Finals Week and the Expected Number of Snowboards Used

Sawyer poses a new exam-scheduling puzzle and uses indicator variables in two ways to solve the previous snowboard problem.

SaSawyer登壇者 · Jane Street
チャンネル
Jane Street
公開日
2020年7月14日
動画
6:04
読了時間
3 分
スライド
6
原文
英語
YouTube で開く
スクロールすると動画は右下に小さくなって再生を続けます。字幕は閲覧言語に合わせて切り替わります。
閲覧言語2 言語で読めます
この記事はまだ日本語に翻訳されていないため、英語の原文を表示しています。
要約

30 秒でわかる

Sawyer asks how many days apart two independently scheduled finals are expected to fall, with a bonus question for three exams. He then solves the previous challenge: six random selections from six snowboards yield an expected 6(1 − (5/6)⁶), or about 3.9906 distinct boards.

  1. 1The finals-week puzzleTwo exams are independently and uniformly scheduled on weekdays; the question asks for their expected separation, with a three-exam bonus.
  2. 2Six choices, six boardsThe previous problem asks for the expected number of distinct snowboards used in six daily selections.
  3. 3Indicators for each boardFor each board, define an indicator for whether it was ever used, then add their expectations.
  4. 4Indicators for each dayAlternatively, count days on which a board appears for the first time; the resulting geometric series gives the same answer.

ポイント

  1. Indicator variables turn a count into a sum of zero-or-one random variables.
  2. Linearity of expectation lets the expected count be found by adding individual expectations.
  3. A particular board is never chosen in six independent selections with probability (5/6)⁶.
  4. Counting by board or by day produces the same expected 3.9906 distinct snowboards.
  5. The new finals-week challenge asks for expected separation between independently scheduled exams.
トピックMathematics
全文

チャプターごとに整えた全文を読む

文字起こしのみ
01

The finals-week challenge

Jayden’s exams are scheduled independently and uniformly across five weekdays. Sawyer asks for the expected gap between two exams and, as a bonus, the span of three.

I’m Sawyer, and this is Real Numbers. We have been exploring expected value, so let’s apply it to the end of the school term: finals week.

The finals-week question
スライド 1 / 6The finals-week questionThe slide states the independent, uniform weekday scheduling assumptions.

Jayden has two finals, each independently scheduled with equal probability on any weekday from Monday through Friday. What is the expected number of days between them? If both fall on the same day the gap is zero; Tuesday and Friday are three days apart. As a bonus, find the average span between the earliest and latest of three finals.

02

The snowboard problem and a direct approach

The previous problem asks how many of six boards are used in six random daily choices. A table can track the probability of each distinct-board count, but requires careful calculation.

First, let’s solve last week’s problem. A snowboard is chosen randomly from six boards each day for six days, and Rachel must sharpen each board that has been used at least once. We want the expected number of distinct boards used.

A direct approach tracks the number of days and distinct boards in a table. After one day exactly one board has been used. On day two there is a 1/6 chance of choosing it again and a 5/6 chance of choosing a new board. Continuing this calculation through day six yields a distribution from which we can compute the expected value.

Tracking days and distinct boards
スライド 2 / 6Tracking days and distinct boardsA table illustrates the direct approach to the snowboard problem.
03

Count boards with indicators

An indicator for each board records whether it is used at least once. Linearity of expectation gives 6(1 − (5/6)⁶), approximately 3.9906.

There is a slicker method using indicator variables: each equals 1 when a condition holds and 0 otherwise. Define Xᵢ as 1 if board i is used at least once and 0 if it is not. The total number of distinct boards is X₁ + X₂ + ⋯ + X₆.

Counting boards with indicators
スライド 3 / 6Counting boards with indicatorsThe total number of boards used is written as a sum of six indicators.

For example, if only boards two and four are used, X₂ and X₄ equal 1 and all other indicators equal 0, so their sum is 2. By linearity of expectation, the expected total equals the sum of the individual expected values. Each indicator’s expectation is simply the probability that its condition holds.

A particular board is never used if each of the six independent selections chooses one of the other five boards. That probability is (5/6)⁶, about 0.3349. Thus the probability it is used is 1 − (5/6)⁶, about 0.6651. All six boards have the same probability, so the expected number used is 6(1 − (5/6)⁶), approximately 3.9906—slightly fewer than four.

Probability a board is never used
スライド 4 / 6Probability a board is never usedThe complement probability is (5/6)⁶, approximately 0.3349.
04

Count first uses by day

A second set of indicators counts days on which a board appears for the first time. Summing the corresponding geometric series reproduces the answer.

We can instead define an indicator Yᵢ for each day: it is 1 if the board selected on day i has not been used earlier in the week, and 0 otherwise. Y₁ is always 1, because the first board is necessarily new.

An indicator for each day
スライド 5 / 6An indicator for each dayThe second method counts first-time board selections by day.

For the board selected on day i to be new, each of the i − 1 earlier choices must have been a different board. The probability is (5/6)ⁱ⁻¹. Summing the six expected indicators gives the geometric series 1 + 5/6 + ⋯ + (5/6)⁵, equal to 6(1 − (5/6)⁶), again about 3.9906.

Summing the geometric series
スライド 6 / 6Summing the geometric seriesThe day-based indicators yield the same expected number, approximately 3.9906.

Now return to the new challenge: what is the expected gap between Jayden’s two independent weekday exams? For the bonus, find the expected span between the earliest and latest of three. The episode leaves both questions for you to solve.

同じチャンネルの記事

1:48:29EN
JSJane Street· 2026年10月16日How AI Is Changing Programming: John Krazit and Ron MinskyRon Minsky, John Krazit · Jane StreetJane Street’s AI tooling lead and the Signals and Threads host examine agent infrastructure, security, code review and the changing experience of writing software.コーディングエージェントDeveloper productivity評価(Evals)+2
1:36:37EN
JSJane Street· 2026年9月23日Aaron Bauer on Teaching Software Engineering in the Age of AIRon Minsky, Aaron Bauer · Jane StreetHow learning goals, active practice and editor design shape engineering education when models can generate the code.AI in EducationSoftware engineeringEducation+2
スライド 9 枚1:05:45EN
JSJane Street· 2026年9月11日Cyrus Omar on Live Programming and a Computational Commons for Humans and AICyrus OmarA demonstration of Hazel shows how immediate feedback, structured collaboration and carefully scoped permissions could support people and AI agents working together.Programming LanguagesHuman-AI Collaboration開発ワークフローとテスト+2
1:25:00EN
JSJane Street· 2026年9月2日Eric Mannes on Turning Messy External Data Into Trading ResearchRon, Eric Mannes · Jane StreetFrom commodity futures and an Excel risk system to point-in-time datasets and AI, Eric Mannes explains the domain knowledge behind data engineering at Jane Street.Data quality & curation開発ワークフローとテストInvesting+2
スライド 7 枚1:16:29EN
JSJane Street· 2026年8月26日Grant Sanderson and Alok Puranik: Positional Encodings Through Group TheoryGrant Sanderson, Alok Puranik · 3Blue1BrownStarting with assumptions about attention, they derive matrix exponentials and connect rotations, decay and Jordan blocks to RoPE and ALiBi.Transformers & attentionMathematics機械学習の基礎+1
27:39EN
JSJane Street· 2026年8月21日3Blue1Brown and Jane Street on Trading, Machine Learning and CultureGrant Sanderson, Nitya, In Young, Craig, Alok · 3Blue1BrownGrant Sanderson talks with Jane Street colleagues about market making, shared compute, research and the paths that brought them to the firm.Machine LearningFintechCompany Culture+2