Full lesson previewThe Rating Formula: How Chess Ratings Predict Who Wins
Math60 minutesGrades: 12th
Overview
Use logarithms to see how chess ratings predict a winner.
Description
Your senior learns the Elo expected-score formula E = 1/(1 + 10^(-D/400)), where D is the rating gap, graphs the exponential piece 10^(D/400) and the S-shaped curve it creates, and uses logarithms to show that a gap of about 191 points gives a 75 percent expected score. They then play three quick games with family members starting from sample ratings of 1200, update everyone's numbers after each game with K = 32, and discuss what a single number can and can't say about a player.
Objective
Your teen will graph exponential and logarithmic functions with intercepts and end behavior, solve 10^(-D/400) = 1/3 with a logarithm, and apply the Elo update rule to real games.
Standards
F-IF.7eMath · Grade HS · High School: Functions — Interpreting Functions Graph exponential and logarithmic functions, showing intercepts and end behavior, and trigonometric functions, showing period, midline, and amplitude.
F-LE.4Math · Grade HS · High School: Functions — Linear, Quadratic & Exponential Models For exponential models, express as a logarithm the solution to abct = d where a, c, and d are numbers and the base b is 2, 10, or e; evaluate the logarithm using technology.
Materials
| Item | Quantity |
|---|
| Chess set | 1 |
| Graphing calculator or a free graphing app on a phone or computer | 1 |
| Notebook | 1 |
| Pencil | 1 |
| Phone timer (for 5-minute games) | 1 |
Vocabulary
Elo rating
A number that estimates a player's strength, named after physics professor Arpad Elo, whose system the U.S. Chess Federation adopted in 1960 and the international chess federation (FIDE) in 1970.
Expected score
The average points a player is predicted to earn per game against an opponent, counting a win as 1, a draw as 0.5 and a loss as 0.
Logarithm
The exponent you need: log base 10 of 3 is the power you raise 10 to in order to get 3 (about 0.4771).
K-factor
The multiplier in the update rule that controls how far one game can move a rating.
Logistic curve
An S-shaped graph that rises slowly, then quickly, then levels off between two limits, here 0 and 1.
Preparation
- Set up the chess board and open a graphing app or calculator.
- Write "Teen 1200" and "Parent 1200" (and a third family member if playing) at the top of a notebook page; these are sample starting ratings for practice.
- Read the Formula Sheet, Practice Problems and Answer Key in Additional Sections.
Introduction
Tell your teen: a 1500-rated player sits down against a 1700-rated player. The rating system already 'knows' how often the lower-rated player should win. Ask them to guess the percentage, and then promise that by the end they'll be able to calculate it exactly.
If two players have the same rating, what should each one's chance be?
Should beating a much stronger player move your rating more than beating an equal one? Why?
Steps
1Step 1
Meet the formula
(10 min) Your teen copies the Formula Sheet into the notebook and uses the graphing calculator to fill in a table of expected scores for gaps of -400, -200, 0, 100, 200 and 400. They write one sentence describing the pattern, such as 'every 400 points multiplies the odds by 10'.
2Step 2
Graph the exponential and the S-curve
(12 min) On the graphing calculator your teen graphs y = 10^(x/400) and marks its y-intercept (0, 1) and its end behavior (it approaches 0 to the left and grows without bound to the right). They then graph y = 1/(1 + 10^(-x/400)), sketch both in the notebook with labeled asymptotes y = 0 and y = 1, and explain how the exponential piece bends the S-curve.
3Step 3
Solve for a 75 percent edge
(10 min) In the notebook your teen sets 1/(1 + 10^(-D/400)) = 0.75, simplifies to 10^(-D/400) = 1/3, and solves D = 400 x log(3), which is about 191 points. They graph the logarithmic form D = 400 log(E/(1 - E)), find its intercept at E = 0.5, and describe what happens as E gets close to 0 or 1.
4Step 4
Play and update
(18 min) Your teen plays three 5-minute games on the chess set against family members, timed with the phone timer. After each game they use R_new = R_old + 32(S - E) in the notebook to update both players' ratings, showing every step, and check that the two players' changes add to zero.
5Step 5
Test the model
(10 min) Your teen finishes Practice Problems 1 to 8 in the notebook and checks the Answer Key with you. Together you discuss what the model can't capture, such as a player having a bad day, improving fast, or being strong at quick games but not long ones, and your teen writes two limitations in their own words.
Conclusion
Ask your teen to explain, without notes, why the rating gap for a 75 percent edge needs a logarithm to find. Then ask whether they trust a 191-point gap to predict their next family game, and why or why not.
Make your own lesson plan in minutes
Describe a topic and Resource Portal AI writes the objectives, materials and steps for your group.
Try Resource Portal AI