Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

This is an automated message added by the articles system.

A new article titled That's How I Roll - A Scientific Analysis of Dice has been added to the dakka articles system.

This message thread is for the discussion of content in the article. Any updates to the article will be automatically added to this thread and (if you are logged in), you can get automatic email updates as and when the article changes by clicking 'Email me when this thread changes' below. If you are the author or have added a lot to the article then it is a good idea to do this to ensure that you maintain control over the content of the article.

If you have anything to add to the article, then just jump in and edit it by going to the actual article page and clicking 'edit' (the link can be found just above the article). If there is no edit link then the article is locked for now, so just add your comments or content to this thread and if they are appropriate and the article owner checks this thread, they should get merged in.

If there is something in the article that you wish to debate, then this is the place to do it. Just hit the reply button and get chatting. You need to be registered and logged in to post in the forums so if you are an anonymous article editor then now would be a great time to register and join in dakka's great forum discussion!

Holy cow- AWESOME article. I'll definitely be buying some casino dice.

Just dont use them for your LD tests

Good to know. I have actually seen a player do GW for leadership and casino for everything else. Interesting.

My reaction to this article was "wow, why would anyone make such a garbage product." (cheap dice, not the article!) I then proceeded to order some casino dice, but now a part of me is regretting that decision.

I began to wonder if a disadvantage shared by all is really a disadvantage. If I use Casino dice against an opponent who doesn't have them, I've given myself an advantage simply because of the equipment I'm using. I'll probably find myself offering to let my opponent use my dice.

At the end of the day it comes down to how the game was designed. If the game was designed with lots of playtesting then it probably does not matter as the developers would have it in mind, even if it was just subconsciously considered. If the developers used theory rather than playtesting then it would have an impact.

The real goal of this article is to ensure that there are not cases of players abusing the different types of dice for their own game in competitive settings. So long as everybody is using the same dice it should all be alright in my opinion.

Yeah I've run into guys that split their dice.

At tournaments, I've had GW come over and give the guy a new set of dice (on me) and insist he play with it in his next game (game 5).

Just once, and it was years ago, but I always watch now.

Sad thing is, even cheating he still lost.

Ah well, they never do prosper.

This article is garbage.

His data prove that the 36 dice that he used for the test will roll a "1" 29% of the time.

He used a high # of rolls but the sample size is still just 36 dice of each type.

In addition the tests don't simulate the way people actually roll dice during games. Individual rolling? No thanks.


Suggestions:

First, the dice you use for the test should be taken from several different production lots.

Put 1000 dice in a big-ass cork-lined tumbler, tumble them for 0.5m and dump them out on a huge felt-covered piece of plywood and count the 1's after that. Repeat it 3 times.

Next, repeat the same test with 100 dice. Do 30 replicates (so your total # of rolls remains the same as with the 1000 dice test.)

Repeat with 10 dice. Do 300 replicates.

The data from these tests would take into account the effect that the dice have on each other while being rolled and poured out onto the table.

Now who's got 1000 dice?

Wow, my local scientist weighs in (hi, Brian).

One wonders what the details were of the studies which led casinos to the conclusion.

I like the article. I am a 7th grade math teacher and when we get to the chapter on probability we do die trials. Some of the dice are rolled over 800 times. I do the trials with all of my classes then we compile the data and analyze it. I too found that the small dice rolled poorly. I tried out several different types of dice & I found that the vehicle destruction dice from GW were the most fair. The small chessex dice were the worst. I have since retired the chessex dice and started using the larger sized GW dice & I love them! I can't wait to test them out when we get to probability next year.

Cheers,
rebreather

Brian, check out the original warseer thread where the guy posted his original findings, and you will see what you have suggested has already been debated there.

The one thing that makes it very clear that there is indeed a lack of randomness is simply that casinos do not permit the usage of these types of dice in accordance with their own studies. Though a greater sample size is always desirable, there are limits to what people with no funding can achieve, and I am glad someone even went this far for our hobby.

Brian's point is valid though. You can't start making conclusions from bad science just because it's hard to properly execute and fund good science.

1,000 rolls per die is decent though, so what this study does do pretty well (assuming the math that is left out is done properly, and the method used to roll was decent) is show that a given set of store-bought dice can perform non-randomly.

That part is fine, but to then go on making assumptions about all small dice, chessex dice, or GW dice is just bad science.

Edit: I tried the warseer thread and it died after the first page, so I'm not sure what went on there.

awesome i think i am gonna get some new dice

Apologies for thread necromancy...

But seeing this article piqued my interest, and led me to doing a quick test on my own; while not having students to send out to do my bidding, I took two sets of 16mm Chessex block D6s (for a total of 24 dice), threw them 25 times (using a typical game-room style dice throw; cupped all dice, shook for a bit before tossing down and slightly forward), and counted ones, just for a quick attempt at verifying this (or, at least, seeing the trend for myself).

Out of 600 throws, my Chessex drilled-pips rounded-corners dice yielded 97 ones. 100 would be the expected amount; 97 is close enough to be indistinguishable from fair.

