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.

Ghaz wrote:If its not equal, then its not random because one (or more) models have a greater probability of being chosen. Stacking the odds in one model's favor of being chosen is not random.

False. Look up random.

1) Put all of your opponents models into a barrel.
2) Spin barrel around multiple times to make sure everything is shaken up good and proper.
3) Ask an impartial observer to reach in and select a model.


Hey, it's considered random enough for the lottery, why isn't it good enough for your warhammer?

The dude who said cards before may be onto something. Unless its a blob of 5 guard squads each of which has a commissar, it will be enough.

Random is defined as "without pattern". If one or more models have a greater chance of being chosen you have set up a pattern that shows they are being chosen more often. Hence I stand by my statement that it is not random.


Automatically Appended Next Post:
http://education.yahoo.com/reference/dictionary/entry/random

JB_Man wrote:Unfortunately, multiple d6s wouldn't work very well because then you can never get 1...

Which is solved by subtracting by X, where X is the number of d6s used - 1.

Ghaz wrote:Random is defined as "without pattern". If one or more models have a greater chance of being chosen you have set up a pattern that shows they are being chosen more often. Hence I stand by my statement that it is not random.

No, they only have a higher chance of being chosen. It is still determined by random chance.

Are you trying to say that noting in 40k is random? Because every event is weighted in one form or another.
High/Low probability decided by a roll of the dice is still random.

No, because once again you've set up a pattern that will show up in the results.

Are you trying to say that noting in 40k is random? Because every event is weighted in one form or another.

Who said everything in 40K is random? The dice rolls may be, but the results of WS10 versus WS2 will in the long run show a pattern that's anything but random.

Ghaz wrote:No, because once again you've set up a pattern that will show up in the results.

Are you trying to say that noting in 40k is random? Because every event is weighted in one form or another.

Who said everything in 40K is random? The dice rolls may be, but the results of WS10 versus WS2 will in the long run show a pattern that's anything but random.

1
1
3
4
5
6

I roll a (modified) d6. The probability that I roll a 1 is higher than any other number, but the outcome is still randomly determined. It is possible, though unlikely, that I will never roll a 1. Randomness is not tied to probability.

Sorry Ghaz, but a weighted random number generator is still random.
If you can prove that it isn't, there will be a whole heap of very upset mathematicians and statisticians upset that you have debunked their life's work in the field of probability theory.
Or are you saying that.. a leadership test isn't random because it is weighted towards 7? ... that roulette wheels don't randomly determine red, black or green?

That being said, depending on the number of dice rolled, it is biased towards certain numbers and as such shouldn't be used to determine your opponents models.

I would suggest getting a D100 (or 2 D10's) and counting along your opponent's models until you reach that number. If he's deployed, you might need to count as if reading a book, ie left to right moving down the board, or similar.

Trasvi wrote:Sorry Ghaz, but a weighted random number generator is still random.
If you can prove that it isn't, there will be a whole heap of very upset mathematicians and statisticians upset that you have debunked their life's work in the field of probability theory.
Or are you saying that.. a leadership test isn't random because it is weighted towards 7? ... that roulette wheels don't randomly determine red, black or green?

That being said, depending on the number of dice rolled, it is biased towards certain numbers and as such shouldn't be used to determine your opponents models.

I would suggest getting a D100 (or 2 D10's) and counting along your opponent's models until you reach that number. If he's deployed, you might need to count as if reading a book, ie left to right moving down the board, or similar.

Very true there are 2x Revolvers one has 3 bullets, one has 1. if you need to play Russian Roulette there is still a random chance you'll die/live. They don't have to be weighted evenly as long as there is a chance .

I roll a (modified) d6. The probability that I roll a 1 is higher than any other number, but the outcome is still randomly determined.

You're confusing the result on the dice with the result that the roll gives you. It is the result that must be random. Try the aforementioned WS10 versus WS2 one hundred times. If it were truly random, then the results would be near equal. I believe you will not be the case.

Warhammer 40K is not a random game. It is a game of odds like any casino game.

Randomly select a model eh? Ask a passer-by to pick a random model. It doesn't get more random than human nature.

For a really numerous unit, like Ork Boys or IG blobs, start by selecting a model in the center, then roll a scatter die and a D6. The scatter die indicates the direction and the D6 tells you how many models in that direction to count. If you roll a hit, the starting point model is chosen. Viola!

That gives the starting point model a 1/3 chance of being chosen. If I was the Necron player, I'd choose to start on the Power Klaw Nob.

xlEternitylx wrote:Randomly select a model eh? Ask a passer-by to pick a random model. It doesn't get more random than human nature.

Actually, it does. Humans are awful at picking at random.

MasterSlowPoke wrote:Actually, it does. Humans are awful at picking at random.

"Ooo, I randomly select this large flashy model with the cape and the big sword"

If you only have d6, a consistant way for random numbers is as follows.

Take however many models you must randomize, and break them down into groups of 6.

Take those groups, and break them down into groups of 6.

If you have more than 36 models, form groups of 36.

Now roll a d6 to figure out which of the largest groups are chosen, and work your way down. If you roll a group with no models in it, start over.

