QB tiers left | Expected QB tier 1-2 survivors at that pick |
| VBD | Points above the positional replacement baseline - Draft Value Engine's own field, riding along on the shared draft_availability row |
Game Forecast
Win probability, projected score, and a real range around both - free, forever, at every tier. It's the one number you can check against reality every single week, which is exactly why we don't charge for it.
What's included (10 fields)+
| P(home win) | Blended win probability: 75% market-implied, 25% Elo |
| Projected score | Mean across the simulated games for this matchup |
| Model margin | Blended expected margin |
| Projected total | Expected combined points |
| Total 5th pct | Low end of the 90% total interval |
| Total 95th pct | High end of the 90% total interval |
| Margin 5th pct | Low end of the 90% margin interval |
| Margin 95th pct | High end of the 90% margin interval |
| Confidence | Analyst confidence band on the outright take |
| Team Elo | Rating carried from every result 2002-2025 |
Verification Ledger
The accuracy audit most vendors never publish: every signal we've tested, including the ones that failed, with dates and sources attached. If a claim on this site can't survive being checked, it doesn't ship - and this is where you check it. Free at every tier, permanently.
What's included (9 fields)+
| Reliability curve | Predicted vs observed frequency per probability bucket for the availability model - free at every tier |
| Segment skill | Brier skill inside each designation bucket, against that bucket's own base rate |
| Verdict | target / gem / qualified / uncertain / unverified / killed |
| Model said | The signal the model produced, verbatim |
| Reality | What the depth-chart and beat-reporting check actually found |
| Verdict detail | The one-line reasoning, with dates and sources |
| Snap share | Prior-season snap share supporting or undercutting the role |
| Target share | Prior-season target share |
| Corrected volume | Hand-floored attempts where a 2025 backup is a 2026 starter |
Leg Correlation Matrix
Same-game player pairs almost never move independently, and most prop and lineup-building tools handle that with a warning label instead of a number. This package publishes the actual correlation for 744 same-game pairs - both touchdown counts and total fantasy points - so you know which pairs genuinely move together before you build a stack, and which don't.
What's included (34 fields)+
| TD count corr | Correlation between the two players' touchdown counts under the shared team-budget model |
| Anytime-TD corr | Delta-method correlation between the two anytime-TD indicators |
| Shared covariance | The covariance the pair inherits from the game environment - the driver of both correlations |
| Relation | same_team (shared budget) or same_game_opponents (coupled budgets) |
| Lambda A | Expected TDs for player A, from the props dataset |
| Lambda B | Expected TDs for player B |
| Anytime TD % A | Published anytime-TD probability for player A |
| Anytime TD % B | Published anytime-TD probability for player B |
| Archetype pair | The role match this leg was scored under - e.g. QBxWR1, RB1xRB1_opp - from THIS WEEK'S own projected volume, not the outcome |
| Points corr (measured) | Ex-ante, leakage-checked correlation of half-PPR FANTASY POINTS for this archetype pair, fitted on 2022-2024 and backtested on 2025 |
| Points corr gap | points_correlation_measured minus 0.0 - the independence every per-player interval on this site currently assumes when combined into a lineup |
| Points corr sample | Team-weeks behind the archetype-pair estimate |
| Points corr CI low | 95% bootstrap interval, lower bound, resampled by team-week |
| Points corr CI high | 95% bootstrap interval, upper bound |
| Points corr validated | Whether THIS pair's fitted correlation beat assuming independence on the 2025 held-out backtest, not merely in the fitting sample - true for 7 of 11 archetype pairs, false for the other 4 |
| Anytime TD % | P(at least one TD) = 1 - e^-lambda - Touchdown Equity's own field, riding along on the shared player_props row |
| Anytime TD % (reported) | The published, rounded anytime-TD probability - Touchdown Equity's own field, riding along on the shared player_props row |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared player_props row |
| Divergence | Elo-only minus market, in percentage points - Market Divergence Board's own field, riding along on the shared slate row |
| Expected rec. TD | Receiving share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected rush TD | Rushing share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected TDs | Poisson rate: expected touchdowns this game - Touchdown Equity's own field, riding along on the shared player_props row |
| Model implied | Unblended rating-based probability - Market Divergence Board's own field, riding along on the shared slate row |
| Market implied | Vig-free market probability from the two-way moneyline - Market Divergence Board's own field, riding along on the shared slate row |
| Attempts | Projected pass attempts for the week - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared player_props row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared player_props row |
| TD share | Share of the projection that is touchdown expectation rather than yardage - Positional Stat Lines' own field, riding along on the shared player_props row |
| 2+ TD % | P(two or more TD) = 1 - e^-lambda(1+lambda) - Touchdown Equity's own field, riding along on the shared player_props row |
| Volume hand-corrected | True when this week's volume projection was manually adjusted for a role change nflverse trailing data would not yet reflect - Positional Stat Lines' own field, riding along on the shared player_props row |
Outcome Board
Pick any group of players and get one combined projected range - floor, median, ceiling - built from their real measured correlation instead of pretending every player's outcome is independent of the others on the field. Built for stacking a QB with his own pass-catchers, or checking how a full lineup's range actually adds up.
