16 famous candies, seeded by how often they won in FiveThirtyEight’s poll of 269,000 matchups. Pick a winner in each matchup until one champion is left.
Cocoa got 4× pricier. So some candy quietly got less chocolate.
Cocoa, world priceUS candy and gum prices
1 · The price shockBad weather and plant disease hit Ghana and Côte d’Ivoire, which grow most of the world’s cocoa. By 2024, cocoa jumped from about $2,000–3,000 a ton to $12,000.
2 · What candy makers didSome swapped cocoa for cheaper sugar, oil, milk and nuts. Reese’s new Valentine’s Mini Hearts had less than 2% cocoa and were sold as “chocolate candy,” not milk chocolate.
3 · Read the wrapper“Milk chocolate” means real chocolate. “Chocolate candy” or “chocolate flavored” can mean very little cocoa. Ingredients are listed biggest first.
Market idea: when one ingredient gets expensive, makers look for substitutes. Shoppers who read labels can push back by choosing what they buy.
7 of the 10 most-loved candies were invented before 1950.
Reese’s Cups (1928), Snickers (1930), Kit Kat (1935) and M&M’s (1941) are still America’s top sellers. Here’s every candy brand in our list by the decade it launched, with that decade’s most-loved candy.
Launch years from each brand’s Wikipedia article. Most-loved = highest share of head-to-head wins in FiveThirtyEight’s poll. Top sellers: Instacart.
Trend · startups vs. giants
New “better-for-you” candy brands grow fast, and then the giants buy them.
2020SmartSweets (low-sugar gummies, started 2016) sells a majority stake to investor TPG, reportedly for $360M.
2021Hu (simple-ingredient chocolate, started 2012) is bought by Mondelēz, maker of Oreo.
2021Lily’s (low-sugar chocolate) is bought by Hershey for about $422M.
2022Feastables, MrBeast’s candy brand, launches and lands in big stores fast.
Market idea: big companies can buy what’s growing instead of inventing it. That’s one way startups pay off.
A blogger predicted it would take 400 to 500 packs. Pack 464 matched.
Possibly Wrong · Experiment
Try this
This is the birthday problem in a candy wrapper. How many people do you need in a room before two probably share a birthday? Guess first, then look it up (it’s 23).
A math blogger predicted it would take roughly 400 to 500 packs to find two with the exact same colors, then spent 82 days counting 27,740 Skittles to check. The match showed up right on schedule.
Squashed spheres fill more of a jar than balls do. Physicists proved it with candy.
Physics World · Science study
Try this
Guessing the candy jar? Find the jar’s volume, multiply by 0.7, and divide by the volume of one M&M. Then compare with your plain guess.
Physicists poured M&M’s into containers and found they fill 68 to 71% of the space at random. Round balls only fill about 64%. Being a little squashed lets each M&M touch more neighbors. They published it in Science in 2004.
By Paul Chaikin and colleagues (Princeton & Cornell). Open it ↗
A statistician’s colors didn’t match the official mix. Then Mars admitted it has two factories.
R-bloggers · Statistics
Try this
Sort a few fun-size bags by color. Does your mix look like one factory, the other, or neither? How many bags would you need to be sure?
A statistician scooped M&M’s from the office break room for weeks. His colors didn’t match the mix Mars had published, and a chi-square test said that probably wasn’t luck. Then Mars explained: two factories make M&M’s, each with its own color mix.
By Rick Wicklin (SAS), re-run in R by Chuck Powell. Open it ↗
A machine said 364. Kids said 252. Another machine said 411. Who’s right?
Tootsie · Experiments
Try this
Why don’t they agree? Write your own rule for what counts as one lick, then test it on 3 pops. Is your answer a single number or a range?
Purdue engineers built a licking machine: 364 licks. Their 20 human volunteers: 252. A University of Michigan machine: 411. Junior high students at Swarthmore: 144. Same candy, four answers.
800 people guessed. Most were wrong. Their middle guess was off by 9 pounds.
Wikipedia · Classic study
Try this
At a party, have everyone guess the candy jar without talking. Line the guesses up and take the middle one. Did the crowd beat the best single guesser?
In 1906, about 800 people at an English fair guessed the weight of an ox. Most were off, but the middle guess (1,207 pounds) was within 1% of the real weight (1,198). It’s the reason candy-jar contests are secretly a math lesson.
Growers turned 1,203 pumpkins into a chart that weighs with a tape measure.
BigPumpkins.com · Data tables
Try this
Measure around your own pumpkin, then weigh it. Do it with 5 pumpkins and plot circumference against weight. Is it a straight line? (Hint: think about volume.)
Giant pumpkin growers can’t put a 1,500-pound pumpkin on a bathroom scale, so they built tables from 1,203 real pumpkins to turn tape-measure readings into a weight. They also tell you how often the estimate misses: about 76% land within 10%.
