Coin toss not so random after all, says groundbreaking study

[frauenfelder]

Originally published at: Coin toss not so random after all, says groundbreaking study | Boing Boing

…

[the_borderer]

@frauenfelder

The implications of the findings suggest the possibility that activities relying on coin flips — like gambling and certain random number generation processes — may need to adjust to account for this newly discovered bias.

Sportsball games would be the most obvious place to start.

Law 8 - The Start and Restart of Play | IFAB

• the referee tosses a coin and the team that wins the toss decides which goal to attack in the first half or to take the kick-off
• depending on the above, their opponents take the kick-off or decide which goal to attack in the first half

Law 10 - The Outcome of a Match | IFAB

• Unless there are other considerations (e.g. ground conditions, safety etc.), the referee tosses a coin to decide the goal at which the kicks will be taken, which may only be changed for safety reasons or if the goal or playing surface becomes unusable
• The referee tosses a coin again, and the team that wins the toss decides whether to take the first or second kick

[Doctor_Faustus]

Yeah, a coin toss starts every cricket match as well, to determine which side gets to choose whether they bat or bowl first. I think, maybe baseball, too?

[VenTatsu]

@frauenfelder

Coins of 46 different currencies were flipped by hand and caught in the palms of 48 student participants to record the landing side.

The lack of randomness in caught coins has been know for a while, the common recommendation if your looking for more true random result is to allow the coin to strike the ground and come to a rest. (but that has some issues too.)

@frauenfelder

This same-side bias was first predicted in a physics model by scientist Persi Diaconis.

This study seams to be less ground breaking and more confirmation of a believed truth.

[danimagoo]

@Doctor_Faustus

I think, maybe baseball, too?

No, there’s no coin toss in baseball. The visiting team always bats first in the inning. The team that bats last has a slight advantage, so that advantage is always given to the home team. In All Star games and the World Series, I believe the advantage alternates every year between the two leagues.

ETA: For the World Series, there are 7 games, so in one year, 4 of the games the American League team will be the home team and bat last in more games. The next year, the National League team will get that advantage. But the first 2 home games are at the “home” league, the next 3 at the “visiting” league, and then the final two back at the home team. So the advantage is minimized as much as possible.

[Doctor_Faustus]

That’s a shame, because there is a lot of strategy in deciding to bat or bowl first in cricket, depending on the weather, the opposing side’s composition and the pitch conditions (which in themselves are based on recent weather, the hosting country’s groundskeeping traditions, climate and recent groundskeeping decisions). Although I guess that’s less relevant in baseball anyway, since the ball doesn’t bounce and stadiums have more controlled conditions, so pitch conditions aren’t relevant.

[Jesse13927]

This is all the more reason why you should always carry a few dice on your person. (You only need to roll one; the rest are because you will inevitably lose several over the course of a day.)

[jaded]

I’ve read lots of papers on cryptography that reference the randomness of “tossing a theoretically fair coin” when they’re describing a 0.5 probability.

In cryptography a true 0.5 probability is often a critical requirement, as any results that are skewed even slightly can offer a massive benefit to an attacker. Using statistical analysis they can pry an encrypted message open one bit at a time, and with modern computing resources being capable of testing billions or trillions of guesses every second, we can assume that attackers have access to whatever equipment they need.

I doubt that many authors will find this study relevant, as they’re trying to impart the impression of true randomness rather than describing a physical manifestation of it. But it always pays to be precise, and this study just took away that precision.

[the_borderer]

Both rugby codes, shinty, hurling/camogie and Gaelic football too. I don’t know about other sportsball games.

Coin tosses used to decide more than who kicks off too.

1969-70 EUROPEAN CUP: Celtic on road to final after flip of a coin and Battle...

[euansmith]

Escape and Evasion Dice

Self Defence Dice

[MikeR]

@Doctor_Faustus

That’s a shame, because there is a lot of strategy in deciding to bat or bowl first in cricket, depending on the weather, the opposing side’s composition and the pitch conditions (which in themselves are based on recent weather, the hosting country’s groundskeeping traditions, climate and recent groundskeeping decisions).

And the most important cricket question of all - milk in the cup first or after?

[rob_dobbs]

@VenTatsu

The lack of randomness in caught coins has been know for a while

Heads

[bashful]

So we should flip them on edge now, right?

[nick_warr]

The cups aren’t cast iron are they?

[Nebulous]

Whether they let the coin hit the ground and then come to a stop was my first question too.
Next, since they have the coin caught by hand, was it the coin tosser catching the coin, or another person? I suspect that having a second person catching the coin would improve the randomness by elimination of an individual’s potentially synchronized reflexes.

[Murgatroyd]

The Conventional Wisdom ™ when I was growing up was that pennies and dimes flipped “tails”, while nickels and quarters flipped “heads”. I’m sure that was all backed up by decades of empirical study.

[timd]

What does this mean for the game of two-up, (on the one Aussie day it is legal)?

[mous]

the variation depends on the amount of thumb pressure applied. the study mainly shows that the pressure is consistent among participants(just enough to show it was flipped). I suspect more variation of age and gender would alter the results. as a kid I tended to flip the coin a lot higher than I do now. more time in the air adds randomness.

say is this new reply format obscure or what?

[Dachshund_lover]

That’s why one person tosses the coin and the other calls it in mid-air.

[FGD135]

The coin toss is not random at all if you are fast and have nimble fingers.