The Thirty-Day Regimen

Chapter 15 gave the instruments. This chapter puts dates on them: thirty days, twenty to forty minutes each, everything drawn from material already in this book. No equipment beyond a notebook (the journal), a timer, and occasionally coins, matches, or a deck of cards.

How to read the references. Puzzles are named by chapter·number: 5·6 means Chapter 5, canon puzzle 6; C·30 means Appendix C — the Reserve Rack — canon puzzle 30, same ritual with no chapter warming you up. The number after the dot is the puzzle's canon number, printed in its heading where you first met it. Nothing is restated here — every reference lands on a self-contained ritual with the statement, the pause line, and the solution below it. Cold means: work from the statement only, with the solution covered. The spoiler discipline from How to Use This Book is now load-bearing.

You have seen these puzzles. That is the point, not a flaw in the plan. The regimen does not test whether you remember answers; it tests whether the move comes to hand — whether you can rebuild the path, name the pattern, and feel the recognition signal fire before the memory of the answer arrives. Re-solving from a blank page is a different act from recalling, and the journal (which demands the reasoning in writing, not the answer) keeps the two honest. Where memory does contaminate, the clock catches it: rebuilding is fast when the move is owned, halting when only the answer is.

The rules, once:

1. Twenty to forty minutes. Set the timer. Stop at forty — the edge, not exhaustion.
2. The journal is half the training. Every miss gets the five-line entry from Chapter 15, and the missed move gets named.
3. Struggle rule: twenty minutes of real motion before any peek; then one hint's worth, then back in.
4. A miss is upsolved the same day or it enters the review deck. Nothing is abandoned.
5. A skipped day is skipped. Do not double up; the deck keeps score, not the calendar.
6. AI only per Chapter 15: variants and post-commitment attacks. Never its solution first.

Deck day (each week's last day) means: re-solve, cold, every miss the journal has collected, and log the re-solve times. Calibration set means: five Fermi questions with 80% intervals, scored, per Chapter 15's protocol.

Day zero, ten minutes. Copy the five-line entry template onto the journal's inside cover. Bookmark Appendix A; it will be consulted daily. Then run calibration set 0 — five questions, five intervals, scored and dated — so the curve has a baseline nobody can later flatter. The regimen measures change, and change needs a before.

The anatomy of a session

The same thirty minutes, dissected, so the schedule below can stay terse:

• Minutes 0–3 — read and frame. Read the statement twice. Write one line: what kind of problem this smells like, and which move the smell suggests. This line is the day's real bet; the journal will grade it.
• Minutes 3–23 — motion. Representations tried, cases enumerated, small instances computed. When stuck for ten of these minutes, run Chapter 15's constraint drill: every rule you are obeying, marked STATED or ASSUMED.
• Minute 23 — the hint decision. Still nothing? One hint's worth: reread the puzzle's Move line only — the bolded name, not the paragraph — then back in.
• Last 5 minutes — the ledger. Solved or not, write the entry: the move, named; the signal, in a sentence; the miss class if it was a miss. On a miss, schedule the upsolve for today or file the card.

Two people can run the regimen on one book, and it improves: alternate days as solver and verifier, with the verifier holding Chapter 15's rules — one-sentence hints on request, attack after commitment, never the solution. Explaining a move you just used is the cheapest known way to find out whether you own it ✦.

Week 1 — Frames and deduction

The insight muscles of Chapter 4, then the propagation discipline of Chapter 5.

Week 2 — Numbers and chance

Chapter 6: conservation, recursion, digits under constraint, and the probability moves that punish good intuition.

Week 3 — The olympiad moves

Chapters 8–11. The format shifts: each day, re-derive a worked example cold — statement only, path from scratch — then run its drill. These are the sixteen tools' home gyms.

Week 4 — The boss room, and the mirror

Chapter 7's monsters, one per day — a single ★★★★ is a full session — braided with the calibration thread. The week ends by measuring you.

