Probability Calculation

**JohnD** · April 5th, 2018, 12:39

Hoping someone can help figure this out, my brain is more than a little sleep deprived.

If I have an event that has a 14% chance of occurring every time I generate a random result, then across 100 rolls, this event should occur 14 times. With a small sample size, maybe a few more or a few less. Correct?

Then what is the likelihood/probability of said event not occurring even once over said 100 random results?

**shadzar** · April 5th, 2018, 12:43

a 1 in 7 change of something happening not happening after 100 tries?

without the math, the chance of it not happening is very low says probability, however statistics is a different story. it could statistically happen 0 times or 100 times, which isnt very probable.

**Bidmaron** · April 5th, 2018, 12:44

2.82e-7. Highly unlikely

**dulux-oz** · April 5th, 2018, 13:17

The trouble is, 100 is a very small sample size (not withstanding previous posters calculations).

We were taught at University (3rd year Statistics) that while a truly random sample size of about 32 can give a statistical result with an error margin of only plus/minus 5%, you really need sample sizes in the 1,000s to 10,000s to generate statistically significant results with errors of only 1% or less.

So while Bidmaron is correct in saying its highly unlikely, its still possible for the result obtained to be correct.

However, I'd like to know more about how the 14% chance was determined in the first place, and how the 100 results were generated, and based on those two answers I'd be looking at some sort of biases, such as imperfect dice, bad maths, human error, or something similar.

John?

**damned** · April 5th, 2018, 14:06

Dice dont give a toss about history.
They dont care what the previous roll was.
They dont care what any of the previous rolls were.
When they roll they only worry about their own roll.

If you roll a d6 100x and dont get a single 6, the next time you roll a d6 there is still an 83% chance you wont get a 6.

**JohnD** · April 5th, 2018, 14:51

Yeah so, I'm leading towards a misrepresentation of the percentages.

All I have is a claim that probability of A is 14%. Not A is 86%. Either A or Not A happen on every generation of a result. There is no insight nor information on how the "random" outcomes are generated.

In response to inquiries I'm told the probability is actually just 2.8%, which is contradicted by the initial verbiage stating 14%, verbiage which is strangely no longer available (shocking). At the moved goalposts rate, it certainly is very believable that for example you could roll d100 100 times and not get any results of say 4, 29 or 85.

Going on history, I'm inclined to assume a willful misrepresentation of the odds, especially since the verbiage has been removed now.

Nothing to do with FG, just thought there are a lot of knowledgeable people here who probably unlike me have slept since Monday. Thanks guys.

**dulux-oz** · April 5th, 2018, 14:59

Originally Posted by JohnD

Yeah so, I'm leading towards a misrepresentation of the percentages.

All I have is a claim that probability of A is 14%. Not A is 86%. Either A or Not A happen on every generation of a result. There is no insight nor information on how the "random" outcomes are generated.

In response to inquiries I'm told the probability is actually just 2.8%, which is contradicted by the initial verbiage stating 14%, verbiage which is strangely no longer available (shocking). At the moved goalposts rate, it certainly is very believable that for example you could roll d100 100 times and not get any results of say 4, 29 or 85.

Going on history, I'm inclined to assume a willful misrepresentation of the odds, especially since the verbiage has been removed now.

Nothing to do with FG, just thought there are a lot of knowledgeable people here who probably unlike me have slept since Monday. Thanks guys.

Intriguing - I would love to know more, but its cool if nothing more could (or should) be said

Have a good one, John - and maybe get some sleep?

**shadzar** · April 6th, 2018, 02:02

Originally Posted by damned

Dice dont give a toss about history.

sounds like something Andrew Clay would have said during one of his shows.

**Bidmaron** · April 6th, 2018, 04:54

The probability in 100 rolls is .86 to the 100th power, which is how I arrived at that tiny *** number I previously posted.

**damned** · April 6th, 2018, 08:53

By my calculations not rolling 3 specific numbers out of 100 in 100 rolls is very close to 5%.
Not unreasonable odds.

Thread: Probability Calculation

Thread Tools