Lab 10 / Uncertainty

Why fair sequences form streaks

How unusual is the longest run in one fair sequence?

7 min synthetic data fixed seed

Generate binary sequences, count their runs, and compare independent trials with a mechanism that alternates too often. Change one control at a time, then open the values under the chart before interpreting the picture.

Interactive synthetic lab
runs in this browser

View the values behind this chart

Loading the deterministic simulation...

SEED

What this display answers

Independent fair outcomes do not need to alternate. Runs and clusters are valid outputs, and longer sequences create more places for a memorable streak to occur.

The repeated histogram puts the longest run from one displayed sequence into context. Raising the switch chance above 50% creates too many alternations and shortens typical longest runs even though heads and tails can remain balanced.

What to notice

  1. At a 50% switch chance, read the displayed sequence and count its runs before comparing it with the repeated summary.
  2. Increase sequence length. Typical longest runs grow because the sequence offers more opportunities for one.
  3. Set switch chance to 80%. The mean run count rises sharply and long same-symbol streaks become less common.

The model behind it

The first symbol is equally likely to be H or T. Every later symbol switches from the previous one with the selected probability.

A 50% switch chance is equivalent to independent fair Bernoulli trials. Other settings deliberately introduce serial dependence and are included for comparison, not as models of a physical coin.

Where the result stops

One ordinary run count does not prove randomness, and one long streak does not disprove it.

The browser generator is deterministic from its seed and is not intended for security, gambling, or prediction.