Were Squig Dice Balanced?
Vložit
- čas přidán 5. 09. 2024
- Sometimes you just have to find the answer to a question.
Oh and when talking about how close the results were from the average it should be 0.0017 rather than 0.0027. It doesn’t affect any of the results, i just messed up there. You may have noticed i’m very out of my depth here lol
Find us elsewhere!
🔵 Patreon: / snipeandwib
🟣 Merch: my-store-be810...
🔵 Discord: / discord
🟣 BlueSky: bsky.app/profi... / bsky.app/profi...
🔵 Twitch: / snipeandwib
The irony here is that rolling less 5's makes these the worst possible dice you can use for the various orc, ork, goblin and gloomspite armies because 5+ tends to be your armour save (or Waaggh invuln).
*fewer
That was also my first thought, but I wondered if the higher than average chance of rolling 6 might balance that out. Sure, that would give them an advantage on triggering rules that require rolling a 6, but only a slight one.
Not to mention to-hit rolls.
But absolutely perfect for the 3+ Stunty dodge rolls for Blood Bowl Goblins 😂
Or Space Marines. Hit on 3+, 3+ armor saves...
NEW SQUIG LORE JUST DROPPED! Squigs instinctively distrust and avoid fives of anything, but like threes.
That's why they bite off fingers!!!
@@Phillipbain that and their crunchy
🤣🤣😂
This is because of their chewing habits. If it takes more than 3 chews to eat something then it isn't good for them to eat, as such they have grown suspicious of anything numbering higher than 3, especially 5 since typically a stikk bomb takes 5 chews before it blows up.
These are the best dice Gdubs ever made. My friend bought 4 pairs of them for his orcs and whenever he rolled them all he would make sure to drop them from an appropriate enough height to cause absolute chaos!
Surprised you haven’t been swamped by comments, love what you do mate
That's the only proper way to play Orks.
Ugh spork
They desperately need to rerelease them
Thanos, with his squig Dice Rolling 3.498: "Perfectly ballanced, as all things must be."
I kinda love the fact that the orkiest dice are so unreliable but still somehow work as intended. Gw tapping into that waaargh energy with enough orks believing it works so it works.
I asked my friend to attach them to bases and paint them and now I have twenty more squig herd. :)
I think I've seen more squig dice used as squig models than as actual dice.
Minecraft Squigs lmao
This is the most niche topic and I love it. And the average being so close to 3.5 is kind of wild.
Glad you enjoyed it, and yeah, I very much did not expect it!
who knew that mathmatical averages on dice could be so entertaining. The irony that they are on ork dice does not escape me
Orks have always been a deeply math army. How can you improve the odds of an unlikely event? Thats right, use the simple maxim of "More Dakka"
Suddenly have an urge to have a sticker of an Ork explaining rocket science 😁
@@yeturs69420 Math just works for Orks because they believe it works
Id like to think squig dice are literally live squigs bred in the shape of dice
Hey, breeding your livestock into cuboid shape is an effective (if unethical) method for easy storage and transport. That is, unless the walls of your cargo hold are sharply jutting from every direction (as orc structures tend to be internally).
They've done it with vegetables in real life, im sure the orks would figure out how to do it with squigs
They put baby squigs (larva? how do those things reproduce?) in those little plexiglass boxes like melons
@@RadicalGarry😅😅😅
@@VisonsofFalseTruths to answer your "how do they reproduce" question, orks, grots/gretchins and squigs are all literally mushrooms so they all grow from spores.
I’m reminded of a joke . Three statisticians were out hunting. The first missed the deer by 3 feet to the right. The second missed by 3 feet to the left. The third explained “we got em!”
This video is like a perfect demonstration of how to present a scientific study, including introductory info and question, methods, results, (a very good) discussion, and a conclusion. I might share this with my biology students. This is excellent.
This is the type of investigative gaming journalism that we want to see!
Hell yes, roll over Coffezilla, a new player is in town! :)
As someone who took 1st place in my High School's Math Fair two years in a row with dice probability testing projects that each involved thousands of rolls, I fully endorse this video.
Like a plucked flower, I roll and enjoy my squig dice knowing they will fade away.