Every Halloween, the Census counts the doors. And the kids. And the spooky towns.
Census · Facts for Features · Government data · free data
Try this
72.0 million kids and 132.7 million homes: about how many kids per home? Why might your street feel very different from that average?
Every year the Census Bureau posts Halloween facts: how many homes there are (potential stops), how many kids (72.0 million), and towns with spooky names like Sleepy Hollow, Illinois.
One kid cornered the Reese’s market by trading first. Charleston Chews never moved.
NPR · The Salt · Reporting
Try this
Spot the market ideas: who used supply, who used timing, which candy acted like pennies? Then try one rule at your own trading floor.
A reporter watched 25 kids, most about 11, trade after trick-or-treating. One cornered the Reese’s market by trading early. One worked hard to unload extra candy from her own house. Charleston Chews barely moved at all.
The price stayed 5 cents for 69 years. What changed was how long you worked for it.
Human Progress · Analysis
Try this
Find your own time price: how many minutes of chores or allowance pays for one full-size bar? Is candy cheap or expensive for you?
A Hershey bar cost 5 cents from 1900 until 1969. That sounds cheap until you divide by wages: an unskilled worker earned about 9 cents an hour in 1900. By 2022 the bar cost $1.15, but only about 4.5 minutes of work.
Two sides, two sets of real numbers, one factory move. Decide who’s right.
NPR · Potato Grower · Two sides
Try this
Both sides used real numbers. What would you need to know to decide who’s right? Who gains from each story?
In 2002 the maker of Life Savers announced it was moving production from Michigan to Canada and blamed high U.S. sugar prices. The sugar growers’ group did its own math: at most $4.8 million a year saved on sugar, but about $10.2 million on wages.
By Jackie Northam (NPR) and the American Sugar Alliance. Open it ↗
When cocoa hit $12,000 a ton, some candy quietly changed what was inside.
The Guardian · Reporting + data
Try this
Pull the monthly cocoa price yourself from FRED (series PCOCOUSDM). By what percent did it rise from its 2022 average to its peak?
Bad weather and disease in West Africa sent cocoa from about $2,000–3,000 a ton to $12,000. Some candy makers quietly swapped cocoa for cheaper ingredients. This is the story behind our Candy trends chart.
About 300 dentists did it in 2007. More than 2,500 did by 2014.
NPR · The Salt · Reporting
Try this
Weigh your extra candy. What would a buyback pay you? Now compare with what that candy cost in the store. Who gets the better deal, and why might you still say yes?
A Wisconsin dentist started paying kids $1 a pound for extra candy and shipping it to troops overseas. About 300 dentists joined in 2007. By 2014 it was more than 2,500.
A 2005 law kept Halloween evening lighter. Candy makers were called “winners.”
NBC News · Reporting
Try this
Look up sunset time in your town on October 31. What would it be one hour earlier? How many minutes of daylight trick-or-treating does that change?
A 2005 law pushed the end of daylight saving time into November, so Halloween stayed lighter in the evening. NBC listed candy as a “winner.” The candy industry is widely thought to have lobbied for it; its trade group says that’s not so.
Every year, thousands of shoppers say what they’ll spend. Candy always tops the list.
NRF · Data hub
Try this
Divide this year’s candy spending by the number of U.S. kids (72 million, from the Census). Is that the cost of one kid’s haul? What’s missing from that math?
Every year the NRF surveys thousands of shoppers about costumes, candy and decorations. Candy is the item the most people buy. The yearly numbers make a great line chart.
Illinois grew 630 million pounds in 2022. Most of it went into cans.
USDA ERS · Government chart · free data
Try this
What fraction of the top six’s pumpkins came from Illinois? Draw it as a pie with six slices (Indiana 160, California 120, Michigan 90, Pennsylvania 90, Virginia 50).
The top six pumpkin states grew about 1.2 billion pounds in 2022, and Illinois grew 630 million of them. Most Illinois pumpkins go into cans, not onto porches.
By Gregory Astill, USDA Economic Research Service. Open it ↗
Two candies at a time, 269,000 times. Reese’s won 84% of its matchups.
FiveThirtyEight (archived) · Poll + open data · free data
Try this
Download the spreadsheet. Do chocolate candies win more often than fruity ones? Make two lists and average each.
Instead of asking “what’s your favorite?”, FiveThirtyEight showed people two candies at a time. Reese’s Peanut Butter Cups won 84% of its matchups. The data that seeds our Sweet 16 comes from here, and it’s free to download.
Votes for JOY minus votes for DESPAIR. The raw survey is free to download.
The Science Creative Quarterly (UBC) · Survey + open data · free data
Try this
Make your own Net Feelies at home: ask 10 people JOY, MEH or DESPAIR for 5 candies. Which candy is the most divisive (lots of both)?
