Fermi problems challenge us to estimate quantities with limited information, relying on logical reasoning and approximate data. These methods, pioneered by physicist Enrico Fermi, are invaluable in science, engineering, and everyday problem-solving.
This framework codifies the methodology from Lawrence Weinstein's Guesstimation books into a structured set of laws that enable effective human-AI collaboration on estimation problems. The mechanical laws (LAW0-LAW11) handle arithmetic and formatting. The estimation laws (LAW-INTERPRET, LAW-ESTIMATE, LAW-BOUNDS, LAW-PROCEED, LAW-DECOMPOSE, LAW-VALIDATE, LAW-REPORT) address the critical judgment layer: knowing when and how to guess, and when to ask for help.
Key insight: The hardest part of Fermi estimation isn't the math—it's having the confidence to make reasonable guesses and proceed with imperfect information.
Collaborative model: The human provides judgment (setting bounds, validating assumptions, sanity-checking results). The AI handles mechanics (applying laws, arithmetic, unit conversions, generating comparisons). When stuck or uncertain, the AI asks—it does not invent data or stall.
LAW0: Follow a structured five-step approach:
- Step 1: State your interpretation of the problem and confirm if ambiguous (LAW-INTERPRET).
- Step 2: Determine the known facts and confirm their accuracy.
- Step 3: Make reasonable assumptions for unknown information.
- Step 4: Perform calculations explicitly and transparently.
- Step 5: Report results in metric units, ensuring adherence to scientific notation (LAW1).
LAW1: Use scientific notation consistently. Round coefficients to the nearest integer and ensure exponents are non-negative by adjusting the units appropriately.
LAW2: Convert all numbers to scientific notation before performing operations.
LAW3: Calculate geometric means using the arithmetic averages of coefficients and exponents, applying adjustments for odd sums of exponents.
LAW4: To multiply numbers, multiply their coefficients and add their exponents.
LAW5: To divide numbers, divide their coefficients and subtract the exponents.
LAW6: To add or subtract numbers, align exponents first, then add or subtract coefficients.
LAW7: For square roots, adjust odd exponents by multiplying the coefficient by 10 and subtracting 1 from the exponent before calculating. Example: √10⁷ = √(10 × 10⁶) = 3 × 10³.
LAW8: Perform calculations in convenient units but report results in SI units.
LAW9: Avoid negative exponents by converting to smaller units, ensuring quantities remain expressed in positive powers of 10.
LAW10: Incorporate lessons from worked examples to improve problem-solving accuracy.
LAW11: Avoid unnecessary rephrasing of the problem; instead, work directly with the question at hand.
These laws address the judgment layer—knowing when and how to guess, and when to seek clarification.
LAW-INTERPRET: Before solving, state your interpretation of the problem.
Write one sentence summarizing what you will calculate, including:
- The specific quantity being estimated
- Key assumptions about the scenario
- Numerical values you've extracted from the problem statement
If the problem is ambiguous or could be read multiple ways, ASK for clarification before proceeding.
Triggers for asking:
- The problem statement contains vague terms ("fast," "large," "a lot")
- Multiple reasonable interpretations exist (e.g., "speed of sound" vs "speed of light" when both are mentioned)
- Critical context is missing (time frame, location, scale)
- The stated scenario seems physically implausible
Example application:
Problem: "The Flash runs at near the speed of light. What force is needed to reach 10% the speed of sound in 10 ms?"
Interpretation: "I will calculate the force to accelerate a ~100 kg person from rest to 34 m/s (10% of 343 m/s) in 0.01 seconds. Note: The problem mentions 'near the speed of light' but asks about 10% speed of sound—should I use a higher target velocity?"
LAW-ESTIMATE: For every quantity you use, you MUST do one of the following:
-
State as known — a fact that stands to reason without justification (e.g., speed of light ≈ 3×10⁸ m/s, human height ≈ 1.7 m, Earth's population ≈ 8×10⁹). These are anchors any technically literate person would accept without argument.
-
Bound with reasoning — construct explicit upper and lower limits from physical or logical arguments ("certainly more than X because..., less than Y because..."). Take the geometric mean as your estimate.
-
Ask — if you cannot do (1) or (2), stop and request human input before proceeding.
"I'll estimate X as Y" without justification is not permitted. If you find yourself reaching for a number you cannot justify by either method (1) or (2), that is a signal to ask, not to guess.
LAW-BOUNDS: Bounds should typically span 1-3 orders of magnitude.
