Probability & Statistics

Four Coins Against Five

I toss four coins and you toss five. You win only by getting strictly more heads than I do. That extra coin sounds like a small edge, so how likely are you to win?

solvedmedium1 min

I toss four fair coins and you toss five. You win if you get strictly more heads than I do. What is the probability that you win?

Reveal solutionHide solution

#A tails mirror

Let HH be your heads and GG be mine, and set D=HGD = H - G, so you win exactly when D1D \ge 1. Now turn every coin over to its opposite face. Heads and tails are equally likely, so the flipped table of outcomes is exactly as probable as the original. Your head count becomes 5H5 - H and mine becomes 4G4 - G, so the difference transforms as

D=HG    (5H)(4G)=1D.(1)D = H - G \;\longmapsto\; (5 - H) - (4 - G) = 1 - D. \tag{1}

Because the flip preserves probabilities, DD and 1D1 - D have the same distribution.

you win-4-3-2-1012345
Flipping every coin sends the difference D to one minus D, mirroring the bars about the dashed line at one half. The winning scores, a difference of one or more, sit entirely on one side of that line and carry exactly half the total weight.

#Reading off the answer

The reflection D1DD \mapsto 1 - D centres the distribution on 12\tfrac12, so

P(D1)=P(1D1)=P(D0).(2)\PP(D \ge 1) = \PP(1 - D \ge 1) = \PP(D \le 0). \tag{2}

Those two events are complementary, because the integer DD is either at least 11 or at most 00. Complementary probabilities that are also equal must each be one half,

P(you win)=P(D1)=12.(3)\PP(\text{you win}) = \PP(D \ge 1) = \tfrac12. \tag{3}

#Why the extra coin matters

The reflection centre sits at 12\tfrac12 precisely because you hold one more coin than I do. That centre falls strictly between the losing scores D0D \le 0 and the winning scores D1D \ge 1, splitting the mass into two equal halves. With equal coin counts the centre would sit at 00, a genuine tie outcome, and the clean halving would break.