Truly poetic! Godspeed
not to be that guy but, can you not just repaint the faces? >.>"
These dice would be my absolute kryptonite. See, back in 4th/5th edition, I ran an uncommonly shooty ork army. Shoota Boyz and Lootas were just so good then, so I leaned into it. More fun than just jamming nob bikers down everyones throat. Thing was, I discovered that my luck tended toward extreme ends of the scale, producing alot of rolls of 1,2, 5 and 6, and very little else. Mathematically, it averaged out (like these dice ironically), but when you're chucking literal buckets of dice and getting 10-15% better results than you should, you turn people into Swiss cheese.
These dice would undo all that. And the worst thing is, I should love them. The erratic bouncing, the pure chaos of it (and not the armored spiky kind), but had they been available then, they very well could've been the death of everything I'd build and painted in that mad month I spent locked in my basement.
I'm an unprofessional mathematician (but not a statistician), and this video was just dandy imho.
Your experiment design was absolutely fit for purpose and I enjoyed your conclusions, especially about the mean just being one part of a more complex story and your analysis of the validity of rolling dice in batches.
Your sample size was more than fine, the chi square test showed that. I think people are spoilt by large sample sizes these days, and in some cases even misled (choosing large sample sizes can be an example of p-hacking.)
I wouldn't bother with control dice, either, unless you can find some precision dice that have been pre-tested and certified somehow, and even then what you'd really be testing is how your rolling method appears to affect those particular dice compared to how they were originally tested.
My one almost-gripe was that I thought you could maybe have added confidence intervals to your frequencies of dice rolls, but then you handed over to your mate Matt for his Pearson Chi-squared test and I was more than satisfied, also because that's a better choice than confidence intervals.
Squigs for the d-squig, Meat Cubes for the Meat Cube Throne.
They could make great minis or mini accessories if you don’t want to use them as dice. Maybe as actual meat cubes in a dystopian food production kind of way
I remember when these were released and thought they were super weird. Did not even know they were rubber, thought they were solid plastic.
As a mathematician, I think you did a pretty good job. Surprisingly entertaining for a video about dice
Unexpected but delightful! I just have one question: did the bouncy novelty factor wear off after rolling the dice so many times, and if so, how long did it take?
Shockingly no. I'm still very amused by them hah
@@SnipeandWib Sometimes it's just the simple things in life
@@SnipeandWib by far the most important finding.
This is the sort of scientific dedication and rigour that deserves to be nominated for an Ig-Nobel prize.
Better or worse than the Idoneth dice that were both perfectly transparent AND had soft-coloured abstract markings on the sides?
When I was in school, I didn't go to my Statistics math class most of the time (I had all my math credits needed, but had to take a math class) because I didn't think I'd ever use statistics in an important way during my life.
You've proven me wrong. This was important.
if you went to the statistics class you would know he should have done significance tests!
My friend has these dice. He really liked them and has been using them to this day. The fading numbers part is very true.
I wonder if painting a layer of varnish over the "pips" would help with the paint rubbing issue? Anybody know how varnish and rubber interact??
Ah, but would the weight of the varnish slightly skew the probability in favour of the sides with more markings on?
@Blacknight8850 not if you just cover the whole dice in varnish.
@@RipOffProductionsLLC well hold on now, the sides each being sculpted differently would therefore have differing surface areas and recesses that would hold onto more varnish! Clearly much more research is needed
As my boss used to say "There are lies, damned lies and statistics"
Your boss was William?
This is the sort of keen excel-based data analysis I yearn for. Thank you for your service.
Do you remember the weird nurgle dice that came out prior to these? They had a weird lighter plastic used as a cage for an inner foam that made for a weird weight feeling to them, and ive always been curious were they actually balanced either, but never enough to have tested them
All I know is that those dice were the most horrible things to read out of any dice I've ever played against xD
My man really rolled these 3000 times? subbed.
I didn't think I needed this video this morning, but now, after a gruelling day at work, suddenly this video has made all the difference between a bad day and a good evening. Thank you for your mathematical exploration!
Love the Space Truckers reference in a video of novelty dice! Could do another video on other ones, like those Lumineth Realm Lords or Skaven dice?