- If bounds span more than 4 orders of magnitude, decompose further or try a different approach
- Good bounds are physically or experientially motivated, not arbitrary
- Behavioral observations make excellent bounds (e.g., "A gecko can walk on a ceiling but cannot run on one, so adhesive force is between 1× and 10× body weight")
- Example: "A person uses their phone more than 1 minute/day and less than 1000 minutes/day"
LAW-PROCEED: A factor-of-ten estimate is always preferable to no estimate.
- "Dare to be imprecise!" (Weinstein)
- An answer within a factor of 10 is good enough for most decisions
- If two approaches disagree by more than 10×, investigate why
- Always compare your final answer to something meaningful
LAW-DECOMPOSE: If stuck, ask "What would I need to know to answer this?" Each answer becomes a sub-problem.
After listing sub-quantities, explicitly classify each as:
- KNOWN — fact that stands to reason (no justification needed)
- BOUND — will construct upper/lower limits with reasoning
- ASK — cannot bound with available knowledge, need human input
For any quantity classified as ASK, stop and request input before proceeding. Do not invent data or stall.
The human provides judgment (validating bounds, sanity checks); the AI handles mechanics (arithmetic, unit conversion, applying laws).
LAW-VALIDATE: Cross-check estimates using independent approaches when possible.
- Use different decompositions or methods
- Compare to known reference values or real-world data
- If approaches converge, confidence increases
- If they diverge significantly, re-examine assumptions
LAW-REPORT: Final answers must be expressed in base SI units using scientific notation.
- Use meters, kilograms, seconds, joules (not km, mg, μs, kJ)
- Format: coefficient × 10^exponent (e.g., 3×10⁶ m, not 3000 km)
- If the question specified other units, provide the answer in both SI and requested units
- Include an uncertainty estimate when possible (e.g., "within a factor of 3")
- Always contextualize with a meaningful comparison
For each Fermi problem:
- INTERPRET — State what you're calculating; ASK if ambiguous
- DECOMPOSE — Break into sub-problems; classify each as KNOWN/BOUND/ASK
- ESTIMATE — For each sub-quantity, state as known OR bound with reasoning OR ask
- CALCULATE — Apply mechanical laws; show work explicitly
- VALIDATE — Cross-check if possible; flag inconsistencies
- REPORT — SI units, scientific notation, meaningful comparison
When available, consult reference tables for anchor values. Common anchors include:
| Scale | Value | Examples |
|---|---|---|
| 10⁻¹⁰ | Atom | Hydrogen atom diameter |
| 10⁻⁵ | Cell | Human cell diameter |
| 10⁰ | Human | Arm length, stride |
| 10⁷ | Earth | Earth diameter (1.3×10⁷ m) |
| 10¹¹ | AU | Earth-Sun distance (1.5×10¹¹ m) |
| 10¹⁶ | Light-year | 9.5×10¹⁵ m |
| Scale | Value | Examples |
|---|---|---|
| 10⁻³⁰ | Electron | 9×10⁻³¹ kg |
| 10⁻²⁷ | Proton | 1.7×10⁻²⁷ kg |
| 10⁻¹² | Cell | Human cell |
| 10⁻⁶ | Mosquito | ~2 mg |
| 10⁻² | Mouse | ~30 g |
| 10² | Human | 70-100 kg |
| 10³ | Car | Small car ~1000 kg |
| 10²⁴ | Earth | 6×10²⁴ kg |
| 10³⁰ | Sun | 2×10³⁰ kg |
| Scale | Value | Examples |
|---|---|---|
| 10⁻¹⁸ | Light across atom | |
| 10⁻³ | Millisecond | Camera shutter |
| 10⁰ | Second | Heartbeat |
| 10⁵ | Day | 8.6×10⁴ s |
| 10⁷ | Year | 3.15×10⁷ s (π×10⁷) |
| 10⁹ | Human lifespan | ~2.5×10⁹ s |
| 10¹⁷ | Age of universe | 4×10¹⁷ s |
| Scale | Value | Examples |
|---|---|---|
| 10⁰ | Walking | 1-2 m/s |
| 10¹ | Highway | 30 m/s (≈70 mph) |
| 10² | Sound | 343 m/s in air |
| 10⁴ | Earth orbit | 3×10⁴ m/s |
| 10⁸ | Light | 3×10⁸ m/s |
| Scale | Value | Examples |
|---|---|---|
| 10⁻¹⁹ | Photon | Visible light ~4×10⁻¹⁹ J |
| 10⁻¹⁹ | Chemical bond | ~1.5 eV = 2×10⁻¹⁹ J |
| 10³ | Food calorie | 4.2×10³ J per kcal |
| 10⁵ | Soda can | 6×10⁵ J (140 kcal) |
| 10⁸ | Gallon of gas | ~1.3×10⁸ J |
| 10¹⁵ | Nuclear bomb | 10¹⁵ J (kt range) |
| Material | Density |
|---|---|
| Air | 1.2 |
| Wood | 500-800 |
| Water | 1000 |
| Rock | 2500-3000 |
| Iron | 7900 |
| Lead | 11,300 |
| Constant | Value |
|---|---|
| Speed of light c | 3×10⁸ m/s |
| Gravitational constant G | 6.7×10⁻¹¹ N·m²/kg² |
| Planck's constant h | 6.6×10⁻³⁴ J·s |
| Avogadro's number | 6×10²³ /mol |
| Boltzmann constant k | 1.4×10⁻²³ J/K |
The following worked examples demonstrate the laws in action, with explicit narration of judgment calls.
