Dice-O-Matic hopper and elevator (2009) | Hacker News Reader

Dice-O-Matic hopper and elevator (2009)

gamesbyemail.com

246 pointspavel_lishin55 comments

When I play RISK with my kids, the time invested in each roll of the dice is proportional to what's at stake. During a critical close battle, they might shake, blow, and otherwise cajole the the dice for a minute before rolling. Playing on an iPad with its RNG loses all this, so it's very pleasing to think of playing online with real dice.

They should definitely add a fan or compressed air to simulate blowing on the dice. For accuracy, of course.

Premium feature: reserve a set of lucky dice for your roles. $5/month

A few years ago, I found a WebGL physics-based dice roller.

http://a.teall.info/dice/

I just made it throw a few hundred dice in one roll, and chrome promptly packed it's bags on me

It works pretty well on my side, but I get stuck dices that never land

Can you add a table edge and some piece of furniture they can roll under?

I assume for critical rolls (edit:roles) you also sacrifice small animals? Just as effective as talking to the dice!

Cute mechanism, but only running it part time and storing random numbers for later use is asking for a security breach. If you can find out what random numbers are coming up, you win.

Vibratory bowl feeders pretty solve the problem of getting simple objects lined up.[1] Any object that isn't lined up properly gets dropped off the ramps back into the bowl for another try.

[1] https://www.youtube.com/watch?v=Mejn0n4IslY

Agreed, but it sounds like this project was for fun and to make users happy, not to actually guarantee true randomness.

I would wager the amount of wear on the dice, and small flaws in each individual die, produce a non negligible bias to the rolls that could be avoided by using cryptographicly secure random sources.

If they start using these dice to generate cryptographic keys, I'll run for the hills, but so long as they are only used for playing board games over the internet, this is really cool.

> If they start using these dice to generate cryptographic keys, I'll run for the hills, but so long as they are only used for playing board games over the internet, this is really cool.

This just gave me a flashback to making cryptographic keys long ago, and adding in entropy by typing randomly on the keyboard.

What a time to be alive.

> If you can find out what random numbers are coming up, you win.

So an adversary can breach the system and read the upcoming results but they can't mutate the results? That seems like an unlikely scenario.

Also the security implications of this are that some jerk cheats at a board game. Hardly something to lose sleep over.

> Hardly something to lose sleep over.

Easy to write that now, but let's see how you feel the next time your armies in Kamchatka are being invaded.

Oh, I thought this was an online gaming system with real money. Apparently not.

Spoilers: he doesn't actually use the output of the machine to power his site. Not that he couldn't...

This pops up on HN or reddit every 3 or 4 years. I want to say he made it in early 2000s? I was in the cube next to him at the time... my only brush with greatness.

Before each day he generates a pseudo-random shuffle of the physical dice rolls ;)

This is great. Reminds me of the Lego brick identifier/sorter that was posted a while ago. Sure, there's no real reason that true-random bits from Random.org couldn't be used, but I think a project like this is neat precisely because it takes such great strides to bring back a classic physical component of playing board games that the service is otherwise designed to eliminate.

A wonderful machine, I can imagine the sound of it two rooms away.

So now as an engineer and semi-pro backgammon player, the ultimate dice were ones that had the divots drill out and replaced with a different colored plastic. This way the uneven weight of the dice was not a factor. Just wondering...

Has it ever been shown that dice not made as you describe are non-random?

The manufacturers represented them as less biases. I do not have any statistics to back it up.

But it does make sense that the removal of any material would effect the rolling unless very sophisticated things were done to offset the remove of material.

Really? I've never heard of this, and I live in Vegas where fair dice are considered pretty important.

The pips on opposing sides always add up to 7, and each of the faces are balanced (1 is in the center, 2 is on opposing corners, 3 is a line, 4 is all the corners, and 5 is the corners and the center, 6 is two rows).

At least in craps, what the casinos worry about are fake dice, and not hitting the back wall with the dice. The back bumpy wall introduces a lot of randomness. Some people try to put the dice in various initial positions and roll them so that they only roll over one axis, to try to influence the rolls. The casino will throw you out if you don't hit the back wall repeatedly.

Also, over time the sharp corners of the dice get ground down just by friction. Uneven wear is also a problem, which is why they frequently switch out the dice.

The I want to say (counter sunk) dice always have rounded corners too. I think the cost might outweigh the randomness benefit for Vegas.

Probably. Even the extra security features probably make them cost a lot more than cheap dice, and there's a lot more QA like spinning them on various axes to make sure they are balanced.

They make up a bit of that on the backend though, selling the dice to all the tourists. :)

Found what I was looking for!

Precision Dice. http://www.bgshop.com/index-backgammon-dice.htm

They seem to be a backgammon thing.

Ah thanks, very interesting. This part was telling:

"Precision dice with rounded (or cut) corners are recommended where a limited rolling area is available, the cut corners allow the dice to tumble, in particular the smallest size of 1/2 inch (13 mm) are becoming increasingly used by many RPG gamers."

I didn't think about backgammon being a small dice rolling area, but that makes total sense.

If the divots are a different color plastic then they have a different weight than the rest of the die as they contain different amounts of carbon black, camel white etc.

