Dick Burnett  26 Oct 2020 10:37 a.m. PST 
It seems the hobby has reverted to using six sided dice, and lots of them, just as I encountered in the hoary days of the DElim, Aelim,Exchange SO I CRT and the ubiquitous six sided dice! I recall the six sided religion which hunted down heretics who used percentile dice! And the gamers who THREW the dice causing a dice safari! Why the return? I see it's of six sided dice bouncing about in too many Warlord games, Toofatlardies games and others. Six sided has severe granularity problems and there are still gamers who THROW the dice everywhere. We heretics used percentile dice, dice boxes and random number generators What happened? Why the return?! 
MajorB  26 Oct 2020 11:08 a.m. PST 
If you lots of D6s then the results are closer to a normal distribution. 
robert piepenbrink  26 Oct 2020 11:19 a.m. PST 
I thought dice WERE random number generators? And there are ways around the granularity problemnot to mention dice trays. If you're just ranting, knock yourself out. If you want an actual discussion, take a look at Priestly & Lambshead, Tabletop Wargames, Chapter Four, which takes the other side. For myself, I've used one time or another every type of dice not licensed for a particular game, and card draw as well. I'll admit to an emotional satisfaction out of throwing dice I don't get out of drawing chits, or cards, and I don't trust electronic systems. Nothing against percentile dice as long as the "tens" die is marked rather than specified. I don't want to figure out what I rolled: I want to know. But could you please tell me who exactly has been hunted down for using D8's, and the nature of their punishment? 
martin goddard  26 Oct 2020 11:24 a.m. PST 
An interesting topic Dick. As an aside. All PP rules use 6 sided dice and have never used anything else since 1985. I do own various multi sided dice but find them less tactile (D4, D8 particularly), less available ( i need 4 D8 for tonight' s game)and sometimes less readable (especially D20 on a rough surface). On the matter of players who throw their dice, it is probably more of a player problem. The games I play insist on good manners and care of property. As has been pointed out, the major advantage of D6 is that quantity levels out the extremes. This is all just chat. You do have the right to ignore any of it, should it cause offence. martin

GildasFacit  26 Oct 2020 11:49 a.m. PST 
Lots of D6 do not give a Normal distribution. Even a crude approximation to one is only possible with accumulated scores of a large number of dice. Lots of D6 do not 'level out' the extremes. The extremes are still there ! Throw 10D6 and it is still possible to get 10 6's. It matters little that this is a rare event, it can happen. 
jwebster  26 Oct 2020 12:05 p.m. PST 
Well, what and how many dice you use depends on the probability distribution you are designing the game for The other thing going on is simplicity. Looking up a value in a table or on a graph, then rolling percentage dice gets old very quickly. With a bucket of 6 sided dice, it's easy to buff up by adding dice, use +/ modifiers on the dice, roll saves against hits, and so on Rolling a bucket of dice across the table is indeed a bad thing. Lots of options on dice trays Although a bucket of dice seems iffy to us old farts, it's the principle (only ?) mechanism in Games Workshop games … If I might digress, people seem to "desire" a normal (or bell curved) distribution. I'm not sure that this represents combat results that well. There can often be lots or no casualties. Very strange results are very unusual, but I think the middle of the curve should be flatter Another digression – it gets complicated to figure out the probabilities from a bucket of dice, once you add in "saves", rerolls, modifiers and additional dice. I doubt that many designers do the math Keep on rolling John 
etotheipi  26 Oct 2020 12:25 p.m. PST 
Six sided has severe granularity problems Well, yeah, but cut them some slack. Having performed operational test of new (at the time I was doing it) military gear for multiple services and a few different countries, I can tell you that you're doing extremely well if you can nail a statistically significant performance answer into a 20% band. And that, of course is (more or less) controlled test and evaluation conditions, let alone actual operational conditions with tons of unknowns at basically unknown levels. So, yeah a d5 would be fine, but just because a d6 is a little overkill, why give them a harsh time? 
etotheipi  26 Oct 2020 12:26 p.m. PST 
here are still gamers who THROW the dice everywhere. Having been a human being for much longer than I have been a military analyst, I'm fairly confident that being a douchebag is not activity specific. I'm guessing those people are creative enough to express themselves, not matter what dice you are using. 
MajorB  26 Oct 2020 1:28 p.m. PST 
Lots of D6 do not give a Normal distribution I only said lots of D6s are closer to a normal distribution not that they give you one. 
