Why in God's name does DCC require oddball dice, when i already own hundreds of dice?

Razor 007 · August 07, 2019, 07:20:34 PM

Quote from: Pat;1098680That means nothing, I just generated a few sets of 98 d7s (using a spreadsheet's RNG), and ran a chi-square test on the results. This is one of the results:

(Number of 1s thru 7s, respectively): 18, 13, 15, 17, 14, 15, 6

That's only six 7s, out of 98 rolls. Yet that set easily passes the 90% confidence level test (it's about 63%). If you rolled your die and got that set, we'd have to conclude there's no evidence of bias.

If you really want to test your die, run a chi-square test.

No, you just verified my initial results. We each got half or less of the expected 7 results on a d7. That is a strong coincidence. It's indicative of a flawed design.

Now, if your shorted number had not been 7; I'd agree with you.

Pat · August 07, 2019, 07:35:06 PM

Quote from: Razor 007;1098684No, you just verified my initial results. We each got half or less of the expected 7 results on a d7. That is a strong coincidence. It's indicative of a flawed design.

Now, if your shorted number had not been 7; I'd agree with you.

Absolutely not.

As I stated, I used a RNG, a random number generator. Not a physical die. And I wasn't testing the RNG to see if it was fair, I only used it to get a few sample sets that I could run through the chi-square test, to see what type of arrays come up as biased and which do not. So how I generated it is irrelevant, the array I provided with the 6 is not biased. Ergo, you rolling a physical d7 100 times and getting seven 7s doesn't indicate your die is biased.

Razor 007 · August 07, 2019, 07:38:10 PM

Quote from: Pat;1098687I used a RNG, a random number generator.

Evidently, a flawed one....

Just kidding.

GeekyBugle · August 07, 2019, 08:00:32 PM

Quote from: Pat;1098687Absolutely not.

As I stated, I used a RNG, a random number generator. Not a physical die. And I wasn't testing the RNG to see if it was fair, I only used it to get a few sample sets that I could run through the chi-square test, to see what type of arrays come up as biased and which do not. So how I generated it is irrelevant, the array I provided with the 6 is not biased. Ergo, you rolling a physical d7 100 times and getting seven 7s doesn't indicate your die is biased.

But does it pass the casino craps table test for a fair die?

Ratman_tf · August 07, 2019, 08:12:30 PM

I posted a link to this recently about X-Wing miniatures and biased dice.

https://docs.google.com/document/d/1alv05cXh0WhNaFPoq3LNzxmv6veurf6utV1kMQAYDQw/edit

What I took away from it-
Mass produced dice can, and likely do, have a bias large enough to affect an evening's worth of play.
The amount of bias is small (5ish percent) and varies from dice to dice. One dice might favor a "hit" more, and another dice might favor a "blank" more.
The amount of testing to detect bias is far beyond what I'm willing to do (tens of thousands of rolls) to maybe find a dice that rolls 20's, 2% more likely.
Mix up your dice during a game, so that if you do have biased dice, it isn't the same bias over the course of an evening.
Which fits nicely into many gamer's dice superstitions about switching out dice that roll "cold".

GeekyBugle · August 07, 2019, 08:26:50 PM

Quote from: Ratman_tf;1098692I posted a link to this recently about X-Wing miniatures and biased dice.

https://docs.google.com/document/d/1alv05cXh0WhNaFPoq3LNzxmv6veurf6utV1kMQAYDQw/edit

What I took away from it-
Mass produced dice can, and likely do, have a bias large enough to affect an evening's worth of play.
The amount of bias is small (5ish percent) and varies from dice to dice. One dice might favor a "hit" more, and another dice might favor a "blank" more.
The amount of testing to detect bias is far beyond what I'm willing to do (tens of thousands of rolls) to maybe find a dice that rolls 20's, 2% more likely.
Mix up your dice during a game, so that if you do have biased dice, it isn't the same bias over the course of an evening.
Which fits nicely into many gamer's dice superstitions about switching out dice that roll "cold".

If I remember correctly there was a company called GameScience selling what they claimed were fair dice for TTRPGs.