Two scientists have ranked Halloween candy for years with a tongue-in-cheek score called Net Feelies: votes for JOY minus votes for DESPAIR. The raw survey answers are free to download as a spreadsheet.
Each candy plotted by price and by love. The bargains jump right out.
FlowingData · Chart
Try this
Find a candy that is cheap but loved, and one that is pricey but unloved. If you were handing out candy on a budget, which corner of the chart would you shop from?
A data designer plotted the FiveThirtyEight candies by how much they cost and how much people love them, using candy names instead of dots. The good deals jump out.
What everyone is searching for this October, ranked by Google.
Google Frightgeist · Interactive
Try this
Pick one costume. Sketch the graph you’d expect for searches from September to December. When does it peak? Why does it crash?
Every October, Google’s Frightgeist ranks the costumes, movies and party ideas people are searching for most. Searches aren’t sales, but they move fast and they’re a fun first look at data.
Whoever eats the poisoned corner loses. The first player can always win. On most bars, nobody knows how.
UCLA Math · Play online
Try this
Play on a 2-row bar until you find the winning move. Then try a square bar. (Big hint on Wikipedia, after you’ve tried.)
Take turns biting a chocolate bar. Each bite takes a square and everything above and to the right of it. Whoever eats the poisoned corner loses. Mathematicians proved the first player can always win, without being able to say how on most bars.
A computer scientist proved Candy Crush is as hard as the toughest puzzles in computing.
arXiv · Math paper
Try this
Read Smithsonian’s plain-English version first. Why can checking a winning set of moves be easy even when finding one is hard?
A computer scientist proved that winning a general Candy Crush board puts it in the same family as some of the hardest puzzles in computer science, like planning the best delivery routes.
A free text game where candy piles up every second. Spend it wisely.
Candy Box 2 · Free game
Try this
Keep a log: candy per second before and after each upgrade. How many seconds until an upgrade pays for itself? (Investors call that the payback period.)
A text-art adventure where candy piles up every second and you decide what to spend it on. It was made by a 19-year-old student in France and its code is open source.
31 in base 8 equals 25 in base 10. That’s the whole joke, and it’s great.
The Big Apple · Math joke
Try this
Write 25 in base 16 and in base 2. Then invent your own joke: which date in base 16 equals a date in base 10?
Read “OCT” as octal (base 8) and “DEC” as decimal (base 10). 31 in base 8 is 3 eights plus 1, which is 25. Martin Gardner printed it in Scientific American back in 1966.
Pumpkins, cats and ghosts drawn with nothing but equations.
Desmos · Graphs to color
Try this
Open Desmos and draw a pumpkin: start with a circle, x² + y² = 25. How would you squash it wider? Add a stem with a line segment.
College students drew pumpkins, cats and ghosts using only equations in the free Desmos graphing calculator, then turned them into a coloring book for a kindergarten class.
By Luke Walsh’s students (Catawba Valley CC). Open it ↗
The Army asked for a bar that tasted “a little better than a boiled potato.”
Hershey Community Archives · Archive
Try this
A regular Hershey bar is about 1.55 ounces. How many would you need to match one Ration D bar’s 600 calories? Check the calories on a real wrapper.
In 1937 the Army asked Hershey for an emergency bar: about 4 ounces, able to handle heat, packed with energy, and tasting “just a little better than a boiled potato,” so soldiers would save it for emergencies.
The second M was Hershey’s president’s son. His share: 20%.
HISTORY · History
Try this
If a company made $1,000,000 profit, what would a 20% owner get? Why might Mars have wanted a partner connected to Hershey in 1941?
Forrest Mars started M&M’s in 1941 with Bruce Murrie, son of Hershey’s president, who got a 20% stake. During World War II they were sold only to the military. Mars later bought Murrie out.
Invented in the 1880s, sold in a box with a rooster on it.
National Geographic · History
Try this
Check the claim: 35 million pounds ÷ 9 billion kernels = how many grams per kernel? (1 pound ≈ 454 grams.) Weigh a real kernel on a kitchen scale. Close?
Candy corn dates to the 1880s and was sold in boxes with a rooster on the front. In 2015 the candy makers’ trade group put yearly production at about 35 million pounds, or roughly 9 billion kernels.
Sugar was rationed to half a pound a week. Then Peanuts and Donald Duck helped bring trick-or-treating back.
HISTORY · History
Try this
Half a pound is about 227 grams. Read the grams of sugar on a fun-size wrapper. How many of those would one week’s ration cover?
Sugar was the first food rationed in World War II, cut to half a pound a week. Rationing ended in June 1947. Candy ads, a 1951 Peanuts strip and a 1952 Donald Duck cartoon helped bring trick-or-treating back big.
Kids who got a candy bar plus gum were less happy than kids who got just the bar.