So, with a block of 16mm dice, thrown like I would down at the game store, I couldn't see this trend. 600 throws isn't a huge amount, statistically, but the trend he's seeing, if it exists, should show up pretty obviously with 600 throws.

Now, throwing out experimental error (as I'm going to assert that 600 throws is enough that his claimed 29% 1s on Chessex dice *should* be noticeable), what I'm left with is:

1) The 12mm and 16mm dice roll differently; perhaps the pips drilled and corners rounded are different ratios of the overall volume of the two sized of dice.

2) We were using different batches of dice (although the Chessex dice I used were purchased around '05 if memory serves, not too far off from the time of the experiment)

3) Rolling the dice as a group introduces additional dynamics to the situation which make the dice "fair", and I might see the effect if I could be bothered to throw individual dice hundreds of times instead of a total of hundreds of dice thrown.

But, meh. Repeated experiment, failed to repeat results. Considering that I was following rolling procedure much closer to how the dice get rolled in the game, I'm calling shenanigans on the study.

Interesting stuff, I hadnt thought about your (3) but seems plausible.

lifeafter wrote:My reaction to this article was "wow, why would anyone make such a garbage product." (cheap dice, not the article!) I then proceeded to order some casino dice, but now a part of me is regretting that decision.

I began to wonder if a disadvantage shared by all is really a disadvantage. If I use Casino dice against an opponent who doesn't have them, I've given myself an advantage simply because of the equipment I'm using. I'll probably find myself offering to let my opponent use my dice.

Everyone using the same dice is not a disadvantage, only a longer game (if it's GW dice that is)

While the work this guy put into his study is impressive, in the end it is only a study of the 36 dice he analyzed.

(Also, as legoburner noted, using dice that roll an inordinate amount of 1s for morale checks and nothing else is cheating.)

Orkeosaurus wrote:While the work this guy put into his study is impressive, in the end it is only a study of the 36 dice he analyzed.

(Also, as legoburner noted, using dice that roll an inordinate amount of 1s for morale checks and nothing else is cheating.)

Well, the fact that both the GW dice and Chessex dice ended up with the same results seems (to me, at least) to indicate that it's something to do with that general style of dice (drilled pips, rounded corners).

The problem with the study is that it uses a fairly degenerate method for rolling the dice (and even then doesn't mention things like how far the dice travel on the throw, from how high up they're thrown, etc), and extrapolates to typical game usage from there. My mini-experiment seems to indicate that it doesn't follow.

Weren't the GW and Chessex dice rolled together though?

You can't say for certain it wasn't just the batch of Chessex that threw off the numbers, especially since it doesn't appear that he mentions how many GW dice he rolls.

Hmmm...

Afterwards we calculated the results and the Chessex and GW dice averaged 29% ones. Mind you that this is an average and our high was 33 and our low was 23.

You may be right; I read it as both sets of dice averaged to 29%, with high and low individual dice of 23-33. In either case there's a lack of, well, specificity here.

But it appears 23% ones was the lowest result he got for any of the GW or Chessex dice, so he's getting unfair results for both sets of dice, regardless.

We removed any statistical anomalies and came up with 29%.

This part bugs me.

Anyway, throw some dice, count the ones, see what happens. Doesn't take terribly long if you're throwing them by the batch and only counting ones.

My dice roll fairly well. Their corners aren't very rounded though.

alright, some people have noted that dice in the game setting may be dynamically different than dice in a non game setting. To facilitate this, I will be taking casino dice and my chessex dice and be rolling them simultaneously during 5 games next week. I will first roll the chessex dice, note how the dice land, then roll the casino dice, noting those. I will attempt to roll the same way for each "set" (set meaning a single roll of chessex and casino dice). I will then provide this in a nice big write up for you guys.

See any methodology problems? Anything to add, change? I know 5 games might be a little small, but any more than that and results might take a long time. this is meant to tackle the question of dynamics of dice within a game setting, and see if there is a visable difference. I think after 5 games it will be fairly clear per role how the dice do. Maybe I will include pictures and battle reps with each, make it mildly more interesting.

Grey Knight Luke wrote:alright, some people have noted that dice in the game setting may be dynamically different than dice in a non game setting. To facilitate this, I will be taking casino dice and my chessex dice and be rolling them simultaneously during 5 games next week. I will first roll the chessex dice, note how the dice land, then roll the casino dice, noting those. I will attempt to roll the same way for each "set" (set meaning a single roll of chessex and casino dice). I will then provide this in a nice big write up for you guys.

See any methodology problems? Anything to add, change? I know 5 games might be a little small, but any more than that and results might take a long time. this is meant to tackle the question of dynamics of dice within a game setting, and see if there is a visable difference. I think after 5 games it will be fairly clear per role how the dice do. Maybe I will include pictures and battle reps with each, make it mildly more interesting.

I hope you play orks...

