2026

Chapters

• Chapter 1. Two Conversations · what a year of applications has produced so far; no count claimed
• Chapter 2. FP&A for HVAC · an FP&A work sample, second in a series of one industry at a time; sourced industry figures and a calculator that starts empty
• Chapter 3. The Computer Room · a cooling failure at a bank branch, told from the customer’s side; an account, not a finding
• Chapter 4. A Policy Brief on the Response Gap · a credit-governance working paper read as a policy analyst would: four proposed controls, each with a test, and one run
• Chapter 5. A Proposal With a Test That Can Fail · one decision rule from a research proposal, written before any data and made runnable, with the project’s record of what failed
• Chapter 6. What a Collaborator Found · what a catalysis researcher studied, what one exploratory pass over her data showed, and what her review changed
• Chapter 7. Two Postings, One Employer? · two recruiting-coordinator postings read side by side: what they share, where they differ, and why the record cannot yet call them one role
• Chapter 8. A Preprint Against a Kernel · a 247-page theory that cites five Lean-verified theorems: what the five say, a checklist of what is checked and what is not, and what a full check would cost
• Chapter 9. A Licence Before a Student · what New Jersey requires before a one-person company may sell instruction: the statutory test, the three exclusions, and the fees payable before a first student exists
• Chapter 10. Four Pillars, One Mission · a vision for a social justice institute’s four program areas, tied to its mission, with the strongest case against each; names the institute, as an exception, because it is written for a live application
• Chapter 11. The Books They Manage · what a state workers’ compensation insurer’s own annual reports show about the bond book behind one posting: the 2025 bond table, the policy rules, and seven years of realized gains and losses; names the insurer, as chapter 10 names its institute
• Chapter 12. From Problem to Plan · what a Newark nonprofit’s own Forms 990 show about the finance work behind one posting, read as problems found, analysis, recommendations, actions and follow-through; names the institute, as chapters 10 and 11 name their organizations
• Chapter 13. The Recommendation Record · a work sample for a Principal Investigator posting on FINRA’s Recommendations team: one month of published disciplinary actions read against Regulation Best Interest and Rule 2111, with a worked test on a synthetic account; names FINRA and two firms, as chapters 10 to 12 name their organizations; not investigative experience
• Chapter 14. Options From First Principles · a work sample for a Quantitative Trader posting on an index-volatility desk: Black–Scholes price, Greeks and implied volatility derived and checked on synthetic quotes
• Chapter 15. A Fair Price for a Resource That Is Not Free · resource rent, a levy on revenue against a share of rent, and who pays for a data dividend, worked on invented firms and checked by script
• Chapter 16. When Bail Is Not Posted · a descriptive audit of New York pretrial data on bail posting and early plea dispositions, the $1 bail code and a judge check that fails.
• Chapter 17. The Credit Behind the Annuity · a work sample for an insurance domain analyst posting at EY: what the NAIC’s year-end 2025 data show about U.S. insurers’ private credit, the bounds they allow and seven questions for a client risk team; names EY and the NAIC, as chapters 10 to 13 name their organizations; claims no insurance experience

What this record is

Each row is one application, or one set of application materials, grouped by the industry of the employer and the type of role. Employers are left out on purpose. The point of the record is the shape of a search: how wide it was, how the roles moved between fields, and how many there were. Naming a company would add risk for the people who work there and nothing to that picture.

2026 applications

What this record does not show

• It does not show how many applications were rejected, ignored or still open. Outcomes are recorded only where the employer has written back: no outcome is recorded in this table so far, and a blank is not a rejection.
• It leaves out applications made through job boards with a single click, so the true total is higher than the number of rows.
• It does not explain any outcome. A search of this width, across this many fields, in this labor market, cannot be read as a judgment on the person applying.

How it gets completed

Each row needs three confirmed facts: whether the application was sent, the date it was sent, and what happened. Until those are filled in, the table above stays marked as unconfirmed. Applications from 2027, if the search continues, go in Book XXVII on the same terms.