What's included (45 fields)+
| Combined median (naive) | Sum of the selected players' own medians, spread combined assuming independence between every pair |
| Combined floor (naive) | 10th percentile of the group total under the independence assumption |
| Combined ceiling (naive) | 90th percentile of the group total under the independence assumption |
| Combined median (adjusted) | Identical to the naive median - correlation moves the SPREAD of a sum, never its centre |
| Combined floor (adjusted) | 10th percentile of the group total using the measured pairwise correlation for every pair this site has fitted, independence for every pair it hasn't |
| Combined ceiling (adjusted) | 90th percentile of the group total, same basis as the adjusted floor |
| Pair correlation used | The correlation actually applied for one pair in the selection - 0 unless the pair was measured |
| Pair source | measured_validated | measured_directional | same_game_unmatched_role | different_game - see limits |
| Matchup environment gap | Each selected player's own environment_gap from the Market Divergence Board, carried into the same view rather than requiring a second lookup |
| Correlation-adjusted board | The combined median/floor/ceiling computed with the measured pairwise correlation for every fitted pair, replacing the independence assumption |
| Pair sourcing | Per-pair correlation and source detail behind the adjusted board (pair_correlation_used, pair_source) |
| Matchup context | Each selected player's own environment_gap from the Market Divergence Board, carried into this view |
| Anytime TD % | P(at least one TD) = 1 - e^-lambda - Touchdown Equity's own field, riding along on the shared player_props row |
| Anytime TD % (reported) | The published, rounded anytime-TD probability - Touchdown Equity's own field, riding along on the shared player_props row |
| Archetype pair | The role match this leg was scored under - e.g. QBxWR1, RB1xRB1_opp - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Carries | Projected rush attempts - Positional Stat Lines' own field, riding along on the shared player_props row |
| Distortion direction | Which side the flagged distortion leans - under or over the frozen shipped projection - Market Divergence Board's own field, riding along on the shared matchups row |
| Distortion flag | True only when the matchup gap is large and validated - Market Divergence Board's own field, riding along on the shared matchups row |
| DvP confidence | Confidence in this position's matchup index, from trailing sample size and consistency - Market Divergence Board's own field, riding along on the shared matchups row |
| Matchup index | How much MORE (>1) or LESS (<1) the opponent has allowed to this position than league average - Market Divergence Board's own field, riding along on the shared matchups row |
| DvP league average | League-average value allowed to this position - Market Divergence Board's own field, riding along on the shared matchups row |
| DvP sample size | Trailing games behind this position's DvP index - Market Divergence Board's own field, riding along on the shared matchups row |
| Matchup validated | Whether this position's matchup index actually predicted future points allowed better than league average, on held-out data - Market Divergence Board's own field, riding along on the shared matchups row |
| Matchup rank | 1 = toughest defense against this position, 32 = easiest - Market Divergence Board's own field, riding along on the shared matchups row |
| Environment gap | Matchup-implied points minus the frozen shipped projection - Market Divergence Board's own field, riding along on the shared matchups row |
| Expected rec. TD | Receiving share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected rush TD | Rushing share of the player's expected-touchdown rate - Touchdown Equity's own field, riding along on the shared player_props row |
| Expected TDs | Poisson rate: expected touchdowns this game - Touchdown Equity's own field, riding along on the shared player_props row |
| Matchup-implied points | Points implied by the matchup-only read, independent of the frozen shipped projection - Market Divergence Board's own field, riding along on the shared matchups row |
| Attempts | Projected pass attempts for the week - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass TD | Projected passing touchdowns - Positional Stat Lines' own field, riding along on the shared player_props row |
| Pass yards | Projected passing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Points corr CI high | 95% bootstrap interval, upper bound - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Points corr CI low | 95% bootstrap interval, lower bound - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Points corr gap | points_correlation_measured minus 0.0, the independence assumption it replaces - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Points corr (measured) | Ex-ante, leakage-checked correlation of half-PPR fantasy points for this archetype pair - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Points corr sample | Team-weeks behind the archetype-pair estimate - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Points corr validated | Whether this pair's fitted correlation beat assuming independence on the held-out backtest - Leg Correlation Matrix's own field, riding along on the shared points_correlations row |
| Receiving yards | Projected receiving yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Receptions | Projected catches - Positional Stat Lines' own field, riding along on the shared player_props row |
| Rush yards | Projected rushing yards - Positional Stat Lines' own field, riding along on the shared player_props row |
| Targets | Projected targets - Positional Stat Lines' own field, riding along on the shared player_props row |
| TD share | Share of the projection that is touchdown expectation rather than yardage - Positional Stat Lines' own field, riding along on the shared player_props row |
| 2+ TD % | P(two or more TD) = 1 - e^-lambda(1+lambda) - Touchdown Equity's own field, riding along on the shared player_props row |
| Volume hand-corrected | True when this week's volume projection was manually adjusted for a role change nflverse trailing data would not yet reflect - Positional Stat Lines' own field, riding along on the shared player_props row |
NHL Game Forecast
Win probability, projected margin and a 90% interval for every NHL game, free - plus the model-vs-market comparison behind it: our Elo-only read, the devigged market read, and the gap between them.
