Lab · c16.s01.L1
Biased-coin entropy
A biased coin's entropy is the probability-weighted average of the surprises of its two outcomes, and that average is largest, at one bit, when the coin is fair.
Question: How much information is in one symbol?
Frame 1 of 9: One time in a hundred: headsAt p = 0.01 heads is almost never the outcome, but when it comes it is worth 6.644 bits: the tallest bar on the chart. Its share is only 0.066 bits because it happens so seldom, and tails adds 0.014 bits, so the stack reaches 0.081 bits, the smallest average on the grid. The gold line marks one bit of surprise: heads towers above it, yet the average sits below it.
One time in a hundred: heads. At p = 0.01 heads is almost never the outcome, but when it comes it is worth 6.644 bits: the tallest bar on the chart. Its share is only 0.066 bits because it happens so seldom, and tails adds 0.014 bits, so the stack reaches 0.081 bits, the smallest average on the grid. The gold line marks one bit of surprise: heads towers above it, yet the average sits below it.
Keyboard: Space plays or pauses, Left and Right step, Home goes to the first frame, End goes to the last frame.
Text description of this lab
A slider sets the probability that a coin lands heads. Two vertical bars show the surprise of each outcome: heads, common, short; tails, rare, tall. A shaded part of each bar is its probability-weighted share, and those two shares stack to a point on the entropy curve. At a fair coin the point peaks at one bit; toward either edge the rare outcome grows tall but its share shrinks, and the point falls to nearly zero.
Return to the section that hosts this lab
Caption status: re-simulated. Every lab declares one of published, re-simulated, schematic, synthetic or interpolated.