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.

If you haven't heard it's based on a dice-type roll-off. Each weapon rolls a single die, D6 for bolter/chainswords, D8 for Plasma gun, D10 for melta gun or heavy bolter, etc. Then the defender rolls an armor die. Marine's armor die is a D6. Some models have higher armor, like a D10 for a Terminator. Players compare the highest rolled on each side. If the Attacker has the high dice the Defendant takes a hit. Then Players compare the next highest dice, etc. Defender wins ties. I think its a good system. Ambush Ally's Force-on-Force and Tomorrow's Wars uses a similar system.

I guess it gets into the small arms can hurt everyone area, but I feel like it would be really wierd to apply to large amounts of dice rolling.

Also, if a marine's armor is a D6, I cannot see this being applied to the larger game of 40k as a whole. A 6+ would just be all 1's (4+ being a d4, 5+ being 50/50 in my mind).

I mean, it works for Prospero, with its smaller subset of models, but as always, that's a bit much random for me. (But my dice hate me, so I might be a wee bit biased.)

Edit: I also generally dislike things that remove interaction between model stats. Then again, everybody rolling on any given turn or during any given interaction is a good thing, so meh? Still, it would be a bit fiddly for large numbers.

 Red_Ink_Cat wrote:
I guess it gets into the small arms can hurt everyone area, but I feel like it would be really wierd to apply to large amounts of dice rolling.

Yeah, Force-on-Force uses more dice but not like 40K. FonF is kind of a skirmish game. In F-on-F you discard results of 3 and less. You also don’t line-up dice results from highest to lowest, you just call-out your results stating with your high dice then next highest, etc. I think it could work for bigger games, but maybe it would work best in a Kill Team like game.

 Red_Ink_Cat wrote:
Also, if a marine's armor is a D6, I cannot see this being applied to the larger game of 40k as a whole. A 6+ would just be all 1's (4+ being a d4, 5+ being 50/50 in my mind).

For 40k I would give Bolters a D8, Las rifles, Shootas, and Guardian shurikens would be D6. In fact I would do dice-types by army. Marines would be primarily a D8 army; their weapons would be mostly D8’s, some weapons getting two or three D8. Orks would be primarily a D6 army; their weapons would be mostly D6’s, some weapons getting two or three D6. Like that.

 Red_Ink_Cat wrote:
I also generally dislike things that remove interaction between model stats.

Actually there’s still an interaction with stats but the stats are the dice-type. I like the dice-type role-off because its does away with the S and T stat. Its also clean and quick. Instead of three roles and six different stats its two stats (dice-types) and a role-off. That’s it. And, as you say, both players are involved throughout. Its just a better mousetrap.

I dislike it greatly, for two major reasons.

The first is a statistical one. The system makes Power Armor much less protective than in 40K, dramatically undercutting the protection of Power Armor. I feel that 40K has already gone too far in making Power Armor largely ineffective, and this makes it worse. Barring winning ties, the defender takes hits in about half of all shots. It's a standalone game, so it doesn't directly undercut 49K, but it further sets the standard that Space Marine armor isn't particularly effective, which is the exact opposite that the fluff tells us.

Second, I dislike the current level of randomness in 40K, and this is even more random. While I don't care for the setting or background, one of the things I really like about Warmachine, for example, is that Armor Values can commonly be so high that some weapons are simply ineffective against certain targets--period. Tabletop 40K insists on giving grots with grot blastas a chance of taking down terminators (which would be fine if that chance were minuscule), but instead, even the best armor fails 1 time out of six. This goes even further in making 'lucky shots' more common, instead of the ridiculously improbable circumstance that the fluff and background present them as.

It may work for the game, as the game is fundamentally 'superhumans vs superhumans' and you aren't necessarily gaming out each bolter round against your enemy's breastplate. You may want to abstract out firefights with bolters firing full clips down to single dice rolls for a faster, more playable experience. But I don't like it in general, and if I wanted a higher level of abstract gaming than TT 40K, I would go with Epic. In a game that features markedly less models than most games of 40K, I would expect greater detail, not less.