Four complete examples are included below with permission. For the full set of 14 examples that informed this framework, see the Reference List at the end.
Problem: In the movie Spider-Man 2, Spider-Man stops a runaway New York City six-car subway train by attaching his webs to nearby buildings and pulling really hard for 10 or 20 city blocks. How much force does he have to exert to stop the subway train? Give your answer in newtons and in tons (1 ton = 10⁴ N). How does this compare to the force that you can exert?
Decomposition hints:
- What is the kinetic energy of a subway train at normal speed?
- What are its mass and velocity?
- How much work is needed to stop the train? (Remember that work = change in kinetic energy or W = ΔKE.)
- The stopping distance is about 1 km.
Solution:
Since the work done by Spider-Man to stop the train is equal to the train's initial kinetic energy, we need to estimate the mass and velocity of the train. We will then need to estimate the stopping distance in order to calculate the force exerted.
Estimating mass: A subway car is about the same size and weight as a semi-trailer (18-wheeler) truck. This is between 10 and 40 tons. We'll use 20 tons (or 2 × 10⁴ kg). There are six cars on a train so that the mass of the train is 6 × 2 × 10⁴ kg = 10⁵ kg.
Estimating velocity: They certainly go faster than 20 mph and slower than 100 mph. Since it is not that far between subway stops, subways travel at only about 40 mph (20 m/s).
Thus, the kinetic energy of a subway train is:
KE = ½mv² = 0.5 × 10⁵ kg × (20 m/s)² = 2 × 10⁷ J
Estimating stopping distance: There are 20 blocks per mile in Manhattan. Thus, 10 or 20 blocks is about 1 km or 10³ m. (It's certainly more than 100 m and less than 10 km.)
Thus, Spider-Man needs to exert a force:
F = KE/d = (2 × 10⁷ J)/(10³ m) = 2 × 10⁴ N
Converting to tons: F = (2 × 10⁴ N)/(10⁴ N/ton) = 2 tons
Comparison: A force of 2 × 10⁴ N is the weight of 2000 kg or 2 tons. For a superhero who can lift cars, this is quite possible (although definitely not easy). A human could definitely not do it.
Wow! Hollywood got the physics correct, in a superhero movie no less! Hurray!
Key techniques demonstrated:
- Comparison to familiar objects (subway car ≈ semi-trailer truck)
- Physical reasoning for bounds (not far between stops → can't reach highway speeds)
- Geographic knowledge (20 blocks per mile in Manhattan)
- Meaningful comparison at the end
Problem: What is the surface area of a typical bath towel (include the fibers!)? Compare this to the area of a room, a house, a football field.
Decomposition hints:
- Think about the little fibers in a really fluffy towel; how many are there in a square centimeter or square inch?
- What is the area of a large towel? How many total fibers does it have?
- What is the surface area of each little fiber?
- Consider the length and thickness of each fiber.
Solution:
That's obvious, surely! A large rectangular towel 1 m by 2 m has a total surface area of 4 m² (including both sides), right? (In US units, a big towel may be as large as 3 ft by 6 ft.)
Wrong, actually, unless it is a very worn-out towel. New towels have many little fibers that can absorb a lot of moisture (recall the old puzzle—what gets wetter the more it dries?). Unless you're a fan of the Hitchhiker's Guide to the Galaxy, you won't have brought your own towel, so nip off to the bathroom and examine one; quickly now, we're dripping all over the floor.