The Space Truckers fandom has been crying out to be recognised and I knew I had to be the one to do it :p
And I'd be down to do some of the other novelty dice but I don't have any of the others sadly, hopefully they're not as expensive to pick up nowadays as the squig nice.
@SnipeandWib I've had a quick eyeball for dice, and those Lumineth ones aren't too bad price wise.
The Skaven black burrow ones though... yikes!
Use squig dice if you like getting on your hands and knees and going under the table to look for them many times a match
You went above and beyond with this one!!
I just want to say, thank you for your C64 Midi sound scrolls at the end, always a treat!
To see how accurate your experiment is you should recreate it with regular dice and compare the measured results of each
I love how Russian meatcubes became so popular that they permeated other scenes
Interesting video, I reacently met someone when playing a game that had a set of these dice. They are great fun, but in terms of rolling, they were off. I am liking the Beast Wars Megatron on your desk as well, nice to see :-)
I love those dice. Got 2 sets of them and I use them every time I send my squig army into battle.
Wait - I can have a whole army of entirely squigs?!
@@byrongehrig8429 Yep :) Well, there's gonna be a couple goblins, either to "direct" or ride the squigs, but yes, you can. And yes, you're limited yourself quite a bit, but there's still a good choice of units. Loonboss on Giant Cave Squig, Squigboss, Squig Herd, Squig Hoppers, Boingrot Bounders, Mangler Squigs, even Colossal Squig and Squigbobba, if you're okay with ForgeWorld units.
Got them at a narrative event and used them once. They’re impossible to read and ricochet across the club. They got retired after 1 round.
From a stats perspective one way you could do this is to bootstrap it. You only have to roll 100 times for example then use the site stat key. To create a bootstrap, the site can generate more samples based on the data given. Law of large numbers suggests that the more samples you have the closer to the mean you’ll get as outliers cancel each other out from either side of the range. So with a potential 2k more samples you could be looking at even closer rates to normal dice
Wow this is the closest I’ve come to using what I learned in my stats class and it’s for squig dice
Not very useful in this case to bootstrap considering the comparison to a fair dice is trivial. It might have saved him some time, but the results would be the same regardless, and by not using bootstrap you avoid (mostly) the very annoying question of "Would throwing a bunch of dice together at a time make the throws not independent?"
What your friend did was a Chi-Square test! As a someone currently studying stats I was pleasantly surprised to see it pop up.
I have a statistic subject for uni coming up, and listening to this made it easier to understand, so thank you. 😂
Granted, I have not rolled them 3000+ times and recorded data, but I distinctly recall similar concerns over the Nurgle Dice with the plastic cube around the rubbery center. Anecdotally, however, they seem to give me much more range than standard dice, which seem to favor 1's and 2's...
This was one of the most interesting videos you guys have ever made
I own these and they’re great. Have had tons of fun bouncing them around the living room. Never used them in a game.
I feel like their best use is as Squig minis.
The smart thing to do to save the paint is put them in a board game dice popper, in my opinion
So just to add a tuppenceworth to your statistical analysis...
Your friend is right that if in an A/B test one cell produced 500/3000 positives (hypothetical fair dice) and one cell produced 426/3000 positives (observed squig dice) this would have a confidence level of 99.59, e.g. random chance would generate this result less than 0.5% of the time as you say...
That doesn't take into account the fact you were effectively running 6 A/B tests in parallel. This affects the significance of your result.
To offer an analogy - the chance of rolling a 6 is 1/6. But if you roll 6 dice the chance of at least one of them showing a 6 is far higher.
Similarly if an outcome has a probability of
Fun fact the opposing sides of a die always add up to 7. 3&4 are opposites, 1&6 are opposites, and 2&5 are opposites.
I think this test really needs a control. Let's see what happens when you roll a set of normal GW dice 3000 times and compare the results. The results show that these dice are good enough to use but the material is so delicate that their balance could change on a whim. Rubber will morph from heat or squeezing them. I also would like to see if the saline test works on rubber dice. Instead of rolling thousands of times get a cup of water and completely saturate it in salt. This will make the dice float. If it has any bias the dice will turn so that it's heaviest side is down because of gravity. A fair dice will stay on whatever face you place it on.
