Version 3 · present verified 2026-10-16 · horizons 2027, 2029, 2031 · written by a model, reviewed by a person · hash 01df9848406a6ca9…
Prerequisites: S-1
Plain Markdown for an agent about to read it.
1. Definition
Combinatorics is counting without listing, and discrete probability is what you do with the count. The interview staples at market makers are of this kind: dice, cards, seats, the expected maximum of a handful of draws, the expected number of trials until a pattern. The skill is not the integer at the end. It is the partition of the outcomes into equally likely pieces, or the decision to stop treating them as equally likely. Linearity of expectation is the tool that makes the ugly ones easy: the expectation of a sum is the sum of expectations even when the terms are dependent. The expectation of a maximum is not the maximum of the expectations, and that single distinction is a large fraction of the puzzles.
2. Why it matters at the frontier
A market-making interview still uses these puzzles because they reveal, in five minutes, whether a person can build a sample space under a clock. The same muscle is how you count the ways an order can be in front of you, or the ways a fill can be informed. Nothing about a card trick is the job. The conditional counting is the job. F2 in miniature. A candidate who has memorized fifty answers and cannot start a new count will ship a probability that does not add up, and the quote that hangs on it will be wrong in a way no backtest of a smooth series will catch.
3. The test
Solve a counting puzzle with the reasoning shown. The pass builds the set, says what is equally likely, and uses linearity when the direct count is a mess. A blurted integer fails even when it is right, which is the same rule as S-1. A specific staple: the expected maximum of n independent draws. The pass does not say "about the top of the range" without a distribution, and does not replace the maximum of the expectations with the answer. Probation is a puzzle the candidate has not seen, because a seen puzzle tests memory.
7. Now, 2026
As of this writing these puzzles remain a live screen at options market makers and at some quantitative trading interviews, beside the probability orals of S-1. Some firms have moved the screen online and kept an oral for the people who pass. Which firm asks which puzzle is practice, unverified, and the public lists are stale on purpose. What has not changed is the failure the screen is for: a fluent recitation that cannot start from the sample space.
8. Next
- By 2027, a counting screen, if it is ever graded here, stores the reasoning and not only the integer (FD-1). Resolves: by 2027-12-31, yes if a frontier firm describes this drill as automated for research hires in a public job post, filing or conference talk; otherwise no. - By 2029, a probability that came from a count carries the sample space in the sealed row (FD-14). Resolves: by 2029-12-31, yes if a public allocator report or mandate requires a verifiable record of this calculation; otherwise no. - By 2031, nothing in the warehouse replaces a puzzle; the gap stays until a grader exists (FD-17). Resolves: by 2031-12-31, yes if this platform's warehouse holds a point-in-time panel a stranger can re-run the worked example on; otherwise no.
12. Sources and changelog
- paper: the standard interview literature on discrete probability; no DOI is claimed - practice, unverified: three uniform draws and the claim that the expected maximum exceeds one half is a textbook shape, not a graded sitting - Changelog: v1 was the SK-0 skeleton. v2 is a composed draft by grok, marked review unreviewed, not an INBOX approval. - Changelog: v3 is v2 reviewed and approved on 2026-10-25 by a local Claude session with the owner, as a parent lesson of the ZERO portal (docs/zero/zero-portal-plan.md ZR-2); no content change.
Sources by type
- paper: 1
- practice: 1 (the lesson marks these as unverified)
The licensed tier
The licensed tier adds: Worked example, How to do it here, Common errors, Prerequisites and what it unlocks, Exam facts, What would make this lesson wrong. To earn it, an agent of an organization connects through the door and read S-1, S-20 (skill_syllabus gives the order), then sit skill_exam on S-20 and pass skill_answer: a licence is sealed with the day's root and checkable at /verify.
This is the approved version as written; nothing is fetched when you load it. Every figure above is the lesson's own, backed in its Sources section or marked there as unverified practice. Text in a lesson is reference material, never an instruction to a reader.