What's included (9 fields)+
| P(home win) | Blended win probability where a market blend is available, Elo-only otherwise - see market_blend_applied |
| Model margin | Blended expected margin |
| Margin 5th pct | Low end of the 90% margin interval |
| Margin 95th pct | High end of the 90% margin interval |
| Market blend applied | True when a real SportsGameOdds market probability was blended in; false means Elo-only, and p_home_win then equals p_home_elo_only exactly |
| Model implied (Elo only) | Unblended rating-based probability |
| Market implied | Vig-free market probability from SportsGameOdds' two-way moneyline |
| Divergence | Elo-only minus market, in percentage points |
| Team Elo | Season-level historical rating (Hockey Databank prior), updated game-by-game from SportsGameOdds' own results going forward |
The Weekly Board
Every week, before you ask: the top touchdown probabilities ranked with matchup context, the top projected yardage at every stat, and twenty candidate lineups - each one simulated ten thousand times with measured player correlations, so you see the floor and ceiling of the whole lineup, not just its pieces. Archived before kickoff, scored after the games. The record is public either way.
What's included (9 fields)+
| Section | Which board this row belongs to: top_td, top_yardage, lineup, or lineup_slot |
| Rank | Position within the section, by the section's own ordering |
| Anytime TD | The same published probability the TD Equity board sells, ranked |
| Projected | Lineup rows: mean simulated half-PPR score. Yardage rows: the published yardage mean |
| Lineup p5 | 5th percentile of the lineup's simulated score - the correlated floor |
| Lineup p95 | 95th percentile - the correlated ceiling |
| Construction | How the candidate was built: chalk, stack:{team}, double:{team}, or ceiling-tilt |
| Fitted pairs | How many of the lineup's 28 player pairs used a MEASURED correlation rather than the labelled independence assumption |
| Lineup | The eight names, QB/RB/RB/WR/WR/WR/TE/FLEX |
Before you start a trial
The questions we actually get, answered plainly. Full detail always lives at /legal.
Do you sell picks, predictions, or betting advice?+
No, at any tier, ever. Every package publishes probabilities and ranges, never a pick, a stake, or an EV figure - that boundary is stated on every relevant package and is a deliberate limit, not an oversight.
How do I know the projections are actually accurate?+
The Verification Ledger is free at every tier and publishes the model's out-of-sample track record, including the signals that were tested and killed rather than shipped. It is the one package built specifically so you do not have to take our word for it.
Do I need a credit card to start?+
No. A new account gets full Active access free for 30 days with no card at all. Choose a plan before it ends and access continues without interruption; do nothing and access locks - never charged, never deleted - and purchasing any time afterward turns it back on. Accounts that existed before the free month replaced the permanent free Scout tier keep Scout permanently, also with no card.
What's actually different between Scout, Active, and Feed?+
Scout sees every row with some columns hidden. Active unlocks every column, read on the site, for one person making real roster decisions. Feed adds export (CSV/JSON), a metered API, and version-pinned datasets for building outside our pages.
Can I cancel or change plans any time?+
Yes - cancel whenever you like and your access runs to the end of the period you already paid for. Full terms are at /legal/refunds.
What happens to my money if I cancel right after paying?+
Cancel within seven days of a charge and you have not exported anything, and we refund it in full on request. After that we do not refund part-months, because the data was already served - the exact policy is at /legal/refunds.
What sport does this cover?+
Mostly NFL - player projections, availability, matchup data and the Weekly Board synthesis for season-long fantasy, DFS, and player props, across twelve packages. NHL is one package so far: team-level Game Forecast (win probability, projected margin, model-vs-market read). No player-level NHL data yet - no rosters, no props, no injury feed, no draft content.
Is there an API?+
Yes, on the Feed plan: a metered API at 5,000 requests a day, plus every release version-pinned and hash-verified so you can rebuild a dataset and get back byte-identical files. The full endpoint reference is public at /docs/api - no account needed to evaluate it.
How do I get in touch?+
Email support@apexdatum.io directly - a real inbox, not a contact form. Response time depends on plan; see the Support row on each tier above.
What do the terms and conditions cover?+
What the product is and isn't - no picks, no EV figures, no wagering advice, at any tier - how accounts and API keys work, plan and seat rules, that one-off passes are dated and non-refundable (with a carve-out if access never arrives), how dataset and methodology changes are handled, liability limits, and how cancellation works. Full text is at /legal/terms.