Chapter 28 Quiz — Analytics and Offensive Decision-Making
22 questions. All Outlaws and opponent numbers are fictional. Use the illustrative expected-points ladder from 28.2 where needed (1st-and-10 at own 20 ≈ +0.4; own 30 ≈ +1.0; midfield ≈ +2.0; opp 35 ≈ +2.8; opp 20 ≈ +3.8; opp 10 ≈ +4.6).
Part 1 — Multiple Choice
1. Expected points added (EPA) on a play is:
- A. The points scored on the drive containing the play
- B. The change in expected points from before the snap to after it
- C. Yards gained divided by yards needed
- D. The probability of winning after the play
Answer
B. EPA is the change in expected points. Win probability (D) is a different measure, used more late in games.
2. Under the common success-rate rule, which play is successful?
- A. 1st-and-10, +3
- B. 2nd-and-10, +5
- C. 2nd-and-5, +3
- D. 3rd-and-4, +3
Answer
C. Second down needs 60% of the distance; 3 is 60% of 5. A needs 4; B needs 6; D needs all 4.
3. In this book, an explosive play is:
- A. Any play of 15+ yards
- B. A run of 10+ or a pass of 25+
- C. A run of 12+ or a pass of 20+
- D. Any play with EPA above +1.0
Answer
C. Runs of 12+ yards and passes of 20+ — the same definition the companion defensive book uses.
4. The Outlaws' play-action averaged about 8.6 yards per dropback against about 6.2 for straight dropbacks, but only about 8.6 vs. 7.0 on first and second down. The best explanation for the shrinking gap is:
- A. Play-action gets worse on early downs
- B. Selection: straight dropbacks included many third downs, where everyone's yards per attempt drops
- C. Marisol's model was miscalibrated
- D. Defenses stopped respecting the run
Answer
B. Most play-action calls came on early downs; many straight dropbacks came on third down. Comparing like with like removes part of the difference.
5. On fourth-and-1 at the Austin 43 (Week 12), the illustrative break-even conversion rate was about:
- A. 30%
- B. 53%
- C. 70%
- D. 85%
Answer
B. Going was worth p × 1.7 + (1 − p) × (−2.3); punting about −0.2. They're equal at about p = 0.53.
6. Which objection passes the "name it" test?
- A. "I've got a bad feeling about this."
- B. "Their defense is playing well today."
- C. "Our left tackle just left the game, and their best rusher is over his backup."
- D. "We always punt here."
Answer
C. It's specific, observable, and something the model doesn't know. B is already built into the model's assumption of a defense; A and D are not names.
7. Wade Harmon says "Four-down" before third-and-2 in the opponent's territory. The best use of that information is usually to:
- A. Call the same third-down play you'd call otherwise
- B. Consider a higher-payoff call, since a likely failure (an incompletion) still leaves a fourth down you'll play
- C. Punt on third down
- D. Call a draw to set up fourth-and-short
Answer
B. Third-and-short becomes one of two downs to gain the distance. Avoid plays whose likely failure is a sack.
8. A team trails by 14, scores a touchdown, and now trails by 8 before the try, late in the fourth quarter. The analyst-favored choice, still debated, is to:
- A. Kick, to make it a one-score game
- B. Go for two; convert and a later TD + PAT wins; fail and a later TD + two still ties
- C. Take a knee
- D. Onside kick before the try
Answer
B. Going for two at down 8 tells you early whether you'll need two again and gives a chance to win in regulation. Many coaches still kick; it depends on your conversion rate.
9. Marisol's rule of thumb says a success rate measured on about 20 snaps can wander by luck alone by roughly:
- A. 2 points either way
- B. 5 points either way
- C. 20 points either way
- D. 50 points either way
Answer
C. About 20 points either way on 20 snaps; about 10 on 100; about 5 on 400.
10. Why does Marisol code RPOs as their own family instead of "run" or "pass"?
- A. RPOs are illegal in some leagues
- B. Classifying by outcome credits correct reads to the wrong family and makes both runs and passes look better than they are
- C. RPOs never gain yards
- D. The NFL requires it
Answer
B. If handoffs count as runs and throws as passes, the conflict that made each work disappears from the comparison.
11. In the NFL, a two-point try is snapped from the:
- A. 1-yard line
- B. 2-yard line
- C. 3-yard line
- D. 15-yard line
Answer
B. The 2. The extra-point kick is snapped from the 15 (since 2015). College and high school snap the two-point try from the 3.