Pat · August 07, 2019, 09:13:53 PM

Quote from: GeekyBugle;1098691But does it pass the casino craps table test for a fair die?

All we have is one face's worth of results for a tiny number of rolls, we can't really say whether the die is fair or not. All we can say is there's no evidence that it's unfair.

Quote from: GeekyBugle;1098696If I remember correctly there was a company called GameScience selling what they claimed were fair dice for TTRPGs.

With sharp edges.

If you get enough of their d4s, you have an effective pursuit deterrent.

GeekyBugle · August 07, 2019, 09:23:39 PM

Quote from: Pat;1098704snip

With sharp edges.

If you get enough of their d4s, you have an effective pursuit deterrent.

Ouch! Well then they aren't getting any of my money, not worth it for what? a 2-5% bias tops?

Pat · August 07, 2019, 09:29:18 PM

Quote from: GeekyBugle;1098708Ouch! Well then they aren't getting any of my money, not worth it for what? a 2-5% bias tops?

Regular dice can be pretty biased, the manufacturing processes can leave extrusions and bubbles, and rounding the edges makes it worse.

But people are awful at judging bias. You could have hideously biased dice, and never notice. Or completely fair dice, and because of selective memory, you think they're lucky or jinxed. Unless you rigorously test your dice, with no fudging, you'll never know. And that's actually quite a bit of work.

It's not worth worrying about that much.

GeekyBugle · August 07, 2019, 10:10:23 PM

Quote from: Pat;1098710Regular dice can be pretty biased, the manufacturing processes can leave extrusions and bubbles, and rounding the edges makes it worse.

But people are awful at judging bias. You could have hideously biased dice, and never notice. Or completely fair dice, and because of selective memory, you think they're lucky or jinxed. Unless you rigorously test your dice, with no fudging, you'll never know. And that's actually quite a bit of work.

It's not worth worrying about that much.

I happen to know a little bit about mold making and plastic injection, you can make your dice rounded from the start, IF we could make it here in México with regular mold making processes, then in the USA with cutting-edge tech there's no excuse.

About making them fair, this is also solved from the mold design fase, what you need is to have each side weight the same and to have several really thin injection cones.

Ratman_tf · August 07, 2019, 10:12:49 PM

Quote from: GeekyBugle;1098708Ouch! Well then they aren't getting any of my money, not worth it for what? a 2-5% bias tops?

The sharp edges are part of what makes them fair. So yea, pick either biased dice (small bias at that) or sharp dice.

GeekyBugle · August 07, 2019, 10:15:49 PM

Quote from: Ratman_tf;1098718The sharp edges are part of what makes them fair. So yea, pick either biased dice (small bias at that) or sharp dice.

It's the easiest way to make them fair, not the only one, again it's about precision mold making.

Philotomy Jurament · August 08, 2019, 03:26:09 AM

Way late to this party, and I admit I haven't read the thread, but the funky DCC dice always struck me as a gimmick. Personally, I could do without it, but I don't consider it that big a deal (not a deal breaker, by any means). You can generate the appropriate random numbers with regular dice without difficulty.

Vladar · August 09, 2019, 11:00:55 AM

Considering how hard it was for me to acquire even "standard" polyhedral dice, I must admit that non-standard dice is a pretty big cons for me, though the game itself looks pretty solid.

Brad · August 09, 2019, 11:55:28 AM

Quote from: Philotomy Jurament;1098750Way late to this party, and I admit I haven't read the thread, but the funky DCC dice always struck me as a gimmick. Personally, I could do without it, but I don't consider it that big a deal (not a deal breaker, by any means). You can generate the appropriate random numbers with regular dice without difficulty.

Way late as well, and they are a gimmick, but a fun one. The fact that the app roller exists as a free download sort of mitigates this issue. I prefer actually rolling dice, though, so I finally bought a set from Amazon, think they were $15. Or as you stated, just get a chart and use "regular" dice.

I find this entire argument ridiculous, anyway, as if you bought a boardgame that required some sort of paraphernalia not included (and clearly stated on the box), you'd buy it to play the game.