For example, I really like Deathwatch: Overkill, because I think it does a superior job to standard 40K of capturing the power dynamic between Space Marines and their foes. The marines aren't unkillable, but they are superhuman killing machines capable of shredding vast numbers of foes. I feel like this system does a worse job of representing that than standard 40K does, so I like it less. Even if it is an elegant system, I don't like the way it models the 'reality' of 30K.

There was a game in the box?

hobojebus wrote:
There was a game in the box?

Lol, exalted.

I enjoyed it and it is super fast to learn. Even rolling a D8 or D10 doesn't mitigate my unholy dice rolling but for normal people it worked well for them.

Power armor being less protective doesn't really effect anything in astartes vs astartes combat. They were never expecting to fight each other so the armor not working very well makes snese

I was hoping that it would be compatible with the game from Betrayal at Calth. I envisioned the possibility of combining the games into a larger campaign, and following it across all of the 30k releases. Each game could have stood alone, but there could have been a set of rules for putting them together (Maybe in White Dwarf) That would have added to the valus of buying the whole set.

I hope this isn't a preview of Eighth, as the last two things we need are weaker power and terminator armors and funny dice. I get enough of the funny dice in D&D. I come back to Warhammer to chuck ungodly piles of D6s and knock down my opponents figures with them.


I have a prospero box on it's way to me, so I will withhold further judgement until I try it out. I don't mind elites being able to kill each other more easily, if it fits the pace of the game, in a standalone. These games were meant to be played quickly, by the kids. between schoolwork and dinnertime. Calth and the Deathwatch Overkill games work well in that respect, while still being "40k" enough to keep the flavor. Both games are enough fun for die hard wargamers to play between matches too, so they seem to add to the enjoyment of the boxed sets in our family.

I hope this one is fun, even if it doesn't mesh with the Calth set. I have plenty of funny dice, and I am guessing that the game will come with what's needed to play. You can never have too many different games in the house can you?

O, like Infinity.

Thirdeye wrote:

For 40k I would give Bolters a D8, Las rifles, Shootas, and Guardian shurikens would be D6. In fact I would do dice-types by army. Marines would be primarily a D8 army; their weapons would be mostly D8’s, some weapons getting two or three D8. Orks would be primarily a D6 army; their weapons would be mostly D6’s, some weapons getting two or three D6. Like that.

Actually there’s still an interaction with stats but the stats are the dice-type. I like the dice-type role-off because its does away with the S and T stat. Its also clean and quick. Instead of three roles and six different stats its two stats (dice-types) and a role-off. That’s it. And, as you say, both players are involved throughout. Its just a better mousetrap.

I don't like the idea of making armies all based on one type of dice. It makes the armies with smaller dice more reliable than the ones with larger dice. And outside of D6 and D8, we get into wierd odds territory. With how little GW already cares about balance, adding a really statistics heavy aspect like this would make it a lot more work, and probably end up with an even more wacky system than we have now. Can you imagine having to roll d10's or d12's or d20's? The difference between getting min and max really starts to matter way too much. I would take multiple d4's or d6's instead any day. With casting defects, the more sides you add to the dice, the more chance that they get a bias that's not perfect. I have plenty of d20's that roll 1's and 20's way more than they should. And they average about 6.5 instead of 10.5 like they should.

And dice v. dice is not interaction of stats. It's just your luck v. my luck slightly modified. I like that there are things that cannot hurt other things in current 40k. So different? Yes. Better? Not in my opinion.

Have all the board games, Deathwatch, Assassin-o-rama, Horus Heresy 1 etc all had different combat systems?

Plus the one for the cheap Ork vs Marines game?

 Red_Ink_Cat wrote:
I don't like the idea of making armies all based on one type of dice. It makes the armies with smaller dice more reliable than the ones with larger dice.

Well only the base troopers would get the base dice-type. Commanders would get a the next higher dice-type. For example Ork Boys would be a D6 but Nobs would be a D8, and Kapt'nz would be a D10, and Warlords would be a D12. You would also get extra dice for wargear.

 Red_Ink_Cat wrote:
And outside of D6 and D8, we get into wierd odds territory. With how little GW already cares about balance, adding a really statistics heavy aspect like this would make it a lot more work, and probably end up with an even more wacky system than we have now.