Part 2 — True / False (with a one-line justification)
12. True or false: If a fourth-down decision fails, the decision was wrong.
Answer
False. A 70% play fails three times in ten; grade the decision by its inputs and the odds, not the single result.
13. True or false: A 29-for-29 defensive tendency is more trustworthy in a playoff game than a 23-for-29 tendency.
Answer
False (or at best "not necessarily"). The opponent knows you're counting, and a perfect streak may be a setup; it also tells you nothing about how it breaks.
14. True or false: A five-yard run on first-and-10 is a successful play but adds almost no expected points by itself.
Answer
True. It clears the 40% line, but moving from first-and-10 to second-and-5 barely changes expected points; its value is in keeping the next call open.
15. True or false: Most public analysis finds that, on average, early-down passes produce more expected points per play than early-down runs.
Answer
True — with the selection caveat: teams don't run and pass in the same situations.
16. True or false: In high school, the break-even for going on fourth-and-short is usually higher than in the NFL because kicking is less reliable.
Answer
False. Less reliable kicking and shorter punts make kicking worth less, which lowers the break-even and favors going.
Part 3 — Short Answer
17. In one or two sentences, explain why Wade's question "What's the call if we go?" is really a test of the fourth-down model's first assumption.
Answer
The model assumes a league-average play for the distance. Wade's question forces the coordinator to name the actual play and say whether it's at least that good against the front they'll see — if it isn't, the model's recommendation is built on a false premise.
18. What is a "comfort call," and how can a self-scout find one?
Answer
A play the caller reaches for to manage his own emotions after a bad play. Sort the self-scout by "the play called most often right after a negative play" and compare its success rate with the same down and distance when it didn't follow a negative play.
19. Name the four lines of Marisol's slide template, in order.
Answer
The number (as a count), the sample, "This doesn't know…," and "What I'd do."
20. Give two reasons the Outlaws' Game 2 punt on fourth-and-1 at their own 36 can be defended even though the chart said "go, not by a little."
Answer
Lyle couldn't name a fourth-and-1 play he trusted against that front that day (their runs vs. big nickel had been bad, Rollins was beating Achterberg, and a third-and-1 had already failed), so the model's league-average-play assumption wasn't true; and the head coach priced those named factors honestly rather than kicking out of habit. (The lesson is to build a package that makes the assumption true next time.)
Part 4 — Applied Scenarios
21. (hypothetical) Third quarter, you lead 20–17. Fourth-and-2 at the opponent's 38. Punt would pin them near their 10 (≈ +0.1 for you); a conversion gives you about +2.8; a failure gives them the ball at their 38 (≈ −1.2 for you). Your backup left guard has been beaten twice by the 3-technique. Find the break-even, decide, and name a call.
Answer
p × 2.8 + (1 − p) × (−1.2) = 0.1 → 4.0p = 1.3 → p ≈ 0.33. You need to convert only about one time in three, so go even if the guard problem lowers your chance noticeably. The guard changes the call: avoid an inside run at him; prefer a perimeter or quick-game call (a sprint-out snag away from him, or a quick throw) that doesn't ask him to win.
22. (diagnose) A staff saw "opponent pressures on 15 of 26 third-and-7+ snaps" and called max protection with two-man routes on every long third down against a veteran quarterback's team. The defense rushed four and dropped seven on most of them, and the offense went 1 for 5. What went wrong, and what's the fix?
Answer
Aggregation hid the split that mattered: the defense pressured 8 of 9 against rookies and only 7 of 17 against veterans. Against a veteran, the relevant rate was well under half. Fix: split tendencies by the condition that matters for your team (here, QB experience) and report counts so subgroups can't hide.
Scoring Guide
| Section | Questions | Points each | Total |
|---|---|---|---|
| Multiple choice | 1–11 | 2 | 22 |
| True/False with justification | 12–16 | 2 (1 for the answer, 1 for the reason) | 10 |
| Short answer | 17–20 | 4 | 16 |
| Applied scenarios | 21–22 | 6 | 12 |
| Total | 60 |
54–60: You can run Marisol's Thursday meeting. 45–53: Solid — reread 28.6–28.7 and redo the break-even arithmetic. 36–44: Rework 28.2–28.4 (EP, success rate, selection) before moving on. Below 36: Start again with the hook and 28.1; the rest of the chapter builds on Marisol's honesty rules.