The Engine Key
Every snap is dice. Every die is loaded by your players' ratings and your game plan. This page is the loading, written out — the same arithmetic the replay's Why this play receipt prints for every snap. Ratings run 1–20; "effective" means the rating plus a class-year bonus (seniors +1, freshmen −1), plus conditioning, less fatigue. Nothing here is hidden from the receipt.
One snap, in order
1Your plan calls run or pass for the down and distance (Q32–41); their plan predicted it (Q1–10). The gap between their prediction and what you did is the read — up to ±26 tenths of a bucket, symmetric: a right read is punished exactly as much as a wrong one is rewarded.
2Their Front slider and Blitz % resolve to one front for the snap: Run-stop, Balanced, Pass-rush or Blitz. Your formation mix draws a formation.
3Everything that matters becomes a shift, in tenths of a yardage bucket: the scheme-vs-front table, your line against their front seven, the carrier or the coverage, the formation, the dropback, a bracket, a second-teamer. The terms are added; the sum is the shift.
4The dice draw a yardage bucket from the play's profile. The shift gets one chance to move it a bucket — |shift| × 100 / 150 percent, capped at 60%. Then the read gets its own chance: min(26, |read|) / 32 for one bucket, and a second bucket only when the read is big. This is the whole design: a good plan and good players nudge the dice every snap, and nothing ever makes one snap a sure thing.
5Yards are drawn inside the bucket, capped at the goal line. Tacklers, assists, fumbles, out of bounds, the clock.
Running the ball
The shift
shift = front(scheme, their front) the matchup table, e.g. a Draw into a Blitz +46, a Dive into a Blitz -18
+ keys & pursuit -(their key on your back - 33)/4, pursuit vs a sweep, misdirection vs a flying pursuit
+ formation I-Form +6 on a run, Spread -4 ...
+ (your OL run-block - their run-stop) × 3 OL = avg(blk, str) of the five starters; run-stop = DL avg(tkl, str) ×3 + LB avg(tkl, awr) ×2, over 5
+ (carrier SPD - 10)
+ second-team penalty if the plan is using backups here
+ home field +5 at home, -5 on the road
Then the dice: the bucket profile for the scheme (Draw, Dive, Trap, Slant, Sweep, Power, Toss, Zone each have their own), the shift's one chance to promote it, the read's chance, the yards inside the bucket.
| Run buckets | loss | 0–2 | 3–5 | 6–9 | 10–19 | 20–39 | 40+ |
|---|
| Dive, per 1000 | 45 | 348 | 385 | 168 | 42 | 3 | 9 |
|---|
| Toss, per 1000 | 90 | 250 | 330 | 205 | 95 | 18 | 12 |
|---|
A 90-rated line against a 5-rated front is +51 tenths before anything else — a 34% chance every carry of one bucket up, on top of whatever the dice drew. That is how the talent ladder's 90s run for 49 a game.
Throwing it
First, the sack (per 1000)
sack = 72 + (their pass-rush - your pass-block) × 4 + blitz 40 + depth (deep throws take longer) + pass-rush front 14 + rollout -18
- (QB AWR - 10) × 3 - read/2 floor 15 per 1000
Then the interception (per 1000)
pick = base by depth (short 12, medium 32, deep 56) + zone 4 or man -2 - (QB ACC - 10) - coverage shift/8 - read/3 floor 3
Then the completion (per 100)
complete = base by depth (short 67, medium 52, deep 31) + weather
+ (QB ACC - 10) × 2 + (receiver HND - 10) - (their coverage - 10) × 2
+ coverage shift/8 + read/2 + home field (+1 at home, -1 away) clamped 5..95
their coverage = DB avg(spd, awr) ×4 + LB avg(awr, spd), over 5
Then the yards: the coverage shift
shift = coverage vs depth zone gives up the short throw and takes away the deep one; squat / level / bail; press vs off; match; a blitz leaves one-on-one (+14 short, +22 deep)
+ their front pass-rush front -13, run front caught light +12
+ formation + dropback + bracket (-bracket%/3 when the bracketed man is the target) + second-team penalty + home field (±5)
| Pass buckets | loss | 0–3 | 4–9 | 10–23 | 24–45 | 46+ |
|---|
| Short, per 1000 | 45 | 135 | 470 | 335 | 15 | 0 |
|---|
| Deep, per 1000 | 13 | 33 | 57 | 281 | 416 | 200 |
|---|
Who is thrown to: your target shares (Q56–60), then a profile fit for the depth. Who tackles: weighted by TKL and how deep the play got. The pass rush: DL avg(str, spd) ×4 + LB speed, over 5; protection: OL avg(blk, awr).
The ball on the ground
fumble (per 1000 touches) = base (runs 24, catches 9, a sack 13 + the sacker's power)
+ ((tackler STR + TKL)/2 - 10) × 2 - (carrier HND - 10) × 2 + 8 when hit behind the line + weather (rain +6, snow +10)
Kicking
field goal % = 113 - distance + (kicker ACC - 10) × 2, and -3 a yard past 45, + weather (-wind/2 over 10 mph, rain -2, snow -5) clamped 3..98
range the AI will try from = 41 + kicker STR yards
kickoffs and punts: kicker STR sets the leg, returner SPD the return; wind takes yards off both
The line
power rating = 0.5 × offense + 0.45 × defense + 0.05 × kicker, on the 1-20 scale
offense = (OL run-block ×2 + OL pass-block ×2 + QB(ACC ×3, AWR ×2, STR) ×3 + RB(SPD ×3, HND) + top-3 WR(HND ×2, SPD) ×1.5 + TE(HND, BLK) ×0.5) / 10
defense = (run-stop ×3 + pass-rush ×3 + coverage ×4) / 10
spread = 3 + 6.0 × (home rating - away rating) home field 3; 5.5 a point on the talent ladder, 8.1 on this season's real games
Set when the week is built, graded in the digest. Home field is one number for the league: the home offense gets +5 tenths of shift and +1 completion on every snap, the visitor the reverse, which comes out to a 3.2-point mean home margin over 400 calibration games. Scorecard so far: the line's mean miss is about 13 points and it lands within a touchdown roughly a third of the time - the same as a real closing line.