Example with 10 man tac squad, with 7 regular marines, 1 sarge, 1 missile, 1 flamer

Group 1
1: regular
2: regular
3: regular
4: regular
5: regular
6: regular

Group 2
1: regular
2: flamer
3: missile
4: sarge
5: start over
6: start over

So, you would roll a d6. Since their is only 2 groups, on a 1-3 its group 1, and on a 4-6 its group 2. Then roll another d6 to see which is hit. If you get a start over, then go back to rolling your group.

With 2d6 rolls, you can generate up to 36 unique results. With 3 d6 rolls, its 216 unique rolls. This is more than enough, using only the dice you normally bring for your army.

Devian, that method is biased, as explained above. All models need the same chance to be picked to be fair, and the models in your group 2 have a 1/8th chance to be picked, while the models in group 1 have a 1/12th chance to be picked. Use the methods I posted earlier in the thread (my D10 is similar to yours, but split into two 5 model groups) or use a random number generator like random.org.

Numbering for the bases could be a good way to do it.

Just paint numbers on the base of the model somewhere and have the corresponding numbers in a bag, then your opponent picks a number until you get one that matches one that is still on the table.

Or paired die sets with matching colors, that way you put a colored die next to each of the models, then roll the corresponding ones, and the highest roll is picked at random.

Or instead of cutesy, convoluted ways, just do it with the simplest, fairest method - either a short dice chain or a quick trip to random.org. It's not like the game room will have a shortage of smartphones lying around.

insaniak wrote:

Tri wrote:You know what would be simpler then that? Pick a number 1-however many choices (right it down), then your opponent picks a unit/ model and you work round clockwise (or what ever) till you get to that number ... that's the model/ unit effected.

I really like this idea. Simple and elegant.

Thanks, but you seem the only other person that want a simple solution

Masterslowpoke, the way described is fair. Check the math. The key is if you roll a 5 or 6 in group 2 you start from scratch, rolling to see what group (1 or 2) you get.

Its like rolling a d4 on a d6 by rolling until you get a 1-4 and rerolling 5-6.

Heres an Idea... I'm serious... Close your eyes and use your finger and just press on one... that should be random enough... no peeking though...

Ghaz wrote:

I roll a (modified) d6. The probability that I roll a 1 is higher than any other number, but the outcome is still randomly determined.

You're confusing the result on the dice with the result that the roll gives you. It is the result that must be random.

The method of determining the model is random. That method has a higher probability to determine a 1 than any other number, but it's still random.

Warhammer 40K is not a random game. It is a game of odds like any casino game.

Odds dictate Probability. From wiki on casino games, "Games available in most casinos are commonly called casino games. In a casino game, the players gamble casino chips on various possible random outcomes or combinations of outcomes."

The odds say that a model with a 2+ save and 2W that gets hit by 6 bolter shots will be fine. Since the determination of the outcome is random, there's no guarantee.

Or can you predict what a die will roll before you roll it? If you can't, the outcome is random. Most computers use an algorithm to determine a random number, which means the number is only pseudo random. Calling something not random means you can predict the outcome - in the situations you presented you can predict that some models have a higher probability of being chosen, but you cannot predict exactly which model.

Whenever I get this rule I usually associate "Random" with "whatever I like"

DevianID wrote:Masterslowpoke, the way described is fair. Check the math. The key is if you roll a 5 or 6 in group 2 you start from scratch, rolling to see what group (1 or 2) you get.

Its like rolling a d4 on a d6 by rolling until you get a 1-4 and rerolling 5-6.

It is like rolling a D4 - each of the results has a 25% chance to be picked. They don't have a 16.7% chance to be picked.

The problem is that you're not rolling a D4 for the other group. Models in the first group have 1/2*1/6 chance to be picked. Models in the second group have 1/2*1/4 chance to be picked. Run the math, and you'll see that starting over is inconsequential.

No no no master slowpoke.

Each of 10 unique models has a 1/12 chance of being rolled. If you roll an 11 or 12, then you reroll the entire thing, generating a new d2 group roll and d6 model count, until you get a result that is not 11 or 12.

MasterSlowPoke wrote:

DevianID wrote:Masterslowpoke, the way described is fair. Check the math. The key is if you roll a 5 or 6 in group 2 you start from scratch, rolling to see what group (1 or 2) you get.

Its like rolling a d4 on a d6 by rolling until you get a 1-4 and rerolling 5-6.

It is like rolling a D4 - each of the results has a 25% chance to be picked. They don't have a 16.7% chance to be picked.

The problem is that you're not rolling a D4 for the other group. Models in the first group have 1/2*1/6 chance to be picked. Models in the second group have 1/2*1/4 chance to be picked. Run the math, and you'll see that starting over is inconsequential.

I had to think about this one for a minute but the aforementioned method does work, each model has 3 out of 36 combinations that would yield it being chosen. 30/36 = a model is chose, 6/36 start over.

Well, I'll be damned.

http://codepad.org/zJmv5Eow

Apologies.

Avatar 720 wrote:Dice can be used in the majority of situations. Just number the models and roll the amount of dice necessary to incorporate the highest number (so for 14 models you need 3D6) and either make it so that any numbers higher (in this case, 15 through 18) count as rolling the highest possible result attributed to a model (in this case 14) or simply re-roll until you reach a valid number.

You don't have an equal chance of getting every possibility. The casino game of "Craps" highlights this exactly.