MISS, BACK, and BYE: How FantasyOMatic Predicts Availability

Most fantasy sites stop at a colored injury badge. It tells you a player is questionable, out, or on reserve, but it does not answer the decision that follows: when should this player become usable again, and what should disappear from the interface until then?

FantasyOMatic separates those questions. Official status remains official. Schedule remains schedule. Our model adds a clearly labeled return-timing estimate, then one availability engine carries that same answer into Rankings, Player Detail, My Team, Matchup Matrix, AI Coach, and Podcast.

The language is intentionally direct:

• OUT, IR, Q, D, and BYE are official or schedule facts.
• ≈ OUR ESTIMATE identifies FantasyOMatic's forecast.
• MISS means we expect the player to be unavailable that week.
• BACK is the first week our model allows the saved weekly projection to return.
• LOW, MED, or HIGH describes evidence strength, not certainty.

What the model predicts - and what it does not

The model predicts return timing after an injury has already been reported. It does not predict who will get hurt next. It does not diagnose an injury, read private medical records, or override a club's designation.

That distinction matters. If a player is officially OUT, FantasyOMatic displays OUT as the current fact. A future `MISS W10 → BACK W11 · MED` timeline is a separate model statement. The interface never quietly upgrades our estimate into an official return date.

The inputs behind the estimate

The current estimator starts with public weekly football information:

1. Primary injury - hamstring, knee, ankle, foot, concussion, shoulder, and more than 30 other normalized categories.
2. Position - the same injury can have a different football cost for a quarterback, running back, receiver, or tight end.
3. Practice participation - full, limited, or no practice can move the estimate when a usable report is available.
4. Time on the report - a first-week listing is different from an injury that has persisted for four weeks.
5. Game timing - a Thursday game has less recovery time; a Monday game has more.
6. Official status and reserve eligibility - OUT, Doubtful, Questionable, IR, PUP, and return eligibility constrain the result.

The calibration uses more than 24,000 historical injury-report stints from 2021 through 2024. Only stints that reached OUT or Doubtful at least once are included in the duration lookup, so a player who remained Questionable while playing every week does not distort missed-time medians.

In plain English, the calculation is:

historical injury duration × position × practice × game timing × time on report - time already elapsed

Official status then sets a floor. An OUT player cannot be transformed into an immediate return by one optimistic modifier. A short-duration Questionable estimate can remain playable, but its weekly projection may still be reduced.

Confidence is evidence strength, not a promise

The confidence badge is deliberately not a percentage. It scores three things: the size of the historical injury sample, whether practice information is usable, and whether multiple report weeks establish a direction.

• LOW usually means a small or unknown injury sample with little current evidence.
• MED means at least one meaningful evidence source is present.
• HIGH generally requires a large sample plus practice direction, repeat reports, or both.

Football recovery is not clockwork. Confidence tells you how much evidence supports the estimate; it does not guarantee the exact week.

A real Week 8 example: Bucky Irving

The frozen 2025 Week 8 testing snapshot gives us a clean example. Bucky Irving carried an official OUT designation for a foot injury.

FantasyOMatic combined the historical foot-injury duration, his running back position, his time on the injury report, and the current status. The resulting public timeline was:

W8 OUT · W9 BYE · W10 MISS · W11 BACK · MED

Week 9 is the important part. Tampa Bay had no game, so BYE wins. The system does not relabel a schedule fact as an injury estimate. Week 10 is the scheduled pre-return game, so it becomes MISS. Week 11 is the modeled return boundary, so it becomes BACK.

This is also why FantasyOMatic never describes a MISS or BYE as a zero Player Rating. Bucky's opponent-neutral talent still exists. What disappears is the Week Rating and projected fantasy points for a week in which he is not expected to play.

What changes across the product

One structured timeline drives every decision surface:

Rankings and Player Detail

The Player Rating remains visible as the neutral talent baseline. An unavailable current or future week shows status instead of a numeric Week Rating or projected points. The player page can still explain the injury, estimate, confidence, and expected return boundary.

My Team, Dashboard, Compare, and Fix Byes

Lineup tools suppress unusable weekly output, reduce remaining-season availability, and surface replacement timing. Compare can still show long-term talent, but it cannot pretend an expected missed game is startable.

Matchup Matrix

The Matchup Matrix (All-Pro) keeps two different questions separate. Matchups Only still shows the opponent effect during a MISS week because schedule difficulty exists independently of health. Apply Player Ratings hides the unavailable rating. Schedule Edge ignores injuries; Player Outlook reflects expected missed games.

AI Coach and Podcast

AI Coach and Podcast (Hall of Famer) receive the same structured availability timeline. Generated advice must say “Our injury model expects...” before presenting model timing. Deterministic validators reject advice that recommends a MISS player, quotes a hidden projection, calls an estimate official, or treats reserve eligibility as guaranteed activation.

How missed games affect remaining value

BYEs and projected absences are not the same calculation.

A BYE is excluded from the scheduled-game average because no game exists. A scheduled MISS remains a zero-effective-output game inside the remaining horizon. FantasyOMatic then tracks how many games are expected to be available and applies that availability to rest-of-season and playoff decision signals.

This lets the product say two true things at once:

• The player may still be talented when healthy.
• The player's usable remaining value is lower if games are expected to be missed.

Production proof, not just a design promise

Before the corrected Week 8 snapshot went live, FantasyOMatic exported all 632 production rows for rollback and rehearsed the full pipeline in isolation.

The final publication contained:

• 39 players with return estimates
• 23 future MISS cells
• 32 BACK boundaries
• 352 schedule BYE cells
• zero unavailable point leaks
• zero unavailable Week Rating leaks
• zero pre-return timing violations
• zero BYE provenance violations

Kickers use the player injury timeline. Team defenses remain schedule-only. Matchup Matrix Schedule Edge stayed unchanged because injury availability must not rewrite opponent difficulty.

What the model still cannot know

Transparency includes the boundaries:

• Complete Wednesday-through-Friday practice trajectories are supported, but the live feed does not always preserve all three days.
• The historical duration table is intentionally reviewed and versioned; it is not silently retrained every week.
• Repeated same-body-part injury is not yet a separate recurrence feature.
• BACK means weekly output may reappear. It does not promise a full snap count or official activation.
• Public reporting can change after FantasyOMatic's latest scheduled refresh.

These are reasons to label the estimate, not reasons to hide it.

How to use the signal

Start with the official status. Then read the timeline and confidence together. A MED estimate that spans two scheduled games should change roster planning more than a generic red badge, but it should never be treated as medical certainty.

Use Rankings to see which weekly values remain usable, open Player Detail to inspect the full availability story, and use My Team (All-Pro) to identify the actual roster decision created by the absence.

The product rule is simple: availability is a status, not a talent score.