That's one of the beauty of the system; the different dice-types replace a bunch of stats and special rules. Its inherently less complicated and easier to balance because there are less stats and less special rules to deal with.

 Red_Ink_Cat wrote:
Can you imagine having to roll d10's or d12's or d20's?

You would still be rolling mostly D6's, unless you had a marine or aspect army then it would be mostly D8's. But remember, marines and aspects would cost more per model so there would be less of them. Also remember I'm mostly suggesting this for Kill Team so there would be a built-in limit on models/dice. The only D20's I would use is for Ork Big Gun and Mek weapons, just to give them that krazy Orky randomness they should have.

 Red_Ink_Cat wrote:
The difference between getting min and max really starts to matter way too much. I would take multiple d4's or d6's instead any day.

I would include D4's. Grots would be a D4 army. But most armies would be base D6. Tomorrow's Wars uses a neat mechanic where you add dice based on tech levels of the army. That's something I would borrow, so tech heavy armies like Eldar and Mechicius would get to role an extra dice for, say, every five-man squad. That represents their high tech weapons.

 Red_Ink_Cat wrote:
With casting defects, the more sides you add to the dice, the more chance that they get a bias that's not perfect. I have plenty of d20's that roll 1's and 20's way more than they should. And they average about 6.5 instead of 10.5 like they should.

I'm not too concerned about that. Its just a game of toy soldiers after all. And you can always object to a player's dice and have him replace it with another.

 Red_Ink_Cat wrote:
And dice v. dice is not interaction of stats. It's just your luck v. my luck slightly modified.

When you're rolling dice there is always luck. That's what makes it a game. When you're rolling against a static number it's just your luck. But when its a role-off against you're opponent its, as you say, your luck v. his luck, which, in my opinion makes for a more fun game. And isn't that what this is all about?

 Red_Ink_Cat wrote:
I like that there are things that cannot hurt other things in current 40k. So different? Yes. Better? Not in my opinion.

Well, understand Porspero is a dumped-down version of what I'm suggesting. For one thing there's no cover aspect. If you add a modifier for cover then yeah, you can get to where things cannot hurt other things.

Maybe make Terminator armour D10, Power Armour D8, Carapace Armour D6, Flak Armour D4 and for anybody else the attack auto falls on a 1

 Kid_Kyoto wrote:
Have all the board games, Deathwatch, Assassin-o-rama, Horus Heresy 1 etc all had different combat systems?

Plus the one for the cheap Ork vs Marines game?

Yep. All different.
B@C and Overkill were my two favorites with B@C being a solid core for a full game of 40k if done right.

Jefffar wrote:
Maybe make Terminator armour D10, Power Armour D8, Carapace Armour D6, Flak Armour D4 and for anybody else the attack auto falls on a 1

Yeah, that’s the idea. I was thinking maybe 2D8 for Terminator armor (pick the highest result), D10 for Artificer/Exarch armor, D8 for Power/Aspect armor, D6 for Just about everyone else. Orks and others would have a simialar thing.

You could also do power fields. Power fields would give an extra save dice. A grade l power field would give an extra D4, a grade ll power field would give an extra D6 save dice, etc. up to a grade V power field that would give a D12 save dice. Again you would role both your normal Armor Save dice and your power field dice and pick the highest result. You could also do save step reductions; as you take hits you lose save dice. Of course this would only apply to models with multiple save dice. Problem here is that it would require some record keeping or chits to mark models which have taken damage (and lost save dice). Again this would be for a Kill Team type game.

Not really a massive fan of the system to be honest. It seems to me like it'd have more variability than a multiple D6 system. I've never really understood why people dislike multiple D6 systems, D6 are fast to sort and roll, it doesn't really take up a lot of time unless you go stupid with it and have 30+ dice needed to be rolled at once.

Does anyone know a quick way of calculating the stats for such a system. I haven't really thought about it too much but the only way I can think is to write a script that calculates all combinations and then uses if statements to sort them.

They hate the cheap dice that roll unevenly and come up with results below the statistical average.

