Why do you have to perform many trials in probability experiments?

Answer 1

Probability determines likely outcomes, not what will happen every time. For instance, if you tossed a coin, the probabilty of coming up heads or tails is even, but there is also a vanishingly small possibility that it wil land on its edge.

If you only tossed the coin once, and it came up heads, you could not infer from that result that "if I toss a coin, it will come up heads." The next 10 tosses of the coin might come up heads, but if the experiment is run enough times, a roughly equal number of heads and tails would be observed.

This is true of any probability experiment. In biological experiments especially, because we are dealing with the unpredictability of living organisms, a watershed has to be determined, below which we can say with reasonable certainty that the probability of a particular outcome is less than, say 0.5%.

If you imagine a bell, the edges flare out, and it rises steeply to a shallow curve at the top. If we place the results of our repeated experiments under that bell, most will fit somewhere under the shallow part of the curve, with the results falling of rapidly to approaching 0 as the values move down to the flare. With just 1 or 2 results, it would be easy to get a distorted picture of a likely outcome.