Estimating fiber density: You don't need to actually go and count the number of fibers per square inch or per square centimeter; in the latter case there must be more than 10 and fewer than 1000, so we take the geometric mean of 10¹ and 10³, which is 10². In a square inch, being about 6 cm², we should expect about six times as many. This will of course vary, depending on where you buy your towels; we are assuming that we are describing one of those very nice towels found in one of those very nice hotels.
Estimating fiber surface area: Back already? Right-oh. Now we need to estimate the surface area of each fiber. We can approximate the fiber as a cylinder or a box. Cylinders are complicated so we'll use boxes. Each fiber is about 0.5 cm (1/4 in.) long and 1 mm (0.1 cm) wide. Each "boxy" fiber then has four flat surfaces, each 0.5 cm by 0.1 cm. Thus, the surface area of one fiber is:
A_fiber = 4 × 0.5 cm × (1 m/10² cm) × 0.1 cm × (1 m/10² cm) = 2 × 10⁻⁵ m²
Total surface area:
A_total = towel area × fibers per area × area per fiber = 4 m² × (10² fibers/cm²) × (10⁴ cm²/1 m²) × (2 × 10⁻⁵ m²/fiber) = 80 m²
That is about 800 square feet: the size of a large apartment or a small house.
This problem is similar in some ways to calculating the length of the coastline. Just as the area of the towel is much greater than its simple area, the length of coast from, say, New York to Boston is much more than the 200-mile driving distance.
Key techniques demonstrated:
- Recognizing hidden complexity (the "obvious" answer is wrong)
- Geometric mean for fiber density (more than 10, less than 1000 → 10²)
- Simplifying geometry (cylinders → boxes)
- Meaningful comparison (≈ size of a large apartment)
- Conceptual analogy (coastline length problem)
Problem: While I was at a Norfolk Tides baseball game, a foul ball landed in the section above me and showered some of my friends with beer. What is the probability of a foul ball landing in a cup of beer during one baseball game? What is the expected number of "splash downs" during all the major league baseball games played in an entire season?
Decomposition hints:
- How many foul balls land in the stands each game?
- What is the size of a cup of beer?
- What fraction of people have beer cups?
Solution:
We need to break down the problem into manageable (or at least estimable) pieces. The first two pieces will be the number of foul balls per game that land in the stands and the probability that a given foul ball lands in a cup of beer.
Estimating foul balls per game: Let's start with the number of foul balls that land in the stands. The number per inning is definitely more than one and fewer than twenty, so we can take the geometric mean of five for our estimate. If you watch a lot of baseball, you can probably come up with a better estimate; I counted several (between three and seven) foul balls per inning landing in the stands. With nine innings per game, this amounts to forty foul balls per game that could possibly land in a cup of beer.
Breaking down the probability: Now we need to estimate the probability that a given foul ball will land directly in a cup of beer. (Note: only beer is sold in open-topped cups.) This means that we need to break the problem into even smaller pieces. Let's assume that the cup of beer is sitting innocently in a cup holder. To hit a cup of beer, the foul ball needs to:
- not be caught by a fan
- land within the area of a seat
- hit a seat whose owner has a cup of beer
- land in the cup
Most fly balls are caught, but many are not. Let's estimate that between one-quarter and one-half of fly balls are not caught. "Averaging" the two, we will use one-third.
Most of the stadium area is used for seating, so let's ignore that factor.
At any given time, more than 1% and less than 100% of fans have a cup of beer in front of them. Using the geometric mean, we estimate that 10% of seats have beer cups.
Calculating the area ratio: A large beer cup is 4 inches (10 cm) across, so the baseball must land in an area defined by
A_cup = πr² = 3(2 in)² = 10 in²
The area of the seat (from arm rest to arm rest and from row to row) is about 2 ft by 3 ft (60 cm by 90 cm), so
A_seat = (24 in) × (36 in) = 10³ in²
Thus, if the ball hits a seat that has a cup of beer, the probability that it lands in the cup is
P_cup = A_cup/A_seat = 10 in²/10³ in² = 10⁻²
or 1%. The metric probability is the same.
(Extra credit question: Which is more likely, that the ball lands in the cup in the cup holder, splashing the beer, or that the fan is holding the cup of beer when the foul ball arrives and splashes it in his or her excitement?)