Why this order

The sequence is engineered, not decorative. Frames come first because every later week depends on noticing self-imposed constraints, and that noticing is fastest to install (a week of ★1–2 puzzles buys it). Deduction follows immediately so that the propagation discipline — no guessing while forced moves remain — hardens before the probability material arrives to punish sloppy conditioning. Numbers and chance sit in week 2 because they supply the regimen's first calibration data and its first honest humiliations; Monty Hall works better on a solver who has already kept a journal for eleven days. The olympiad moves occupy week 3 as deliberate cross-training: chapter-sized tools, re-derived rather than re-read, precisely when the first two weeks' moves risk settling into comfort. And the boss room is last not as a graduation ceremony but because ★★★★ puzzles are composites — 7·8 is an information budget wearing a deduction costume; 7·37 is probability riding structure — and composites are only trainable once their parts are reflexes. The mirror days (29–30) close the loop the journal opened on day 1: the regimen ends by measuring itself, which is the habit it most hopes to leave behind.

Maintenance mode

After day 30, two sessions a week hold everything the month built. Fewer, and the recognition signals fade to trivia; more is optional, not required.

Session A — new material (30 min). One of: a canon puzzle the regimen never scheduled (the bench: 6·22, 6·48 if either went unused, and the rest of Appendix C — C·23, C·24, C·36, C·45 among its seventeen); a Part III ★★★★ attempted for real under the struggle rule; or one problem from the outside world — the doors listed under Further Reading (a Gardner collection, Project Euler, an ARC grid) — given the full ritual: struggle, solve, name the move, journal the miss.

Session B — the deck (20 min). Every miss aged past two weeks, re-solved cold, re-timed. Monthly: one calibration set, and the journal audit — naming rate, curve, top miss class. If a class refuses to shrink, its chapter becomes Session A material until it does. The prescriptions:

The standing rule. Any puzzle met in the wild — interview, pub, timeline — gets the ritual. The regimen ends; the ritual doesn't.

Honest progress measures

Three numbers, none of which can be gamed without noticing.

1 · Misses named. The fraction of journal misses where the missed move got its name from Appendix A. Expect something near half in week 1 and something near all by week 4 — not because you miss less (the ladder should prevent that) but because the vocabulary has become perception. This number is the skill: recognition, measured at the moment of failure.

2 · The calibration curve. Your 80% intervals should capture the truth about 80% of the time — after fifteen or twenty intervals, a hit rate between 70 and 90 is honest work. Beware the failure that looks like success: intervals so wide they always hit train nothing. Precision under honesty is the game — intervals that narrow while the hit rate holds.

3 · Re-solve time on old misses. The deck's clock, falling. A miss that took thirty minutes and three hints re-solved in six minutes flat has entered the hands. A miss whose re-solve stalls at twenty minutes twice running is telling you which chapter to reread — believe it.

Expect the numbers to misbehave, and know what the misbehavior means. A plateau in naming rate around week 3 is normal: the olympiad tools temporarily flood the vocabulary, and perception lags acquisition by a few days. A falling calibration hit rate is usually good news wearing bad news's coat — it appears exactly when intervals start narrowing, and the fix is to narrow slightly less, not to stop narrowing. Regression on the deck — an old miss missed again — is the single most informative event the regimen produces: it marks a move that was memorized rather than owned, and it earns an immediate second journal entry with a different answer on line 4 than last time. If the same puzzle produces the same named miss twice, the name is wrong; the real gap is one level deeper, and a session staring at that gap is worth a week of new material.

And the non-measures, named so they can be refused: streak length, puzzles per day, hours logged, and how smart a session felt — the last being fluency, Chapter 15's documented liar ⚖. Thirty days buys no measurable rise in anything an IQ test totals, and this book will not pretend otherwise. It buys thirty-two moves at recognition speed, a journal that knows your failure modes better than any test, and a calibration baseline most professionals never acquire. That is not everything. It is more than almost anyone has.

The gate

Thirty days is not mastery. It is something better: working habits, and a vocabulary your eyes now run without being asked.

The Sphinx has been waiting at this book's gate since Chapter 1, and she has not aged. She was never really a monster on the road to Thebes. She is every gate with a question on it that punishes the literal reading: the interview problem built to be misread, the too-clean chart, the contract clause whose surface hides its instruction, the confident machine output that is fluent, cited, and wrong. Solve or be eaten — updated terms, same gate.

Her first riddle — four legs at morning, two at noon, three at evening — is, you may notice, still open. Sixteen chapters, and this book has not answered it. Books must close their boxes, and this one closes it on its final page, where the colophon keeps the answer. Go and look whenever you like. But mark what these thirty days changed: you no longer need the answer to get past her. You have the move. Answers fit one lock apiece; a move opens a corridor of them — which is why the index at the back is the only part of this book built to be reread for the rest of your life.

The manual ends. The gymnasium doesn't.