Pick a Number Between 1 and 10: Why Most People Say 7
Pick a number between 1 and 10 and most people choose 7. Learn why human picks are biased, how random number generators work, and how to run a fair draw.
Updated
Asked to pick a number between 1 and 10, more people say 7 than any other number, by a wide margin. A fair random pick would give each number a 1 in 10 chance, but human choices are not random: we avoid the ends, the middle and anything that feels like a pattern. If you need an unbiased number, use a generator, a ten-sided die or drawn slips instead of a person.
Why people pick 7
In informal polls and psychology experiments alike, 7 comes out on top. A well-known 1976 study by Michael Kubovy and Joseph Psotka asked people for a digit and found 7 chosen by more than a quarter of them, where chance would predict one in ten. Exact percentages vary from survey to survey, but the pattern is consistent.
The usual explanation is that people do not try to be random. They try to seem random, and they do it by ruling numbers out:
- 1 and 10 are the endpoints, so they feel too obvious.
- 5 is the exact middle, which also feels too obvious.
- Even numbers such as 2, 4, 6 and 8 feel orderly and predictable.
- 3 and 9 are odd, but 9 is three threes, and 3 is small.
- 7 is odd, prime, not a factor or multiple of anything else in the range, and off-center. It feels like the least patterned choice.
The same thing happens in larger ranges. Asked for a number from 1 to 100, people tend to favor odd numbers that end in 3 or 7, and 37 is often reported as a common pick.
Human bias versus true randomness
A random process has no memory and no preferences. People have both. The most reliable differences:
| Random sequence | Human-made "random" sequence |
|---|---|
| Every number equally likely | Favorites such as 7 and 3 overused |
| Repeats happen often | Repeats avoided |
| Long streaks appear by chance | Streaks broken up deliberately |
| Ends of the range used normally | Ends of the range underused |
| Each pick independent | Each pick influenced by the last |
Repeats are the giveaway. In ten random picks from 1 to 10, getting no repeated number at all is very unlikely: the chance is 10! divided by 10 to the power of 10, which is about 0.04 percent. A person asked to write ten random numbers will often produce few or no repeats, because repeats feel wrong.
This matters whenever a choice is supposed to be fair. A teacher who "randomly" calls on students, or a person who picks a raffle winner from memory, is applying preferences they are not aware of.
How to pick a number between 1 and 10 fairly
Any of these gives each number the same chance.
- A random number generator. Set the range to 1 through 10 and generate.
- A ten-sided die. Tabletop game dice include a d10, usually numbered 0 to 9, with 0 read as 10.
- Ten slips of paper. Write 1 to 10, fold them the same way, mix, and draw one.
- Playing cards. Take the ace through ten of one suit, shuffle, and draw. The ace counts as 1.
- Coins, if that is all you have. Flip a coin four times and read the result as a four digit binary number from 0 to 15. Add one. If the result is above 10, discard it and start again.
Some methods look fair and are not. Rolling two ordinary dice and adding them gives 2 to 12, and 7 comes up six times as often as 2. Asking a friend to "just pick one" brings back the bias described above. And discarding results you dislike until you get one you prefer is a choice, not a draw.
Pseudo-random versus cryptographic randomness
Computers follow instructions exactly, which makes genuine unpredictability awkward for them. There are three broad sources.
Pseudo-random number generators use a mathematical formula. Starting from a number called a seed, each output is calculated from the previous state. The results pass statistical tests for randomness, but the sequence is fully determined: the same seed gives the same sequence. JavaScript's Math.random() and Python's random module work this way. They are fast and entirely adequate for games, simulations, sampling and casual decisions.
Cryptographically secure generators are designed so that seeing past outputs does not help anyone predict future ones. They are seeded from unpredictable events gathered by the operating system. In browsers this is crypto.getRandomValues(), and in Python it is the secrets module. Passwords, tokens and encryption keys must come from this kind of generator. Our guide on how to create a strong password explains why.
True random sources measure physical processes such as electronic noise or radioactive decay. They are used to seed the generators above and for applications such as lotteries.
One more subtlety: even a perfect source can be used badly. Suppose you take a random byte, a value from 0 to 255, and reduce it to 0 through 9 by taking the remainder after dividing by 10. Because 256 is not a multiple of 10, the remainders 0 to 5 can each be produced by 26 different bytes, and 6 to 9 by only 25. The fix is to discard bytes of 250 or more and draw again. This is called avoiding modulo bias, and well-written libraries handle it for you.
Running a fair draw
For a raffle, a giveaway or choosing who goes first:
- Fix the list before drawing. Number every entry from 1 to N, and do not add or remove entries afterwards.
- State the rule in advance: how many winners, and what happens if a winner is ineligible.
- Draw once. No redraws because the result was inconvenient.
- For several winners, draw without replacement so nobody is picked twice.
- Do it where people can see, or record the screen, if trust matters.
For anything with money or legal weight attached, such as a regulated prize draw, use a method that meets the rules that apply to it, which usually means an audited or witnessed process.
Using the Random Number Generator
Our Random Number Generator takes a minimum and a maximum, so for this question you set them to 1 and 10 and press Generate. It can also:
- Produce up to 50 numbers at once with the count slider.
- Return only unique numbers, which is the setting for picking several winners from a numbered list.
- Output decimals, with up to six decimal places.
- Show the sum, average, smallest and largest value when you generate more than one number.
- Copy all results in one click.
For a 1 to 100 pick, change the maximum to 100. It is intended for everyday uses such as games, classroom picks, sampling and settling decisions, not for generating passwords or security keys. For those, use the password generator.
Try it
Think of a number between 1 and 10 first, then open the Random Number Generator, set the range to 1 and 10, and generate twenty picks. Compare the spread, repeats included, with what you would have written by hand. More generators are in the random category.