40 minutes ago, jagsba said:@punkUser : dice aren't perfect, you should share dice to make things more fair
my locals seeing the results: time to test my dice to make the best pool
the real question, what is the best pool? should we be picking pools that tie into our win condition? I wrote up an annoying thing here but whatever, you get the idea from this post TBH.
3 minutes ago, dadocollin said:First, this is awesome.
Second, it doesn't look like casino dice are super expensive. Sure, probably too expensive to put in a core set. But wouldn't it be in FFG's best interests after seeing your research to create a "tournament" set of dice that are manufactured to the same accuracy as casino sets?
The rig for a "Casino cut" d8 compared to a "Casino cut" d6 would be a bit more complex and fiddly (non-parallel and non-perpendicular surfaces require more setup and work than cubes do) with a fair bit more wastage of material.
1 hour ago, Muff2n said:You can still have unfair dice if you use a shared pool. E. G. Ones that roll focus more often plus a focus loving list! Or a low attack hit rate and a high defence dice list.
Absolutely. A list with many ships benefits more from average dice, a list with few arcdodging ships benefits a lot from extremely hot or cold dice (I'd argue more from hot, but only if they consistently arc dodge)
1 minute ago, Kieransi said:the real question, what is the best pool? should we be picking pools that tie into our win condition? I wrote up an annoying thing here but whatever, you get the idea from this post TBH.
if you check the mynock page, i believe you'll find that i wrote it.
2 minutes ago, Kieransi said:the real question, what is the best pool
Ideally? One consisting of those that best reflect the calculated statistical variance for a D8.
1 hour ago, Muff2n said:You can still have unfair dice if you use a shared pool. E. G. Ones that roll focus more often plus a focus loving list! Or a low attack hit rate and a high defence dice list.
Yeah, fly vcx or decimator and bring green dice that favor blanks. "Lets share a dice pool."
Also, sharing someone elses dice is scarier because you know your dice, not like OP, but you are comfortable with their performance. Someone elses dice can favor blanks and you have Imp Aces.
Only way is to get the most balanced dice by putting them through a ringer like the OP, and then tell your opponent that you swear your dice are super balanced and you both should play with your dice and see if they buy that, lol.
Edited by wurms
Just now, Hiemfire said:The rig for a "Casino cut" d8 compared to a "Casino cut" d6 would be a bit more complex and fiddly (non-parallel and non-perpendicular surfaces require more setup and work than cubes do) with a fair bit more wastage of material.
It seems like the easiest thing to do for complete accuracy is just have tournament devices with the app on them at the largest tournaments, but rolling the dice is what makes the game fun, so I'd hate to see it go away.
I'm afraid with the dice sharing people will complain that certain dice favor certain builds. Still, this is all fascinating.
2 minutes ago, Kieransi said:the real question, what is the best pool? should we be picking pools that tie into our win condition? I wrote up an annoying thing here but whatever, you get the idea from this post TBH.
you run off the assumption that people will be sharing dice, I believe the sentiment I heard was more along the lines of 'stack my dice'
It's certainly possible to construct a list that maximizes/minimizes the effects of dice bias if you know their distributions, but I think we can all agree that's a much less pronounced effect than what we have today with the variance seen on these dice. Using a small, shared pool still mitigates the vast majority of the issues and probably gets us to "good enough for X-Wing".
If people want to use house-provided dice (or sadly, the dice app) in cut games that's fine, but we've probably passed the point of diminishing returns there...
1 hour ago, Caduceus01 said:I would love to see this test extended to the various prize dice. The newest black Hyperspace dice most of all just because I feel like most of them I have seen roll terribly. I know mine do at least lol.
There's a brief discussion of that in the document. I considered it but it requires retraining the neural network since the colors are fairly different on a lot of them. It's certainly not impossible, but after a physical inspection I haven't found reason to believe they are any different in terms of manufacturing or tolerances than the regular dice, and indeed the vast majority of mass produced dice in general.
So possible future work but would want some plausible reason why they'd be different (different manufacturer or something) before putting in the additional work.
8 minutes ago, Hiemfire said:The rig for a "Casino cut" d8 compared to a "Casino cut" d6 would be a bit more complex and fiddly (non-parallel and non-perpendicular surfac es require more setup and work than cubes do) with a fair bit more wastage of material.
Not only more expensive, but if you look at what casino dice are meant to do... its meant that they don't roll. I saw an article that I can't find now (Maybe this was it? https://boardgamegeek.com/thread/642255/casino-dice-are-not-boardgaming ) about how you shouldn't use casino dice in place of D6's for board gaming because you don't have the thing that makes them most random: A large table with randomized pyramids on the other end to fully randomize rolls. Or I guess a randomizing machine as the OP built. Machined dice would also tear up play surfaces and we all love our collector mats.
But all of this IS fascinating! Guess the dice app will come in handy.
10 minutes ago, punkUser said:It's certainly possible to construct a list that maximizes/minimizes the effects of dice bias if you know their distributions, but I think we can all agree that's a much less pronounced effect than what we have today with the variance seen on these dice. Using a small, shared pool still mitigates the vast majority of the issues and probably gets us to "good enough for X-Wing".
If people want to use house-provided dice (or sadly, the dice app) in cut games that's fine, but we've probably passed the point of diminishing returns there...
The real answer is damage charts, **** dice
I played that speed x-wing rules someone posted, and the damage chart actually is really good and moved the focus to the more fun part of the game. Only bad thing is some abilities get ****** over.
