Chapter 37 — Exercises

How to use these. No answers are given. Items marked † are the harder ones — they require you to reason across the table rather than to read a row out of it, and several have no single correct response. Work them with the master table (§37.5) and the claim-form table (§37.6) open; work the dossier items with your own dossier open.

Nothing here asks you to decide whether to take anything. If an exercise makes you want to, that is the signal to reread §37.8's safety callout and then talk to a clinician.

Part A — Reading the table

A. Find every row in the master table whose compound is semaglutide. List the claim, the tier, and the chapter for each. How many rows are there, how many distinct claims do they resolve to, and how many tiers are represented? Explain the gap between the first two numbers.

B. Growth hormone appears twice with opposite ratings. Write out the two claims in full and underline the exact words that differ. In one sentence, state the general principle those words illustrate.

C. Locate the three BPC-157 rows. Two are duplicates from different chapters and one is a different claim. Identify which is which, and say why a reference table should keep a duplicate rather than merge it.

D. Which therapeutic block has the highest proportion of ✅? Which has the highest proportion of ❌? State both as fractions, not impressions.

E. Find all four NOT RATED entries. For each, write one sentence identifying the specific word or phrase that makes the claim unrateable.

F. † Two rows in the master table rate an inference rather than a molecule. Find them. Then find a third row anywhere in the table that arguably does the same thing, and defend your choice.

G. The claim-form table has eleven ❌ rows. Group them into no more than three families by the kind of inferential error each one makes, and name each family.

H. Find every row in the table whose Rating cell contains a slash. There are ten. Sort them by which of the three kinds of split (§37.2) each one is.

I. † Pick any five rows rated ✅. For each, decide from the claim column alone whether the endpoint is a hard clinical outcome or a surrogate. Then rank your five by expected durability and justify the ranking.

J. Colistin and daptomycin are ✅ while "antimicrobial peptides will provide a new class of broad-spectrum systemic antibiotics" is 🔬. Explain the compatibility of those three ratings in three sentences, without using the word "but."

Part B — The two kinds of ❌

K. In your own words, and without looking, state the difference between an ❌ where evidence is absent and an ❌ where evidence is present and negative. Then check §37.3 and note anything you left out.

L. Six rows in the master table are marked (N). List them. What do the source chapters have in common that earned the marking?

M. † Choose three unmarked ❌ rows that you believe are also evidence-present-and-negative, and say what you would need to read in the source chapter to confirm it. Then say what you would conclude if the source chapter did not support your guess.

N. Explain why "absence of evidence is not evidence of absence" is logically correct and rhetorically misleading in the peptide market. Use the kitchen-elephant argument or an equivalent of your own.

O. † Chapter 7 rates native GLP-1 as ❌ for a reason that fits neither of §37.3's two categories. Describe the third situation, and construct one more example of it — a claim that would be rated ❌ even though the molecule demonstrably does the biological thing it is claimed to do.

P. Rank these four states of knowledge from most to least informative, and defend the ranking: (i) large trials ran and failed; (ii) no trials have run; (iii) one small trial was encouraging; (iv) the mechanism is well characterized in cells.

Q. † Which is a stronger reason to be skeptical of a compound sold to consumers: that its ❌ is evidence-absent, or that its ❌ is evidence-present-and-negative? Argue both sides, then commit.

Part C — The rating system on itself

R. State the six rules of §37.2 from memory. Check yourself against the section, then write one sentence on which rule you find hardest to apply and why.

S. Rule 3 says never upgrade with mechanism. Find one row in the master table where the mechanism is described as settled and the clinical claim is nonetheless not ✅. Explain what that pair shows.

T. Rule 4 says never downgrade with distaste. Find the ⚠️ row that most tempts you to argue it should be ❌, and write the strongest honest case for keeping it at ⚠️.

U. † Why is NOT RATED not a fifth tier? Construct the strongest argument for making it one, and then rebut it.

V. All four NOT RATED entries are in Part VIII. Write two paragraphs: one arguing this shows the rating system has a well-defined domain, and one arguing it shows Part VIII asks questions the book should not have asked. Which do you find more persuasive?

W. † Take the Chapter 44 obesity-policy claim and rewrite it three times: once so it is rateable and would probably get ✅, once so it is rateable and would probably get 🔬, and once so it remains unrateable for a different reason than the original. Label each.

Part D — Distribution and aging

X. The distribution is 54 ✅ and 50 ❌. Write the two-sentence version of what that does and does not show — including why the four-rating margin is not something to argue from. Then write the one-sentence version of what the structure underneath it shows.

Y. ⚠️ is the rarest of the four main tiers. Explain why, using the idea that a tier can be expensive to reach.

Z. † Eight of the ten splits are ⚠️/❌. What single recurring situation produces that pattern? Give three examples from the table and say what the eight splits collectively tell you about how peptide claims are marketed.

AA. Which five rows in this table do you expect to have moved by 2031? For each, name the direction and the specific result that would move it.

AB. Which five rows do you expect to be unchanged in 2041? Defend each in one sentence.

AC. † Chapter 9 establishes that phase 2 effects usually shrink in phase 3. Apply that to the ⚠️ rows in the metabolic block. Which are most exposed to that effect, and which are least? Does your answer change if you consider that some of those compounds already have late-stage data?

AD. A 🔬 rating that is still 🔬 in ten years should be reread with suspicion, per §37.9. Why? What would that persistence most likely indicate?

Part E — Dossier and drift

AE. Complete the Step 2 drift audit in the dossier section for every claim you have tracked. Do not read §37.5 for any claim you have not yet rated yourself.

AF. Report your three numbers — MATCH, HIGHER, LOWER — and state which of the three patterns in Step 3 you fall into.

AG. † Sort your disagreements into the two piles of Step 4. Then answer honestly: is pile two larger than you expected? What do the entries in pile two have in common?

AH. Write the Step 5 sentence. One sentence, dated, about yourself rather than about peptides.

AI. † Find one row in the master table you genuinely believe is wrong. Write the four-line callout you would issue instead — claim, rating, why, and what would change it — and identify the specific evidence you are weighting differently from the source chapter. Then write one sentence on whether your disagreement runs in the same direction as the rest of your drift audit.

AJ. Take one compound not in this table — anything currently being discussed that this book does not cover. Rate it yourself using the five questions of §37.9's update procedure. Note how long it took and where you got stuck.