Ive been to a few AoS GTs where people use them.
I completely respect it, because they are funny.
I have these dice, and went to a tournament with my Squig army a few years back. I brought the dice, and, when other competitors asked if I had squig dice and I said yes, they were delighted. They're wonderful dice, and fits squigs perfectly. Shame GW don't do stuff like this anymore, and just go for bland, boring dice with pips and a symbol on the 6.
That doesn't actually look too bad, I once tested my own GW dice pool (~50 dice) by rolling the whole pile a few thousand times and ended up with side biases in the range of 15.7% to 17.8%.
Specifically:
1 - 15.71%
2 - 17.03%
3 - 16.79%
4 - 16.27%
5 - 16.37%
6 - 17.83%
I rolled enough times that the confidence interval on all of them was pretty good. My theory was that there's probably a few very biased dice swaying the averages of the entire pool, but I never had the time to evaluate them to that level.
Our dm is a huge 40k nerd, so we bought him those for birthday. He ended up using them as, well, squig minis - orders of magnitude cheaper then actual squigs.
Buddy, I swear I nearly had an anyeurism when I thought you weren't going to measure the frequencies...you saved me
A float test would be a good comparison too, and quite quick to do
5:25 This actually isn't an issue. Sure, if you rolled a million it would be a more confident result, but the stats that your friend and I calculated independently take into account how many trials were done. So the results stands as what it is. Also 3000 is a bug enough sample size to satisfy just about anyone statistically literate.
I'm still caught up giggling at squigs have a "mean average roll" xD Thank you for that, Wib!
I would love to see next the death guard rubber pustule dice tested!
Only this channel can make such a random video and get me to click so fast - amazing
1:12 I think bouncy dice sound kind of like an improvement, since regular dice just fail to roll sometimes
If you had dice that was equally likely to roll 3 or 4 but would never land on 1,2,5 or 6, you'd still get an average of 3.5.
I guess you’re just going to have to do another video where you roll a standard set of D6s several thousand times to create a baseline. You know, so you can use it to reference future videos rolling different types of dice, several thousand times each. But really, this was a fun watch, and I wouldn’t be upset if this became the start of a series.
I wonder if the place has any effect since rubber may change dynamics due to atmospheric pressure for example, humidity maybe? But this is very interesting, those are very different and interesting, a coat of something may protect the paint but you are into Warhammer you know about that more than me hehe, thanks for sharing this, greetings from Colombia
A lot of the specialty dice from that era are too weird but I love them all none the less.
I wonder if having more mass on the sides with more details (5 and 6, having more paint AND raised edges) leads to a higher likelihood of being the face on the bottom.
I knew the 7000 times joke was gonna come so I was like oh, but it still got me when it landed 😂
The bounce factor makes these even more chaotic, and the lack of weight in the material (that being rubber) adds to that even more while making weight altogether a lesser factor in reducing their randomness (which means all those little uneven bumps on them don't matter as much).
If these had this design but were plastic I think they'd be far more biased, this chaotic quality to the rubber makes sure they're as random as possible.
As someone quite interested in statistics for my day job, this was lovely.
With normal dice, you can test their balance using a cup of water. Put it in the water and spin it. If it is imbalanced, it will always float with the same side up.
Might work with these too.
Problem: it is made of rubber and have lots of uneven surface, meaning the float test wont tell you to much
@@SioxerNikitaI also never trust opaque dice as they could have bubbles inside and you'd never know, fucking up the roll.
Always buy clear.
Perfect for wound marker dice, for a use without having to roll them
That study was necessary, thank you.
Cool video. I'm not big into dice or tabletop games but this was an interesting watch
I've seen conversions to make them squig herds, which was really cute.
Warhammer gamer answer: Then we paint it good as new.
You don’t need a control group for things like dice. Think of it this way; a D6 die isn’t an object with potential unknown properties. It is specifically any object that has an equal chance to return one of six things. This is why we consider Mesopotamian “gaming sticks” to be D4s. You aren’t testing how much like “real dice” these are, there’s no such thing, you’re testing how well these real objects align with the definition of dice.
