Guide

Why ten guesses tell you little

6 minute read Published Updated By Spirited AI

Sixth's first result card carries one line: five guesses tell you little, a hundred start to tell a story. Ten are barely better. It is the most important sentence in the app. This guide gives the arithmetic behind it, mode by mode, then three habits of mind that make small samples feel more meaningful than they are.

Ten guesses in Coin

In Coin the chance of a hit is 1 in 2, so ten rounds expect 5 hits with a standard deviation of about 1.6. Seven of ten feels like a 70 percent success rate. It happens to a random guesser about 17 percent of the time, one session in six, and as a Z-score it is only 1.3. Eight of ten happens about 5.5 percent of the time, about 1 in 18. Even ten of ten, which feels like proof, is 1 in 1,024.

Ten guesses in Symbols

In Symbols the chance is 1 in 5, so ten rounds expect 2 hits with a standard deviation of about 1.3. Four of ten doubles the expected rate and feels strong; luck reaches it about 12 percent of the time, 1 in 8. Five of ten is a Z-score of 2.4 and 1 in 30. On a Symbols day in the daily challenge that is a very good day, and it still arrives by chance to one player in thirty.

Ten guesses in Numbers

In Numbers the chance is 1 in 9. Ten rounds expect only 1.1 hits, so a single lucky hit changes everything. Three of ten is a Z-score of 1.9 and happens about 9 percent of the time, 1 in 11. Four of ten is a Z-score of 2.9 and happens 1.8 percent of the time, 1 in 54, from luck alone. Long odds make small samples lumpy, which is why Numbers and Full Deck sit out of the daily challenge, and why the app makes no significance statement until it expects at least five hits: 10 rounds in Coin, 25 in Symbols, 45 in Numbers, 65 in Full Deck.

The same results side by side

ModeChanceExpected in 10ResultZ-scoreLuck alone gets there
Coin1 in 257 of 101.31 in 6
Coin1 in 258 of 101.91 in 18
Symbols1 in 524 of 101.61 in 8
Symbols1 in 525 of 102.41 in 30
Numbers1 in 91.13 of 101.91 in 11
Numbers1 in 91.14 of 102.91 in 54

Every row would feel like something to the person who got it. None is rare.

What a hundred rounds change

Now stretch Symbols to 100 rounds. Chance expects 20 hits with a standard deviation of 4. To reach the Z-score of 2.0 that 9 of 25 gives, you need 28 hits, a 28 percent hit rate rather than 36. At 1,000 rounds, chance expects 200 with a standard deviation of about 12.6; a modest 23 percent hit rate, 230 hits, is a Z-score of 2.4 and a Sense Score of 136, which luck produces about 1 time in 93. The small edge that vanishes in ten rounds is plain in a thousand. That is the story the Sense Score is built to tell.

The law of small numbers

In 1971 the psychologists Amos Tversky and Daniel Kahneman described the law of small numbers: the intuition that a small sample should look like the population it came from. People expect ten coin flips to come out near five and five, and treat seven and three as a signal. In fact small samples are wild; the ordinary band for ten flips runs from 3 to 7 heads. That is what randomness looks like up close, and why a ten-round session is the wrong place to look for anything.

The gambler's fallacy

The gambler's fallacy is the feeling that a result is due: after four circles in a row, surely a star is next. In a physical deck of 25 Zener cards there was a sliver of truth in it, since once all five circles were dealt none remained. In Sixth there is none. Every round is an independent draw from the phone's random generator, which has no memory. Four circles in a row leave a fifth exactly as likely as before: one in five. Guessing against the run cannot help.

The hot hand

The hot hand is the mirror image: the belief that a run of hits means you are in form and the next guess will land. In 1985 Thomas Gilovich, Robert Vallone and Amos Tversky examined basketball shooting records and reported that runs of made shots were no longer than chance predicts; in 2018 Joshua Miller and Adam Sanjurjo showed a subtle counting bias in that analysis, and the basketball question is open again. For guessing against a random generator it is not. Randomness produces runs: a string of five hits appears somewhere in 1,000 random Symbols rounds about 22 percent of the time, and in Coin a run of ten appears somewhere in 1,000 rounds about 39 percent of the time. The On Fire and Ten Straight achievements mark those moments. They are moments, not evidence of a change in you. Enjoy them and ignore them as data.

Common questions

If ten rounds tell you so little, why is the daily challenge ten rounds?

Because the daily challenge is built for habit, not statistics. Ten rounds a day is easy to keep up, and every daily round feeds your lifetime score. Three hundred rounds a month tell a far better story than any single day.

How many rounds before my score means something?

The score appears at 20 rounds and is rough there. By 100 rounds the standard deviation in Symbols is 4 hits, so a sustained edge of a few percentage points starts to show. By several hundred rounds a score that stays above 130 is unusual; one that drifts back to 100 is what nearly everyone's does. The guide on measuring intuition explains how to read it, and the daily habit guide is about accumulating the rounds.

Is a perfect ten of ten in Coin proof of anything?

No. It is a 1 in 1,024 event, rare for you but routine across many players. If it recurs over hundreds of rounds, the Sense Score will show it. If not, it was luck.

Try it yourself

Sixth is free to download on iPhone and Android. Symbols mode is free, and your Sense Score appears after 20 rounds.