Combining the probabilities: This means that the probability that any single foul ball lands in a cup of beer is
P = (1/3) × (1/10) × (10⁻²) = 3 × 10⁻⁴
With forty foul balls per game, this means that the probability of a foul landing in a cup of beer during any one game is 10⁻². This is not very likely. The probability that we will be directly below the splash is even less likely.
Scaling to the full season: During the entire season, each of the 30 teams plays 160 games, giving a total of about 2,000 games (as it takes two teams to play a game). This means that the total number of beer landings in one season is
B = (2 × 10³ games per season) × (10⁻² beer landings per game) = 20
Because baseball analysts keep meticulous statistics, I am very surprised that they do not appear to record beer landings.
Oh yes. The very improbable detail? According to my friends, the beer belonged to our former governor! ("Now at an improbability factor of a million to one against and falling," D. Adams.)
Key techniques demonstrated:
- Multiplicative probability chain with independent factors
- Breaking "what's the chance of X" into sub-conditions
- Area ratio reasoning (cup area / seat area)
- Geometric mean for uncertain fractions (% of fans with beer)
- Scaling from per-event to per-season
- Reality emerges from combining many small probabilities
Problem: How much radiation damage would we receive traveling to the center of the galaxy?
ANSWER: In order to estimate our radiation damage from space travel, we need to estimate the number of atoms in our path and the damage done to us by each atom.
Estimating the distance: In order to estimate the number of atoms in our path, we need to estimate the distance traveled and the density of atoms. We might remember that the distance is about 10 kiloparsecs, or about 3 × 10⁴ light-years. Alternatively, we might know that there are about 10¹¹ stars in our galaxy and it is about 4 light-years (ly) to the nearest star. In that case, each star occupies a volume of about
V = (4 ly)³ = 200 ly³.
so the total volume of the galaxy is V = 10¹³ ly³. The galaxy is relatively flat, so it could have a radius of 10⁵ ly and a thickness of 10³ ly. That would place the solar system about d = 5 × 10⁴ ly from the center.
Converting to meters: This means that the distance traveled is
d = 3 × 10⁴ ly = (3 × 10⁴)(3 × 10⁸ m/s)(π × 10⁷ s) = 3 × 10²⁰ m.
Counting atoms on our path: The typical density of interstellar space is one atom per cubic centimeter, primarily hydrogen. By comparison, room air has a density of n = (6 × 10²³ atoms)/(2 × 10⁴ cm³) = 3 × 10¹⁹ atoms/cm³. If we consider a longitudinal slice of the spacecraft with surface area A = 1 cm², it will encounter 3 × 10²² atoms on its journey.
Why we need relativistic speed: In order to travel this distance within a subjective human lifetime, our spacecraft must achieve a speed very close to the speed of light. Even traveling at the speed of light, if there is no relativistic time dilation, our trip will take 3 × 10⁴ years, which is slightly longer than our expected lifetime. Fortunately, at speeds close to the speed of light, subjective time passes more slowly than it does for an observer watching from the center of the galaxy. We need time to pass about 10³ times more slowly. In technical terms, this means we need a relativistic gamma factor of about γ = 10³.
Energy of each collision: This means that, from our point of view (i.e., in our reference frame) each interstellar atom will hit us traveling at almost the speed of light with the same gamma factor and will thus have a total energy of
E = γmc² = 10³ × (1 GeV) = 1 TeV.
At these energies, when the hydrogen atom strikes the spacecraft, it will lose its electron very quickly. The bare proton will then pass easily through the spacecraft and our bodies, depositing energy in our bodies at a rate of 2 MeV per centimeter.
Calculating energy deposition: Now we can consider the damage done by all these protons. Consider a volume of 1 cm³ within our bodies. At the density of water it has a mass of 1 g. It will be hit by 3 × 10²² protons, with each proton depositing 2 MeV of energy. This means that the total energy deposited per gram will be
E_dep = (3 × 10²² protons/cm²)(2 × 10⁶ eV/cm) × (1 cm³/g) = (6 × 10²⁸ eV/g)(1.6 × 10⁻¹⁹ J/eV) = 10¹⁰ J/g = 10¹³ J/kg.
Comparing to lethal doses: This seems like a lot, but let's compare it with a few things before jumping to conclusions.
Unfortunately, we would not even survive that long. Radiation dose is measured in grays, where 1 Gy corresponds to an energy deposition of 1 J/kg. A lethal dose of radiation is about 10 Gy or about 10 J/kg. We will accumulate that dose in 10⁻¹² of our journey, or in the first millisecond.