As you said ordinary dices will have deviation too.
It would be interesting to see if the deviation is much worse for these so control group is not useless.
Also allows to control for other factors such as the matt.
These dice somehow made me think of "super meat ball" games.
Meat cube... i never knew i needed 800 of them like the watermelon guy from the math questions
Hmm. Wonder if you could paint a clear coating on the Indicator-Paint bits so that color doesn't rub off.....
I haven't played a game of age of sigmar in a while but during the years I did I always used them for any squig attacks. The bounciness wasn't all that bad, I think I've dropped far more standard dice on the gaming store floor than I ever did squig dice.
These dice remind me of the time when Bricky at Adric said "fun over accuracy".
Are they unbalanced?
Yes.
Are they questionable?
Also yes.
Are they fun, when playing with friends?
Absolutely yes!
wrote a statistics exam last week, so, this was kinda fun for me 🤣
but the average makes sense if the dice favor 3 and 4, since the average should be between those two
Havn't watched the video yet, but those dice are pretty creative in the detail the numbers you roll.
3:15 Tiny correction xP
They deviate by 0.0017 from the standard 3.5 mean average roll. 0.0027 would put their mean up to 3.51, which would imply they're more likely to roll higher, than lower.
But then I've never rolled thee dice before (let alone knew they existed) so who am I to argue with the guy who rolled them 3000 times xD
This is important work. This man deserves our respect.
Now someone needs to make a d12 dice set for the chaos gods that massively favors that gods number while still averaging the exact same as a perfectly balanced dice set.
Given the insane amount of time you put into this question I'm a little scared to ask: at any point did you try... just floating one in a bowl of salt water and giving it a poke to see if it favours any particular sides?
I think it would have been faster to just build a machine like casino dice manufacturers use to test the dice lmao
3:47 Ran some stats for you. Calculated a Chi-squared statistic for you from the numbers at right. Chi-square = 25.164. With 5 degrees of freedom, that gives a p-value of 0.0001295. So assuming no data entry or reading errors, there is approximately a 1 in 7,700 chance of getting this result if the dice are fair/balanced. That is way below the typical significance level of 1/20 at which most researchers would reject the hypothesis that the dice are fair. I hereby declare that the stats say the squig dice are not fair. Well done, sir.
EDIT: Lol independent verification of stats results at 4:48 hahahahahaha
I like the idea of characters of having a final send off. Have them die epically in a campaign, get one last mini. And that’s it.
I thought that Dice are supposed to have a equal chance of outcomes like odds shoud be 1:6 of any one outcome. Aren’t they considered bad if they land on one number more often then others
I absolute love this channel's content.
Could always use them to track remaining wounds on units with 2-6
I'd like to point out that using a dice rolling simulator, your results actually looked pretty typical of what i was seeing. After rolling 3000 dice just a few times, the dice would usually fall around ±25 from the expected 500 with my most extreme difference being this one, with an 88 point difference between 1 and 2:
1 : 537 times
2 : 449 times
3 : 507 times
4 : 492 times
5 : 513 times
6 : 502 times
What does this mean? Well...I'm not sure. Maybe the squig dice are even more balanced than it seems.
Okay, now do the Death Guard dice with their delicious gummy insides and rigid outer coating.
I have the Nurgle dice which were plastic cases with rubber inside and having pusuals represent the values. I don't roll them in games cause they bounce a ton, I would be very curious to see what they would do under the same circumstances.
i felt that the dice werent going to be balanced due to their bouncy nature so i just made them into proxy squig models as figured it would be more fitting both as a proxy and as a reference to the many different types and shapes of squigs.
also is that megatron from beast wars in the background or grimlock?
Hey Snipe and Wib, if I were to paint an entire model with a single human hair (ref: 3rd edition SM codex compliant,) What would you do? ;)
Neeeeerd!!!! 😉😆 I love that you actually went and did this, you legend!
I'm glad that despite your admittance to a lack of ability in data science you only conclude that your dice under your rolling conditions absolutely have this issue, others might conclude it for all of these dice
They need to make a normal dice set like this, I would buy it instantly to add to my collection.
these remind me of those old nurgle dice.