Quick answer: Multiply trick-or-treaters per hour × hours your porch light is on × 2–3 pieces each. A typical US door (20 children an hour for three hours) needs 120–180 pieces, or two to three 75-count bags, 2.2–3.4 kg. UK doors see far fewer: 30–60 sweets, two treat-size multipacks.
Two to three pieces per child. Everything else is arithmetic on a number only you can estimate: how many children reach your door.
Almost every guide answers with a flat bag count — "buy three bags" — which is useless, because a townhouse on a busy street and a farmhouse down a dirt road get the same advice. The per-child figure is the stable part; turnout is local. So the honest answer is a rate:
Trick-or-treaters per hour × hours the light is on × 2–3 pieces per child = total pieces.
That structure isn't ours. It comes from the one widely-circulated formula here, built by Mars Wrigley and the delivery service Shipt, which adds a second term: days until Halloween × household size × pieces your family eats per day. That half is a joke that isn't a joke — buy in early October and you will eat some of it.
What to buy, by turnout
Three hours of trick-or-treating, at 2–3 pieces per child. Bags are 75-count, the size sold as a 39.43 oz Hershey assorted variety bag.
| Kids per hour | Pieces needed | Bags (75 ct) |
|---|---|---|
| 5 — rural, long driveway | 30–45 | 1 |
| 10 — quiet suburb | 60–90 | 1–2 |
| 20 — typical suburb | 120–180 | 2–3 |
| 40 — busy walkable street | 240–360 | 4–5 |
| 60 — destination street | 360–540 | 5–8 |
Four hours rather than three? Add a third. Note how fast this scales: the gap between a quiet street and a well-known one is not one extra bag, it's six.
Estimating your own turnout
There is no national dataset on trick-or-treaters per household — we looked. The closest thing is demographic arithmetic: the Atlanta Regional Commission's census-based count found 441,990 children aged 5–11 across 1,886,751 metro Atlanta households — under one prime-age trick-or-treater per two homes.
That average is not your street, and the gap is the whole problem. Children cluster, and walk to where the doors are close together and the lights are on. Four things move your door off the average:
- Door density. Children have a fixed amount of walking in them. Twenty doors within a two-minute walk get worked over; four houses per mile get skipped.
- How many neighbours participate. If yours is one of three lit porches, you absorb the whole street. The most under-appreciated factor, and it cuts both ways.
- Reputation. Heavily decorated streets import children by car from surrounding neighbourhoods. If that's your block, use the bottom row of the table.
- Weather. Rain thins turnout — everyone agrees, nobody has published a number. Mars Wrigley and Shipt say divide the total by 1.5 if rain is forecast. We found no turnout study behind that figure; it is a guess with a decimal point on it.
Halloween 2026 falls on a Saturday — no school night, later start, longer window. We could find no published data quantifying a weekend uplift, so we won't invent one, but the risk runs one way: plan for the upper end of your range and assume the door goes quiet an hour later than usual.
The sources disagree about who is even handing out candy
This matters directly, because the fewer doors that participate, the more children arrive at yours. Four surveys from the same season:
| Source | Share handing out treats | Sample |
|---|---|---|
| National Retail Federation (US) | 66% | 8,045 |
| Circana (US households) | 58% | not stated |
| Empower (US) | 35% | 2,000 |
| YouGov (Great Britain) | 28% | 2,069 |
The three American figures — 66%, 58% and 35% — were all fielded within a fortnight of each other in September 2025. That is not noise, it's a question-wording gap: the NRF asked celebrants what they plan to buy, Empower asked all adults what they will do. We won't average them into a tidy 53%. The useful reading is the direction of the error — intention surveys run high, so assume fewer neighbours have their lights on than the headlines suggest, and that your door takes more than its share.
The spend figures are firmer: the NRF puts total US candy spending at $3.9 billion in 2025, and Empower found an average candy budget of $58 per household. At the 14 cents a piece implied by Hershey's 290-piece, 85.89 oz bulk bag at $39.99, $58 buys roughly 400 pieces — well past what a typical suburban door needs. Most people over-buy, and nobody in the supply chain is in a hurry to correct them.
Who is telling you this, and what they sell
Every quantitative source on this page has something to gain from a bigger number. Mars Wrigley makes Snickers, Twix and M&M's. Shipt delivers groceries. The National Retail Federation represents retailers. Circana sells sales analytics to manufacturers. None of them are lying, and their figures are the only ones that exist — but nobody in this chain profits from you buying one bag fewer.
The 2–3 pieces figure is the one worth interrogating, because it does the most work here — and it traces to worked examples in the Mars Wrigley material, not to any observational count: Snack Food & Wholesale Bakery demonstrates it at 3 pieces, Food Republic at 2. A 50% swing on the key input, between two illustrations rather than two studies. Take the smaller number if you like: two fun-size pieces is a normal handful and no child has ever complained.
The UK is a different problem
Only 28% of British adults hand out sweets, against 66% of Americans, and YouGov found 51% of Britons don't mark Halloween at all, with 55% of those who do spending under £25 on everything.
Volumes are far lower: a typical UK door sees a handful of groups over about two hours, so 30–60 sweets is realistic — two supermarket treat-size multipacks, which run 12–20 bars each (an ASDA Cadbury treatsize variety bag is 207 g). But participation is concentrated: because so few doors take part, British trick-or-treating runs as a short, dense burst along whichever streets are known to be doing it. If you put a pumpkin out and your neighbours don't, budget like an American.
No UK trade body publishes per-child guidance and there is no British equivalent of the Mars Wrigley formula. The 2–3 pieces rate is imported; the UK turnout figures here are ours, extrapolated from participation rates rather than measured.
Buy in a form you don't mind eating
Leftovers are the normal outcome, not the failure case, so over-buy in a shape that survives: individually wrapped pieces keep for months. And buy by piece, never by weight. Bag sizes are wildly inconsistent — the same Hershey assortment sells as a 100-piece 31.55 oz bag and a 290-piece 85.89 oz bag — and the ounces tell you nothing about how many doorsteps you can serve.
The rest of the night? Our Halloween party quantities guide covers the indoor half, how many appetizers per person handles the grown-up table, and the party drink calculator takes over once the children are home.