Games and odds · free tool

Wingo prediction: what a predictor can and cannot know

A predictor claims to know the next result. See what luck alone does to a record like that, and the check that tells a predictor from a lucky guesser.

Tool · runs in your browserNothing is uploaded

1. Could luck alone have done that?

Type what a predictor, app or channel claims. The tool shows how often a blind guesser would do as well, worked out exactly. It predicts nothing and reads no results.

Use 50 for a two-way call such as big or small. As an illustration, a colour that covers 4 digits of 10 is 40, and one number is 10. Use the chances your own app states.

60%hit rate claimed: 60 right out of 100 calls
2.84%chance that blind guessing, right 50% of the time, does at least this well
1 in 35blind guessers that look this good
28 of 1,000blind guessers in the crowd that would look this good
± 10.0 pointshow far luck alone moves a hit rate over this many calls
132calls before a hit rate this high would happen to blind guessing less than 1 time in 100

A predictor that gets 60 of 100 calls right has a hit rate of 60%. Blind guessing, right 50% of the time, does at least that well 2.84% of the time, so among 1,000 blind guessers about 28 would look this good. If the same hit rate held over about 132 calls, blind guessing would manage it less than 1 time in 100. A hit rate in the past says nothing about the next call.

Blind guessing over 100 calls, right 50% of the time, with 1,000 guessers
Hit rateRight out of 100Chance blind guessing does this wellGuessers that would look this good
55%5518.4%184
60%602.84%28
65%650.18%2
70%70<0.01%under 1
80%80<0.01%under 1

2. Do the popular “systems” beat guessing?

Four rules people swear by, each run over 100,000 fair rounds. A guess is right 50% of the time.

50.26%best score of four popular rules over 99,995 fair rounds
49.74%worst score of the four
0.26 pointsfurthest any rule sits from the 50% that plain guessing gets
± 0.32 pointshow far luck alone moves a score over that many rounds
± 10.0 pointshow far luck alone moves a score over just 100 calls

Four popular rules were each scored on one fixed run of 100,000 fair rounds, every guess made from the rounds before it. All four land within 0.26 points of 50%, which is what guessing gives. Over just 100 calls luck alone moves a hit rate by about 10.0 points either way, so a short record cannot tell a clever rule from a lucky one.

Four popular rules on one fixed run of 100,000 fair rounds
SystemRounds testedRight
Repeat the last result99,99549.95%
Pick the opposite of the last result99,99550.05%
Follow the majority of the last five99,99550.26%
Fade the majority of the last five99,99549.74%

A prediction is a claim about a result that has not happened yet. You cannot check it in advance, but you can check a predictor’s past calls against luck. A hit rate of 60% over 100 calls is something blind guessing reaches 2.84% of the time, so a record like that proves nothing about the next call.

In 100 calls luck alone moves a hit rate by ± 10.0 points, so a rate of 60% sits inside what guessing produces. Some blind guessers always get lucky, and about 28 of 1,000 would match a rate like this one. The one you hear about is the one that did. To tell skill from luck you need a long record written down before each result.

The tool runs in your browser and sends nothing anywhere.

Common questions

Can a predictor know the next colour or number?

Not if rounds are drawn at random and independently, as fair rounds are. A result that has not happened cannot be read from anywhere, and earlier results do not change the next one. We cannot see inside any platform, so this describes fair random rounds.

How likely is it that a predictor’s record is just luck?

More likely than most people think. A blind guesser gets 60 or more right out of 100 about 3 times in 100. The tool above works out the exact chance for the record you enter, so you can see how unusual it really is.

Why does some channel always seem to be right?

Because there are many of them. If a thousand channels guess blind, a few will have a good run, and those are the ones whose screenshots get shared. The ones that guessed wrong go quiet.

Does anything I type here leave my phone?

No. The tool runs in your browser and works everything out on your own device. Nothing you enter is uploaded, and there are no accounts.

Does this page predict the next result?

No. It cannot, and it is built to show why. It makes fair random rounds and counts what happens, and the chances it prints are worked out exactly. Nothing in it looks ahead.

More on the prediction simulator

The full tool