GW's chessex dice are garbage so any system reliant on them is doomed to disappoint.

hobojebus wrote:
They hate the cheap dice that roll unevenly and come up with results below the statistical average.

GW's chessex dice are garbage so any system reliant on them is doomed to disappoint.

Have you tested them?

I have. They are reasonably balanced. Out of a bunch of 53 GW dice rolled over 2000 times the averages were:

1: 15.7%
2: 17.0%
3: 16.8%
4: 16.3%
5: 16.4%
6: 17.8%

So the overall average was a result 3.54 instead of the expected 3.5.

There was a "study" someone a while back that suggested something insane like you'd get 1's nearly 30% of the time but I think their methodology must have been fethed or they kept retesting the same bad die because having tested it myself and read of other people testing it, the results are far more reliable.

The fact you'll probably own a lot of D6 likely helps combat manufacturing flaws.

If I were to guess I'd think the higher numbered dice are probably more prone to flaws that create bias, especially if you only have a few of them.

Yes I had a game where both players used the same GW dice and we kept note 1was the most common result around 40% followed by 6 at around 30% we threw those dice away right after that game.

More recently playing deatbwatch using the dice provided those dice either roll high or low.

There's no consistency just because some get a decent set does not mean everyone does.

I've got proper square edged dice and they've never given me the issues GW dice out of their games have.

So you actually sat down and rolled them a couple hundred times to check it wasn't just a statistical anomaly?

I wouldn't be surprised if you had a bad dice here and there, which is why I don't like when people have "favourite" dice that they insist on using for certain rolls, but as I said, if you have a decent pool of dice I'd be rather surprised if they didn't average out.

The GW dice I tested were just ones I'd gotten in various boxed sets over the years plus one set of GW dice when they came out in a LOTR tin. They were so close to even that I don't care, 1% out is not realistically going to make a difference.

Square edged dice I've used don't roll nicely on a hard gaming surface and I think it'd be much easier to cheat using them.

And if D6 are that bad what makes you think D8+'s are going to be better?

AllSeeingSkink wrote:
Not really a massive fan of the system to be honest. It seems to me like it'd have more variability than a multiple D6 system.

Well of course you can do multiples of different dice-types as well. You can do 2D8, or 3D10’s etc. You can also mix it up: D6D8, 3D6D10, etc. The combinations are endless, it gives you far more variability than multiple D6’s.

AllSeeingSkink wrote:
I've never really understood why people dislike multiple D6 systems, D6 are fast to sort and roll, it doesn't really take up a lot of time unless you go stupid with it and have 30+ dice needed to be rolled at once.

The problem with a D6 is its gives a limited range. That’s a problem for a game based on a universe as vast, varied, and rich as 40K. To make up for the limited range of the D6 you have to have a bunch of stats and dice roles. Sure rolling D6’s is fast and they are easy to sort but it gets a bit monotonous rolling and sorting over and over again.

AllSeeingSkink wrote:
Does anyone know a quick way of calculating the stats for such a system. I haven't really thought about it too much but the only way I can think is to write a script that calculates all combinations and then uses if statements to sort them.

Not too concerned about Mathhammer. I just want a good fast, clean, fun, rich game of Warhammer, and dice dice-type role-off does that.

40k doesn't even use the full range of a D6 And you regain some of the granularity and reduce variability by having multiple stats (a resilient model can be represented by toughness OR armour save OR invulnerable save).

I didn't mind D8/10/12/20's for odd rolls or in a small skirmish game, don't really like the idea for a mass combat game like 40k.


Automatically Appended Next Post:

Thirdeye wrote:

AllSeeingSkink wrote:
Does anyone know a quick way of calculating the stats for such a system. I haven't really thought about it too much but the only way I can think is to write a script that calculates all combinations and then uses if statements to sort them.

Not too concerned about Mathhammer. I just want a good fast, clean, fun, rich game of Warhammer, and dice dice-type role-off does that.

I'm sure you don't, but the question still stands Is there an easy way to calculate it out or is it just a case of writing a program to do it?