True, but that difference is insignificant compared to no material at all. Plus if you were REALLY anal, things could be done to exactly match the density.

That's how casino dice work right?

I think Casino dice just have to be good enough to meet standards, cost vs benefit. See my Precision Dice post below.

Can't be nearly as bad as any pachinko parlor that I've been near. It sounds like a machine shop full of circular saws.

I harbor the belief (likely unjustified) that the hearing loss resulting in hanging out there is as bad as gambling addiction issues. I don't know how people stand it, or why they'd want to.

https://www.youtube.com/watch?v=wHfocxjWbZI

The bright lights and noise levels in casinos are deliberate as they impair rational thought. Study after study on "cognitive load" show that if your brain is being stimulated or worked in one domain it impairs good decision making in other domains.

I'm sure the operators of these places are aware of that too.

What's funny is that with USA casinos, the noise level is loud, but not bothersome to me personally. AFAIK, this is intended to create a "lively" atmosphere.

Whereas the pachinko parlors are far beyond the "lively" level, and well into the "where is my OSHA-approved hearing protection?" level.

> I have used Math.random, Random.org and other sources, but have always received numerous complaints that the dice are not random enough. Some players have put more effort into statistical analysis of the rolls than they put into their doctoral dissertation.

I seriously doubt that high-quality electronic randomness is non-random enough to have a noticeable effect on the outcome of board games. It's nice that the guy is accommodating enough to go to all this effort, but it seems unnecessary. Cool project, though.

I think you can assume he's in on that joke. The goal of this was to build a fun machine, not to actually increase entropy.

The fact that he offers to melt and mail a die to any player who complains of non-randomness is a giveaway he's having fun.

A similar story I recall was that shuffling a deck (by hand) vs randomly sorting cards. It's more likely that the values were TOO random and resulted in a less enjoyable experience (or the player wanted to blame someone other than themselves).

I went looking for a video I recalled about a craps player able to choose what he'd roll. I didn't fint it but I did find: https://www.insidescience.org/news/dice-rolls-are-not-comple...

Bridge is particularly prone to being affected by poor shuffling, for two reasons. Firstly, cards of the same suit are clumped together as part of each trick, and if they are not separated during the shuffle, get evenly distributed back to the players. Secondly, suit distribution is important in bridge - it's normal to describe a hand by the amount of each suit it has.

Players noticed that voids and singletons were more common when playing online or with computer-generated hands. Unlike the board gamers of the article and everyone else who complains about bad RNGs, they had noticed a real effect, but the effect was backwards - the computer hands were better shuffled and more random.

humans have an odd relationship with randomness. iTunes had to de-randomize their shuffle system, because it was too random (i.e. you could get the same song back to back etc, purely randomly). We tend to see patterns that just aren't there. I'm not surprised his player complained.

[1] similar article about spotify: https://www.quora.com/Is-Spotifys-Shuffle-feature-truly-rand...

Insufficient randomness from random.org?

Odd. Seems like the only way it wouldn't be random is if the code transforming the output into die rolls was wrong, or if the inherent unfairness of dice with pips carved out of them is truly desired (which, it seems, could still be modeled from a truly random source).

A cool machine though. I was expecting the captures to take place somewhere other than the lifting chain though, but it makes sense for the setup.

I think the insufficient "randomness" is explained here: https://news.ycombinator.com/item?id=14805265

You're probably right, it's probably more of a psychological issue than an issue of randomness. Perhaps his users simply trust dice more than they do "atmospheric noise with a certification of randomness across sliding windows", even though minor imperfections in dice (and dug-out pips are more than minor imperfections) will cause them to be much less fair.

That said, it would be really interesting to see the referenced thesis-like papers. Or use high quality Vegas dice for awhile and compare the outcome of those to the purely random data.

That's discussing a different dice project - what's the connection to this one?

The psychological effect of people not "believing" true randomness is the same.

I think there might even be a chance that there is bias in the die-recognition system, so random.org would be in fact more random.

I've been playing on gamesbyemail.com for over 12 years. It's amazing what Scott implemented early on in the AJAX era. I mainly play WW2 the Axis & Allies game. Amazing job. Great fun.

He had a lego dice roller before this.

Amazing site.

Thanks for all the hard work on this.

This is awesome. Seems you could make it a business and sell some random bits. Maybe compete with hotbits and random.org.

How could rolling dice be "more random" by any meaningful definition of randomness than random.org? Rolling dice is a pretty trivial physics problem with really hard to observe parameters... parameters that the machine reduces the dimensionality of considerably.

The Internet of Things is amazing.

If you need a really random number, NIST can provide one free of charge:

https://beacon.nist.gov/home

Discussed in 2009: https://news.ycombinator.com/item?id=626092.

The 10fps potatocam footage made me wonder.

What an impressive device! Does it retain statistics keyed to a particular die? So that you could possible identify a 'loaded' cube which sneaked in?

It would be interesting to silently switch back to a PRNG for a few months and see if anyone notices.

It would be cool if the online game showed you the actual photo of your dice roll.

It's magnificent!

Reply on news.ycombinator.com