Grey Knight Luke wrote:alright, some people have noted that dice in the game setting may be dynamically different than dice in a non game setting. To facilitate this, I will be taking casino dice and my chessex dice and be rolling them simultaneously during 5 games next week. I will first roll the chessex dice, note how the dice land, then roll the casino dice, noting those. I will attempt to roll the same way for each "set" (set meaning a single roll of chessex and casino dice). I will then provide this in a nice big write up for you guys.

See any methodology problems? Anything to add, change? I know 5 games might be a little small, but any more than that and results might take a long time. this is meant to tackle the question of dynamics of dice within a game setting, and see if there is a visable difference. I think after 5 games it will be fairly clear per role how the dice do. Maybe I will include pictures and battle reps with each, make it mildly more interesting.

Hrm.

Were I to do this, I think I would (for sanity's sake) use half of one type and half of the other for each throw, just to guarantee they're being thrown the same way.

Of course, I don't have casino dice, and after doing the quick test posted above I'm not convinced they're necessary.

This is no surprise to me, ive tested this things myself, with the sm scouts.

They where from the begining ultra marines and crappy painted, but when i repainted them into camo DA (u can find that picture here!
http://www.dakkadakka.com/gallery/51574-Scouts%20Love%20Camo%21.html)


i always rolled sixes on my
to-wound rolls. Nice isnt it guys?

Interesting Article and Kudos for getting other people (e.g Students) to do alot of that work for you!
I think by analysing this too much we have entirely missed the point. The dice is just a mechanism for generating a value, but the game is more important. Any percentage change will be relatively small (unless your using loaded dice) so really all of this analaysis of something like a dice is not required. If you want perfect results of 1/6 everytime forget it, we live in a dynamic world. Every little change in pressure, moist content, the surface the dice land on and myriad of other factors come into play. We acknowledge this in science all the time by not trying to take into account all of the variables, we assume that a certain things are of a perfect length or are in a perfect vacuum or are a uniform density or surface.
I am not trying to patronise or pick holes, but Dice like all things are flawed. They are not truly random, as per Chaos Theory "the deterministic nature of these systems does not make them predictable". I find it interesting that a value of 29% was returned (4% shy of 1/3!).
Remember those Dice Domes on certain games where you pressed the centre and the dice bounced around in a clear plastic dome but you never actually touched the dice?
Would it make your games more enjoyable if you felt like your dice were perfectly giving you a 1/6 chance regardless of the way you threw the dice?
As a Physics Graduate, i used to worry about big things i could never solve or understand, control or even see (like Quasars or Super Novas) then you just realise it means nothing. I will never know why we are here? These days I just enjoy how wonderful life is for us in the Western Hemisphere, I have a everything I need a enough money to indulge in a frivilous hobby. Dice well, I cannot say I am overly worried if roll too many 1's or 6's...

I've noticed that I tend to roll more 1s with Chessex dice playing 40k then I have with casino dice playing Warmachine over the past couple years. Guess I need to fork out for more Casino dice.

I remember seeing a similar article on Gamescience.com when they advertised their dice.

 Moz wrote:
Brian's point is valid though. You can't start making conclusions from bad science just because it's hard to properly execute and fund good science.

1,000 rolls per die is decent though, so what this study does do pretty well (assuming the math that is left out is done properly, and the method used to roll was decent) is show that a given set of store-bought dice can perform non-randomly.

That part is fine, but to then go on making assumptions about all small dice, chessex dice, or GW dice is just bad science.

Edit: I tried the warseer thread and it died after the first page, so I'm not sure what went on there.

yes, but 29% being significantly larger than an expected 16.6% it is hard to believe that other (similar) dice are any different. You are correct in saying that you can't make that same assumption about all dice. In fact the only thing that you can say with absolute confidence is that the dice that you tested roll non-randomly. But all of science is about testing a controlled group of things and applying that knowledge to the rest of the world.

Side note: we can all just avoid this argument entirely if in a game both you and your opponent use the same set of dice.

I didn't read this before, but one thing I have to note here with regards to gaming. In most cases if both are using same dice which throw more 1's the result of the game is not really affected. Why? Because for both players the effect is same. Meaning 1 is failure. Only in case where one side can't fail with one while other can (in same situation) would there be a difference.

What I'd really like to see (if someone were to do something similar) is to track all numbers. If the the distribution is not equal with other numbers then the difference becomes relevant depending on what army is being used.

Hena wrote:
I didn't read this before, but one thing I have to note here with regards to gaming. In most cases if both are using same dice which throw more 1's the result of the game is not really affected. Why? Because for both players the effect is same. Meaning 1 is failure. Only in case where one side can't fail with one while other can (in same situation) would there be a difference.

This is not true. For example, space marines pay extra points for being BS 4 instead of BS 3 like IG, and that point cost is done on the assumption that going from a 4+ to hit to a 3+ to hit gives you 33% more firepower. If, instead, the difference is less than 33% because a percentage of each number is going to the extra 1s then the marine player is overpaying for their models and at a relative disadvantage.