And the way dice rolls are done at the moment isn't what makes 40k slow and unclean, it's the convoluted rules, vastly too many special rules and exceptions to the core rules along with being a high model count loose formation game so movement phases are slow and clumsy.

While I don't care for the setting or background, one of the things I really like about Warmachine, for example, is that Armor Values can commonly be so high that some weapons are simply ineffective against certain targets--period. Tabletop 40K insists on giving grots with grot blastas a chance of taking down terminators

I know this is subjective, but to me this is not good game design either. I have played many games that are static, and where randomness is slimmed down as much as possible, and the result is that there is never a reason to ever take units that are not at the top of the bell curve.

In this example that would be to take troops that can be hurt vs taking troops that are immune to X% of the weapons in the game.

You'd ALWAYS take the troops immune to X% of the weapons in the game.

Hell even in older 40k when that was a thing you saw exactly this.

I think the D6 is a problem for sure, but I don't like the idea of having invulnerable troops running around because all it will take is 10 minutes of math hammer to find the models that are immune to as much as possible, and you've just narrowed the game down to only seeing those models.

AllSeeingSkink wrote:
40k doesn't even use the full range of a D6 And you regain some of the granularity and reduce variability by having multiple stats (a resilient model can be represented by toughness OR armour save OR invulnerable save).

Yeah, I get that, it just seems a little rube goldbergie to me. With D6 you need three stats. With dice-types you only need one stat, a dice-type. The granularity is build right in.

AllSeeingSkink wrote:
I didn't mind D8/10/12/20's for odd rolls or in a small skirmish game, don't really like the idea for a mass combat game like 40k.

You might be right, that’s why I’m suggesting this just for Kill Team type games. To me that’s where the game should be. When I get together with my buds we like to talk 40K, GW, Sci-Fi, the latest genera movies and TV shows, other games, show off a new paint-job and/or conversion, and then get down to a game. I don’t usually have time for a big long game, so something quick and fun works best. Mostly its been X-Wing for me, but I’d rather it be Kill Team with a dice-role-off combat system. I made up some stats for the game I would like to play. Check it out:
https://lookaside.fbsbx.com/file/Space%20Marines%20%284-15-16%29.pdf?token=AWx2FANz3pyzJ5p3c88ky9PDKI6WhBOFbIXSjj5nHzYZY5K6pUu7ZZLPigOaIq1t6myVXr-QuOquL1kOm2PyeZ4VinpWlcEntvtUnozR_4m_YRG4k_wgEsHxEgm88_o_a9OhTgSuKrQCuDtdpgYr7ytiYf25HV0yqHzvvMHoFO3maA

AllSeeingSkink wrote:
And the way dice rolls are done at the moment isn't what makes 40k slow and unclean, it's the convoluted rules, vastly too many special rules and exceptions to the core rules along with being a high model count loose formation game so movement phases are slow and clumsy.

I certainly agree that there are too many special rules and exceptions to core rules, and the high model count, but I also feel that the game is a bit rotten at its core, and that because it tries to translate the rich 40K universe with a D6.

I like the interaction it creates between the players. You roll off against your opponent for each attacK. You feel more involved with this system.

It may start to break down and slow game down a lot at higher points levels,but it's more fun then sitting through an opponents shooting phase. All you do there is roll your armor saves, if you get them.

Kinda reminds me of Frostgrave.

AllSeeingSkink wrote:

hobojebus wrote:
They hate the cheap dice that roll unevenly and come up with results below the statistical average.

GW's chessex dice are garbage so any system reliant on them is doomed to disappoint.

Have you tested them?

I have. They are reasonably balanced. Out of a bunch of 53 GW dice rolled over 2000 times the averages were:

1: 15.7%
2: 17.0%
3: 16.8%
4: 16.3%
5: 16.4%
6: 17.8%

So the overall average was a result 3.54 instead of the expected 3.5.

There was a "study" someone a while back that suggested something insane like you'd get 1's nearly 30% of the time but I think their methodology must have been fethed or they kept retesting the same bad die because having tested it myself and read of other people testing it, the results are far more reliable.

The fact you'll probably own a lot of D6 likely helps combat manufacturing flaws.

