Imagine you are doing a coin toss. You assume that the coin is a fair coin which has two sides: a picture and a number. When you do the first toss, what appears is the picture. In this stage you feel everything is normal. The probability for the number side and the picture side are same: 0.5.
Then, you do the second toss. And you got the picture again. You might feel something a little bit suspicious, but still you can tolerate it. The picture appears again even the probability in this second toss is 0.25. You do the third toss and get the same result. Of course you will start doubting the fairness since the probability in this stage is 0.12 but it still happens. You do the fourth and fifth toss, and again it is just the same.
At this stage, you would very suspicious right? The probability for the last two toss are 0.12 and 0.06, respectively. However, it just happens like that. The more you try and get the same result, the more you will feel uneasy about it.
The level of this uneasy feeling in mathematical term is called Statistical Significance (P-Value). A p-value of 5% or lower is often considered to be statistically significant. So, if that is the case it means you are 95% confident that the result is not logical so that you are going to reject it. You can change the p-value into some certain percentage so that will be rejected. It can be 10%, for example. But generally 5% is considered the best limit.
Program Link 