Maybe a compromise where you roll a single die for certain procs. The skill tests in Outer Rim actually are really succinct as well, roll two ATK die but you only need 1 Focus/Hit/Crit depending on skill.
5 minutes ago, Jyico said:Not only more expensive, but if you look at what casino dice are meant to do... its meant that they don't roll. I saw an article that I can't find now (Maybe this was it? https://boardgamegeek.com/thread/642255/casino-dice-are-not-boardgaming ) about how you shouldn't use casino dice in place of D6's for board gaming because you don't have the thing that makes them most random: A large table with randomized pyramids on the other end to fully randomiz e rolls. Or I guess a randomizing machine as the OP built. Machined dice would also tear up play surfaces and we all love our collector mats.
So theory of rolling them aside (I agree rolling them down a big table and hitting the backboard is best!), it turns out they roll pretty well even in this sort of apparatus as well. If you saw in the methodology section, I used Casino Dice as one of several controls for the validation of the rolling mechanism and they produce uniformly fantastic results.
Just now, punkUser said:So theory of rolling them aside (I agree rolling them down a big table and hitting the backboard is best!), it turns out they roll pretty well even in this sort of apparatus as well. If you saw in the methodology section, I used Casino Dice as one of several controls for the validation of the rolling mechanism and they produce uniformly fantastic results.
I did and was highly impressed by it
Did the casino dice tear up your clear panel at all during testing? Or the wood?
I recently tested my dice using the salt water method... 24 green and red dice acquired from 3 core sets and several dice packs, and what I can say is... holy moly have I been losing a LOT of games due to GARBAGE dice. Out of my green dice, I had 3 that NEVER rolled paint and the rest were so poor, I only feel comfortable using 5 of the 24 greens. My reds were slightly better where 20 of them rolled acceptably for me to use them during games, but the 4 that are shelved each only rolled a focus once. (note: I rolled each die 100 times)
7 minutes ago, Jyico said:I did and was highly impressed by it
Was wondering how long it'd be before you commented on this! Did you ever get around to chopping open some 2.0 dice, and if so, are they any higher quality than the 1.0 ones you chopped up? I'm wondering if there's ever going to be an official production run of dice made that are pretty balanced so I don't have to weigh and mark mine if I want "normal" dice.
Edited by Kieransi4 minutes ago, Kehl_Aecea said:I recently tested my dice using the salt water method... 24 green and red dice acquired from 3 core sets and several dice packs, and what I can say is... holy moly have I been losing a LOT of games due to GARBAGE dice. Out of my green dice, I had 3 that NEVER rolled paint and the rest were so poor, I only feel comfortable using 5 of the 24 greens. My reds were slightly better where 20 of them rolled acceptably for me to use them during games, but the 4 that are shelved each only rolled a focus once. (note: I rolled each die 100 times)
Yeah, I tested a bunch of die and almost all of them were really biased toward blanks (as shown in gif)
Wonder how well salt water result correlate to actual roll performance.
17 minutes ago, Jyico said:I did and was highly impressed by it
Surprisingly didn't damage the acrylic at all. They did create a bit of sawdust after 20k rolls (which the 3d printed ones did too!), but it wasn't too bad and no visible damage. If they were being dropped from higher initial heights it probably would have been worse
5 minutes ago, Kieransi said:Was wondering how long it'd be before you commented on this! Did you ever get around to chopping open some 2.0 dice, and if so, are they any higher quality than the 1.0 ones you chopped up? I'm wondering if there's ever going to be an official production run of dice made that are pretty balanced so I don't have to weigh and mark mine if I want "normal" dice.
I never chopped any open, but someone here ages ago put them through some sort of MRI type machine and found the holes on the inside. Our group had one player that was pretty notorious for finding and punishing "bad" dice.
I gave away my "perfectly balanced set" to a friend who had dice problems (all of this is anecdotal so it doesn't matter in the grand scheme) but he did win some regional dice with them. I have no idea what happened to the set after that. I gave up and now have one of those eclectic sets of "one of each regional style".
At the last hyperspace tournament I attended, during the cut, my opponent asked to share dice. That was the first time anyone has asked me specifically. Before this article I used to think that was more of a mind trick, but the conclusions of the OP's article are great!
As a note, though shared dice pools are more fair than not -- the particular distribution may be favorable for certain lists or archetypes. For instance, a 5 green pool that favors blanks is a lot worse for a TIE swarm than it is for a dual VCX list. I can't see a world were people are using the app reliably as people seem to really enjoy the haptics of actually rolling.
Edited by Black_EightIt's a shame the dice app isn't free.
(or free with a code in the Core Set or whatever).
3 minutes ago, theBitterFig said:It's a shame the dice app isn't free.
(or free with a code in the Core Set or whatever).
Yeah. Not a huge fan of the app on a phone but a friend of mine (author of the awesome x-wing app Tabled ) brings a tablet with a stand and it works great, I could easily see doing dice that way.
1 hour ago, jagsba said:if you check the mynock page, i believe you'll find that i wrote it.
Kraytvented
3 hours ago, punkUser said:
Please tell me you just earned a PhD in randomness for this. 🤣
I have seen shoddier scientific studies that were supposed to be advancing medicine.
Well, I did expect mass produced dice to be a little biased, but not to this extent. Thank you for providing this info.
Edited by Koing907