Thermal perspective: As it only takes 2 × 10⁶ J/kg to boil water, we would be vaporized more than 10⁶ times over. In a thirty-year journey lasting 10⁹ s, we would be turned to steam within the first 10³ s, or 20 minutes.
Who knew that vacuum could be so dangerous? I don't know about you, but I'm staying home.
*These protons can pass through about 1 km of metal before stopping. We can't stop them with mass shielding.
Key techniques demonstrated:
- Multi-domain synthesis (astronomy, relativity, nuclear physics, radiation biology)
- Converting between unit systems (light-years to meters)
- Relativistic physics integration (gamma factor for time dilation and energy)
- Chain of energy calculations leading to biological impact
- Multiple comparison metrics (lethal dose, vaporization, time to death)
- Reality check reveals impossibility—interstellar travel at relativistic speeds through normal matter is inherently fatal
The following problems from Weinstein's books were used to develop and test this framework. Full solutions are available in the source texts.
| # | Title | Topic | Source |
|---|---|---|---|
| 3 | Life on the Phone | Fraction of people on phones right now | Guesstimation 2.0, Ch. 1 |
| 4 | Flattening the Alps | Potential energy stored in a mountain range | Guesstimation, Ch. 4 |
| 5 | Tilting at Windmills | Power output of a wind turbine | Guesstimation, Ch. 5 |
| 6 | Supersize the Sun | Estimating the Sun's radius using your finger | Guesstimation, Ch. 3 |
| 7 | How High Can We Jump? | Largest asteroid you can jump off | Guesstimation, Ch. 6 |
| 8 | Water Bottle Recycling | How far to walk to recycle a bottle | Guesstimation 2.0, Ch. 7 |
| 9 | More Numerous Than the Stars | Number of cells in the human body | Guesstimation, Ch. 2 |
| 10 | Let's Get One Thing Straight! | Total length of DNA in your body | Guesstimation, Ch. 2 |
| 11 | Oh Say Can You Hear? | Angular resolution of human hearing | Guesstimation, Ch. 8 |
| 12 | A Mole of Cats | Mass and volume of 6×10²³ cats | Guesstimation, Ch. 3 |
Key techniques demonstrated across all examples:
- Geometric mean for bounded estimates
- Comparison to familiar objects
- Multiple independent approaches to cross-validate
- Sensitivity analysis for critical parameters
- Meaningful comparisons to everyday or cosmic scales
- Humor as a pedagogical tool
Fermi problems illustrate the art of estimation, turning vague questions into solvable puzzles using logical approximations. The key principles:
- State your interpretation — Make sure you're solving the right problem (LAW-INTERPRET)
- Dare to be imprecise — A factor-of-ten estimate is always better than no estimate
- Break it down — Decompose until pieces are small enough to estimate
- Bound and guess — Establish upper and lower limits, take the geometric mean
- Cross-check — Use independent approaches when possible
- Compare meaningfully — Put your answer in context
- Ask when stuck — If you cannot bound a quantity or the problem is ambiguous, request human input
By following the laws and studying these examples, you develop the judgment to know when to guess, how to bound, and when your answer is "good enough."
Source material: All worked examples are adapted from Lawrence Weinstein and John A. Adam, Guesstimation: Solving the World's Problems on the Back of a Cocktail Napkin (Princeton University Press, 2008) and Lawrence Weinstein, Guesstimation 2.0: Solving Today's Problems on the Back of a Napkin (Princeton University Press, 2012). Examples 1, 2, 3, and 4 in this document are reproduced with permission under license from the Copyright Clearance Center. The problems, solutions, hints, and explanatory text are Weinstein's work.
Our contribution: The codification of Weinstein's methodology into explicit numbered laws (LAW0-LAW11, LAW-INTERPRET through LAW-REPORT) suitable for AI collaboration, developed through iterative work with GPT-3.0 and refined with Claude. The estimation laws address the judgment layer—the critical insight that AI models need explicit permission to guess and proceed with imperfect information, and explicit instruction to ask when they cannot bound a quantity or when the problem statement is ambiguous.
Recommendation: Buy Weinstein's books. They contain dozens more excellent problems and are a pleasure to read.
The Laws (LAW0-LAW11, LAW-INTERPRET through LAW-REPORT) and the framework structure are released under CC BY 4.0.
The four complete worked examples (Spider-Man, Towel, Baseball/Beer, Galaxy) are reproduced with permission from Princeton University Press and are subject to their copyright.