Zephyr1  26 Oct 2020 1:40 p.m. PST 
D6 'Percentile' Dice: Roll 20 D6's. 1 to 5 = Score as rolled. 6's = Remove one D6 for each 6 rolled (going from highest value to lowest.) Add up your score (results equal or less than zero = zero.) Alternative Method: Roll 2 D10's… ;) 
Extra Crispy  26 Oct 2020 2:00 p.m. PST 
I have no issue with a d6 but I *loathe* games that make you roll buckets of them. They slow the game down as you count out 17 dice, roll them (or more like, flop them down) then sort out how many are 4+. Now I have to find 8 dice, roll them and sort out how many saves I made. Ick. For shits and giggles I made a chart to replace buckets of d6 using a base 6 system. Roll one red d6 (the tens) and one white (the ones). And look up the result on a chart. So now 17 dice becomes: roll 2 dice, read chart. 8 hits! Roll 2 dice, 3 saves! Much, much faster with a well designed chart. But some poeple hate charts too. 
martin goddard  26 Oct 2020 2:09 p.m. PST 
Prob of 10 scores of 6 is .000000016. "Rare" would be an understatement. In other words 1 in 60 million. Infinitesimal might be better? D6 rolled many times is a uniform distribution. Not much point in stating lots of D6 is not a normal distribution unless detail is given of the subsequent stage. Rolling the D6 is usually only the first part of the process. I would suggest that the statement has little validity unless we discuss what is being done with the results of the D6. Are they being added, multiplied and which scores etc. A pair of D6 added give a normal distribution quite quickly. Say 30 rolls? Added to all the above considerations is how many times are D6 rolled. If, for example, 2D6 (added) are rolled for every action and each moving part carries out 12 actions per turn, then the scores will be very normal. Conversely if a player rolls 2D6 one time per turn, then the results will be very scattered. Overall I expect we each have systems we prefer. None of those systems have to pass a "public acceptance" test.
Therefore I suggest that players use any type and number of dice as they wish. There are probably no "best ways"?? 
14Bore  26 Oct 2020 2:25 p.m. PST 
I stick to 2 dice tens and ones. 
robert piepenbrink  26 Oct 2020 2:59 p.m. PST 
Let us suppose I have a choice in my game design of specifying a roll of 3D6's or one D20. Anyone care to guess which system has the greater chance of an extreme result? EC, I think you're being too hard on yourself. I don't really think you lose significant time counting up to eight. And even Mersey stops at 12D6's. Still waiting to hear about the people tracking down and punishing the heretics. 
khanscom  26 Oct 2020 5:40 p.m. PST 
"Let us suppose I have a choice in my game design of specifying a roll of 3D6's…" Do you add the spots or manipulate them in another manner? Decades ago I played a US football game that read them in sequence (lowest score to highest, giving a 3 digit result). Lots of ways to skin a cat. 
Bunkermeister  26 Oct 2020 7:49 p.m. PST 
I use one D6 with arrows and numbers for direction, hit location, and distance and one D20 for the chance to hit. The D20 is a custom I had made: 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, 100. All 5% increments that don't require you to do math, or count multiple dice, or figure out if the red die is the 10's or any of that time consuming stuff. Mike Bunkermeister Creek Bunker Talk blog 
martin goddard  27 Oct 2020 12:03 a.m. PST 
Maybe I can put in a thought here that might help? Some time ago some fire and fury players decided to use 2 x D10 and average the result. i.e you roll 2 x D10, add them and divide by 2. They did this instead of rolling a single D10. Some will say that their idea makes no sense. (uniform distribution fans) Some will say it is a very good idea. (normal distribution fans) I am with the second group. If it makes no sense then that is fine too. martin

parrskool  27 Oct 2020 12:28 a.m. PST 
…..Average dice when I first started with LWS rules ! 
UshCha  27 Oct 2020 1:56 a.m. PST 
LIke they say its personal preference. Buckets of Die are the ultimate turn off for me. The nominal distibution means in fact you can gat generally very smnall variations. If you roill 100 die the 90% variation is quite small. Without manipulation of the D6 die its hard to get repeatable granuality unless you are prepared to do lots and lots of maths of how variable stack up in multi die and factors. So to me I looks at a bucket of die games and would only consider if the guy writing it had a dergee in maths. I only have a degree in Engineering and the effort to translate probabilities is well beyond my tollerance me even as a wargames designer. In addition lots military data is limited to proability's as the samples are too small to get an meaninful estimate, Engineering wise a good sample is at least 20. In addition my own games they represent ststitics of small numbers generaly, for instance we only have perhaps 8 tanks on table and no two shots are alike so a normal distribution is not representative. D^ as a basis to me does not of itself give a decent granuallity, D20 as advised by my son works great. If you do need an approximation of a normal distribution it can be done. Where the granuality is not an issue you can combine effectivelly 2 issues in one. While I am no painter and art is never a first concern, I do hate die on the table. The 3D epresentation to me is key and "dozens of die" strewn across the table looks bad when not being used and puts models at risk of being dispalaced when rolled. 1 D20 is much less risk, in my experience It make resolution of the statistical side the minor effort it is supposed to be within a game (for me anyway). I WILL NOT PLAY with folk who rather sadley (to me) chant and take forever to throw their dir and talk to them. Such folk are interested in the gambling side and that is of no interest to me and almost invariable they make poor players. Like everybody says there are different views none are right for all. The above is my insight into the issue. 