NPR · Hidden Brain · Psychology
Try this
If you’re handing out candy, should the best piece go first or last? Run a tiny test with your family and see.
On Halloween night, Dartmouth researchers gave some trick-or-treaters a full-size bar and others the same bar plus a piece of gum. The kids who got the extra gum rated themselves less delighted. Endings shape how we remember things.
A sociologist checked 60 years of reports. No confirmed case of a stranger’s candy seriously hurting a child.
The Counter · Research
Try this
Why can a story be everywhere but still be rare? Count how many times you’ve heard the warning versus how many real cases exist.
Sociologist Joel Best has checked news reports since the 1950s. He has found no substantiated case of a child killed or seriously hurt by a contaminated treat picked up while trick-or-treating. Scary stories spread faster than the corrections.
By Joe Fassler, on Joel Best’s research. Open it ↗
Cars are the real Halloween danger. Here’s how to read a scary percent.
Computational Complexity blog · Math blog
Try this
If 2.07 per hour is 43% more than 1.45 per hour, check the math. Then: what’s one change (light, reflective tape, a grown-up) that cuts the real risk?
A 2018 study found 43% more pedestrian deaths on Halloween evenings, and it’s real: cars are the actual Halloween danger. A computer scientist shows how a big percent can be a small number, and why both matter.
What is the best Halloween candy, according to data?
In FiveThirtyEight’s candy power ranking, built from about 269,000 head-to-head votes, Reese’s Peanut Butter Cups won 84% of their matchups, the top score. The Candy Hierarchy survey from UBC’s Science Creative Quarterly scores candy a different way, by JOY votes minus DESPAIR votes.
How many licks does it take to get to the center of a Tootsie Pop?
It depends on who is licking. Tootsie Roll lists four studies: a Purdue licking machine averaged 364 licks, Purdue’s human volunteers averaged 252, a University of Michigan machine took 411, and Swarthmore junior high students averaged 144.
Sociologist Joel Best has tracked reports since the 1950s and found no substantiated case of a child killed or seriously hurt by a contaminated treat from a stranger while trick-or-treating. The bigger Halloween risk is traffic: a 2018 study found about 43% more pedestrian deaths on Halloween evenings.
Poured M&M’s fill about 68 to 71% of a container’s space, compared with about 64% for round balls (Princeton and Cornell physicists, Science, 2004). Estimate the jar’s volume, multiply by 0.7, and divide by the volume of one M&M.
Why do math people say Halloween equals Christmas?
Because OCT 31 = DEC 25. Read OCT as octal (base 8) and DEC as decimal (base 10): 31 in base 8 is 3 × 8 + 1 = 25. Martin Gardner printed the joke in Scientific American in 1966.
Not always. In a Dartmouth study on Halloween night, trick-or-treaters who got a full-size bar plus a piece of gum rated themselves less delighted than kids who got only the bar. People judge an experience heavily by how it ends.
Dump out your bag! Tap a candy once for every piece. Ate one? Use the bite button under it.
What’s in your bag?
Counting on the floor? Print the sorting mat, make your piles, then type the totals.
Type how many of each you have. Leave blanks for zero. Use the search above to add a candy that isn’t listed.
Your Candy Report
Gourdon’s Candy Market
After the haul
The candy is dumped, sorted and traded on the floor. Here’s what it weighs, what it’s worth, and how to make it last.
What your haul weighs and is worth
Weigh-in!
0pounds of candy
$0to buy it at the store
0pieces counted
See every weight and price
Candy
You have
Each weighs
Each costs
From
Candy tax and how long it lasts
Most parents sneak a few pieces, and about 3 in 10 charge an official “candy tax.” The most common share is 10%. Set yours, then see how long the rest lasts.
Start with your least favorite piece. Can you reach a full-size bar in five trades or fewer? Every trade has to be a yes from both kids. Write each one on a trade slip from the printable sheet.
The Nov 1 email
A family recap built from every kid’s real bag: how much they got, what it weighs and is worth, and how long it will last.
From: The Candy Market by Real World Math <hello@realworldmath.app>To: you@example.comSubject:
Candy Market Report · Halloween 2026
Each kid’s haul
What they learned about markets
How long will it last?
For grown-ups: check the wrapper
With cocoa prices so high, some candy now has very little real chocolate. “Milk chocolate” on the label means real chocolate; “chocolate candy” or “chocolate flavored” may not. Read the Guardian story. Want real cocoa? Sonoma County has bean-to-bar makers like Volo and Rainy Day.
Ask at dinner
“Why would someone trade two candies for one? Can a trade make both people better off?”
A good answer: people value things differently, so a swap can leave both happier. That’s what markets do every day.
You got this because you asked for your family’s Candy Market Report. Real World Math · realworldmath.app · hello@realworldmath.app · Unsubscribe in one click. We never ask for a child’s full name or location.
