Wednesday, October 26, 2011

Urban Myth - Chessex Dice

A long time ago in a galaxy far, far away……an urban myth was born.
On Dakka Dakka the US-based wargames forum this myth has persisted and is treated as fact.
The myth is that if you use Chessex dice or any round cornered pip indented then you will roll “1” 29% of the time. You can read all about it here.

Twenty-nine percent…..when the expectation is 16.6%. Determined after thousands and thousands of rolls. Any idea how improbable that is. Well statistically it starting to get into the realms of “A Room Full of Monkeys Typing Shakespeare”!

I’m willing to say that you could look at the dice and clearly see the imperfections.

Dice are never perfect. However the chance of picking up a block of Chessex dice and rolling them 28 times then coming up with 29% “1”s or indeed any single number is about as likely as finding intelligent life in Newcastle.

I call “Bullshit”.


  1. You should see some of the muppets on the Privateer forums using this as evidence for exclusively using casino dice during games. Some of them are so extreme as to not want to play against opponents not using casino dice as "they are cheating both you and themselves". Offense intended for anyone that actually believes this.

  2. Honestly, do people really believe that? Also, if anyone was to say I was 'cheating' because I didn't buy into their 'dice superstitions' (e.g. weird die cult), well, I would say that they represent everything that's rotten in gaming.

  3. When Casino dice come with sweet logos of Khorne, Tzeentch, Nugle and Slaneesh, I might give a crap.

  4. I have some dice with an extra 1 instead of the 6. Should be able to replicate the experiment with that.

  5. You are far more likely to produce incorrect results with casino dice, given that you are almsot never rolling them correctly (have fun finding an opponent happy with you bouncing all your rolls off a 'wall').

    I've fucked around with chessex dice enough that I'm extremely comfortable also calling bullshit on this. It's just a rough lesson for a lot of people to learn that trolls can just make up a whole bunch of shit, and people will run off believing them

  6. The findings seem really improbable, and I'd like to see someone else try to replicate them -- perhaps we will at some point.

    The method used to paint the numbers on standard Chessex dice leaves the corners and edges inconsistently rounded. *If* that caused a statistically significant deviation, you'd expect the deviation to be different from die to die -- so some would roll more 1s, others would roll more 5s, etc (this is the GameScience argument).