Weather
Rain: completions −3, fumbles +6 per 1000, run lean +2. Snow: −6, +10, +4. Cold: completions −1, run lean +3. Wind over 10 mph: deep completions −2 per 4 mph, medium −1, one deep bucket down per 5 mph, field goals −½ a point per mph, kicks shorter. The forecast is posted all week; game day rolls the actual from it.
Penalties
Flags are thrown by the rulebook, not by the scoreboard. Two rolls a snap: one before it and one after,
about 11 flags per 100 snaps league-wide before declines — which lands near the real 7.7 a team a game once
the declines come off. Nothing in your game plan raises or lowers the rate; no team is dirtier than another.
before the snap: 42 in 1,000 — the down is replayed either way
False start 45% · Illegal motion 7% · Delay of game 6% · Encroachment 7% five against the offense
Offside 35% five for the offense
after the play: 68 in 1,000
on the offense 42% — Holding -10 · Illegal block in the back -10 · Illegal formation -5
always the previous spot: the play is wiped and the down replayed
on the defense — Pass interference (an incomplete throw only) 15, or a spot foul 5-14 when the contact came early
Defensive holding 10 · Roughing the passer / Face mask 15 · Offside 5
every one but the offside is an automatic first down
Encroachment is the college foul (7-1-4): an offensive lineman in the neutral zone before
the snap, five yards against the offense — it takes the false-start signal on the replay. The defense's version is
offside, which replays the down and gifts five.
Where the yards go from
The spot is the rulebook's call, not the coach's (10-2-2). An offensive foul is always the previous spot
and the play is wiped. A defensive foul on a pass is the previous spot too — which is why an offense that just
gained 51 declines it. A defensive foul on a run that got past the line is added on: the play stands and the
yards go on top. Roughing the passer and a personal foul after the catch are the two exceptions that add on to a completion.
A change of possession voids every add-on.
Declining
Either side takes whichever leaves it better off (10-1-1), and the replay's receipt prints the reasoning in
words — “3rd & 7: the run that lost 2 fell short, the flag moves the chains”. A touchdown always stands.
A flag on a turnover is always taken, because it wipes the turnover. An add-on is always taken, because it is pure gain.
Everything else is weighed: whichever moves the chains, then whichever gains more, ties to the flag. A declined flag still
gets its signal on the replay — arms swept apart, over the veil.
Injuries
who rolls: the players in the collision — carrier, tackler, assist, sacked QB, receiver after the catch, returner — 14 in 10,000 per contact, weighted by position (a back takes a harder hit than a lineman)
off the ball, every snap: one man from each of the trenches, the coverage and the receivers, 4 to 12 in 10,000; the passer after the throw, 6
who it lands on: DB 20% · OL 17% · WR 15% · DL 12% · LB 11% · RB 10% · TE 7% · QB 4% · FB 2% — the shape of a real injury report
how bad: 70% rest of this game · 30% out 1 or 2 weeks · nothing longer, no season-enders
season cap: 2 weeks sent home per player per season — at the cap he can still leave a game, never miss a week
the floor: QB 2 · RB 3 · FB 1 · WR 5 · TE 2 · OL 6 · DL 5 · LB 4 · DB 6 healthy — at the floor, rest-of-game only
immune: kickers and punters
What it costs you
“Out 2 weeks” is the next two games, playoff rounds included; he is back the week after. He stays on your roster and your depth chart, the engine skips him, and the next man at the position takes his snaps, carries and targets. Nothing else moves: no playing hurt, no lingering effect, no injury-prone players, no ratings hit, and coaching still lands on him while he sits. The offseason heals everyone and resets the cap.
What you can do
Set the depth chart. The order behind your starters is who plays when one goes down; an untouched chart plays best available, which is what happens anyway. There is no way to speed a recovery. Expect about six injuries a season, one or two of them costing weeks — a starter missing a game or two a year is normal, a season lost is not possible.
Where to see it
The Game Plan page opens with your injury report before you plan. OUT tags on the roster and depth chart. After the game: the box score lists them, the replay shows the cart, the digest names the week's. League-wide: the Injury Report.
So is it random? Yes, and no. The dice are real: an 89% throw misses one time in nine, and the receipt will show you the one. But every number above is a rating or a decision, and the talent ladder — 48 clone programs on eight rungs, 1,700 games — came out 90-rung 89%, 20-rung 9%, 90 vs 20 54–0 by 63 a game, never inside three scores. Neighbouring rungs are a 67–74% proposition for the better side; a good plan against a poor one is worth 4–6 points, about two-thirds of a rung. The ratings decide the mismatches; your decisions decide the close ones. See
Play Key,
Defense Key.