If I were to guess I'd think the higher numbered dice are probably more prone to flaws that create bias, especially if you only have a few of them.

On top of what a lot of other people have said, is that the standard of the bunch of 53 dice? Because if so, you have 2 variables affecting the distribution of results - the number of dice you have potentially overcoming which ones have flaws and the casting defects in the dice.

Unless you took a table/graph and gave stat bars of how far from your average stats (as seen above) each *individual* die rolled, your statistics mean pretty much nothing other than, if we rolled a sufficiently high number of dice every time, the effects of casting defects would even out over the course of a large number of rolls because there are likely dice with defects over a range and a larger number of "better casted" dice.

If you really wanted an accurate sample, you would need something like 50 *sets* of each brand of dice you wanted to test. Said test would take a long time, and that's assuming your rolling surface has no effect on the roll results.

 Red_Ink_Cat wrote:
Unless you took a table/graph and gave stat bars of how far from your average stats (as seen above) each *individual* die rolled, your statistics mean pretty much nothing other than, if we rolled a sufficiently high number of dice every time

Does it matter though?

It might matter to a casino where you roll the same set of a couple of dice enough times that even a small bias would affect the earnings of a table over the course of an evening.

Most gamers don't have a single D6 they roll over and over again, they have a pool of dice they pull from, and if they are pulling from that pool randomly it doesn't matter if an individual die rolls 1's more often and another die rolls 6's more often, they average out to be even.

There may be a couple of dice in my pool that are biased, but not enough to significantly sway the average. When I did my test people had the idea that chessex dice rolled 1's more often than any other number by a large margin and were discussing buying more expensive dice because of that, I was just proving that wasn't the case.

That's why I specifically said I don't like it when people have "favourite" dice, because intentionally or not you could be picking out dice that are individually biased.

Maybe 53 was too large of a pool if most people don't own 53+ dice, it was literally just what I had sitting in my dice tin at the time I wouldn't really expect a pool of 20 dice to be much worse (the number GW currently sells in a block) but may be worth testing.

When it comes to D8/10/12/20's then they are more likely to have a small pool of dice and then individual flaws are likely to be more significant.

Here was the results of my rolling, each series represents the result (1, 2, 3, 4, 5 or 6) and each data point represents a set of rolls. So within a 3 sets of rolls (159 individual rolls) there was only a few percent between them.


It's actually not that hard to prove a set of dice is or isn't biased, if you care enough it only took me an hour or two to do that test. You can roll the dice thousands of times if you want, but after couple of hundred rolls you have pretty high confidence within a few %.

Said test would take a long time, and that's assuming your rolling surface has no effect on the roll results.

Rolling surface can make it easier to cheat, but if you roll them randomly it shouldn't have a large effect and I know other people have tested on different surfaces and gotten similar results to mine. But some surfaces are susceptible to cheating, eg. if you roll the dice so the 1 is always up in your hand, you'll inevitably get some bias in your results and that bias will be dependent on the surface.

Casino dice are definitely going to be more reliable, but they're designed to be rolled on felt surfaces with a large momentum, they don't roll well on gaming surfaces that are hard and people don't usually fling dice hard on a gaming table lest they destroy their models. Casino dice are actually quite easy to cheat with if you roll them on a hard surface and only roll them a short distance.

@AllSeeingSkink

To most people, probably not.

Btw, I like the data. Was there something different that caused the funky skew in your first few throws there?

I was just pointing out that depending on the situation, your results can be a bit misleading. I would take a set of dice from one box as each sample - I would guess that most people have only one box of dice and they probably bought the big box (so a sample of 36 for chessex)- and go from there. But that still only proves it for one sample. To definitively prove that chessex (or any specific brand for that matter) do not have an inherent bias towards 1's in their production method, we would need maybe 50 boxes of dice per producer. Otherwise, there is the chance of just having gotten a unaveragely balanced ser. And to ensure that the throw did not effect the values, I would probably use a dice tower (eliminate variables and all).

I do agree with your assessment of casino dice and dice with more sides. Most people I know have maybe 6 of the things if they have a "lot" and having had cheaters in my DnD group, that tactic works well (he almost always rolled 15-17). Another reason to make a dice tower.

I almost went home and tested my d6's just for fun (2 36 boxes of chessex, 1 box of GW, 1 36 box of a competitor brand, 1 16 box of chessex, a set of random whites, and various dnd dice) and to see how different each set was.

So, as for viability of a system with a bunch of different dice - probably bad as far as statistical variance of dice rolls. But still probably better than the statistics of DnD rolling

 Red_Ink_Cat wrote:
@AllSeeingSkink

To most people, probably not.

Btw, I like the data. Was there something different that caused the funky skew in your first few throws there?

Just the nature of statistics. That's a cumulative chart so it's showing the more you roll, the closer you'll get to the true averages. My first roll of 53 dice had 8 1's, 14 2's, 9 3's, 8 4's, 8 5's and only 6 6's, but after a few rolls the averages started to balance out.

You can calculate a confidence by using a binomial distribution.

I was just pointing out that depending on the situation, your results can be a bit misleading. I would take a set of dice from one box as each sample - I would guess that most people have only one box of dice and they probably bought the big box (so a sample of 36 for chessex)- and go from there. But that still only proves it for one sample. To definitively prove that chessex (or any specific brand for that matter) do not have an inherent bias towards 1's in their production method, we would need maybe 50 boxes of dice per producer. Otherwise, there is the chance of just having gotten a unaveragely balanced ser. And to ensure that the throw did not effect the values, I would probably use a dice tower (eliminate variables and all).

True, the best thing would be to go and buy 10 sets of dice and test each of them, but it'd take about a day to do any cost $140AUD at GW's price, so yeah, I'm not going to do that Alternatively going through dice individually and determining how much each one is biased and by how much which would be quite time consuming.

I did these tests years ago in response to an article on Dakka which suggested GW's chessex dice rolled 1's 29% of the time which people were quoting as gospel. I contended that to create such a huge bias would require severely weighting the dice (beyond small manufacturing flaws) and/or rolling them in such a way to bias the results (like using the exact same motion with the die in the same orientation each time, which would naturally create a bias and it's how people learn to cheat at dice games). So I performed my own test and found much more reasonable numbers. I remember a couple of other people also tested their dice and got similar results to me where 6's were actually slightly favoured but not by much.

Maybe I should buy another set of dice to test as my GW dice have been accumulated over many different boxed sets and maybe quality has varied over that time. They're all in 1 big tin now so there's no way I can isolate which dice came from which set.

I do agree with your assessment of casino dice and dice with more sides. Most people I know have maybe 6 of the things if they have a "lot" and having had cheaters in my DnD group, that tactic works well (he almost always rolled 15-17). Another reason to make a dice tower.

I almost went home and tested my d6's just for fun (2 36 boxes of chessex, 1 box of GW, 1 36 box of a competitor brand, 1 16 box of chessex, a set of random whites, and various dnd dice) and to see how different each set was.

Dice tower is definitely a good idea.

If you want to test the variability of a whole set of dice it doesn't take too long, just roll the set, group them together, write the numbers in to an excel spreadsheet and roll again. If you plan to test each die individually prepare to write off a weekend doing it

So, as for viability of a system with a bunch of different dice - probably bad as far as statistical variance of dice rolls. But still probably better than the statistics of DnD rolling

One of the things I like about rolling lots of dice is that it reduces variability somewhat. I'd be in favour of maybe upping the wounds of many infantry choices in 40k to reduce variability (Terminators for example suffer from horrible variability, they have a 2+ save so on average it takes 6 wounds to kill them, but I think we've all had days where an entire Terminator squad is wiped out by only a handful of shots). I'd suggest most elite units with a 2+ save should possibly be bumped to 2 wounds just to cut down on variability.

Ideally in a wargame like 40k (which typically takes several hours to play a "standard" sort of game) I like minimal variability, enough such that outcomes aren't predetermined, but not so much that a few bad rolls determine the outcome of the game. Games which are faster to play I think can get away with more variability, because I don't care if I win or lose a game due to bad luck if I can turn around and play another game of it in half an hour.