For New Players: Understanding the dice...

[Cheesehead]

Here is the Scenario...

You just spent about an hour or so creating the team of your dreams, carefully naming your players, and have great expectations of grandure. You expect to learn, suffer a few close losses, and eventually rise to stardom. Your biggest expectation is to be treated fairly.

So, your first match seems to go well. You set up your first block of the game. You roll the dice... Double skulls. You immediately click the reroll, and uh ohhhh... Double skulls again. The injury roll is not nice, and results in death.

"This is bull!", you scream (lots of colorful expletives left to your own imagination)... This game obviously cheats only you, so you rage quit and never play again.

Well, hold on a minute... This happens to everyone that plays. Here is why.

First, you have to understand basic odds. Flipping a coin, its 50/50 heads or tails. The next flip is independant. 50/50 again. The odds of the flip care nothing for history. The understanding of odds will not help you win the lottery, or improve your odds at all. The odds simply do not care.

The blocking dice are basically six siders with a minor exception. On each die, there is an attacker down (skull), defender down (explosion looking splat thing), defender stumbles (a splat with a ! in it), a both down result (a splat with a skull), and two pushes (arrows). So, you get a lot of pushes compared to everything else.

So double rolling skulls is impossible, right? No one complains when you get double splats (you will notice this), besides maybe the defender cringing a bit. Why? No one ever rerolls doubles splats unless they sneeze or something.

I am saying this so you will observe the rerolls more frequently. Watch how often they come up pretty much the same in other cases besides just focusing on the skulls.

The odds of getting two "ones" or two "skulls" are 1 in 36. It will happen, and don't be suprised if it is once or twice a game.

When you click the reroll button, it simply rerolls the dice. Rolling double skulls again is not the same as rolling four double skulls in one single roll. The rolls are independant of one another. Once again, history doesn't matter. It does however, at alarming rates, repeat itself.

Here is a link to a decent and simple article to understand this principle...

http://mathcentral.uregina.ca/QQ/database/QQ.09.06/s/matt1.html

Getting upset and screaming "foul" will not change RNG. It doesn't care. When you get double skulls, please know that you will get to watch in slight satisfaction of it eventually happening to your opponent as well.

Blood Bowl will never be fair. Eventually, the dice will smile upon you.

[Dode]

http://steamcommunity.com/sharedfiles/filedetails/?id=527754743

[Dear next of kin...]

These ideas are true, but only in theory. No system in which a single variable or pattern of variables constantly repeats can be considered random. Possibilites are inconsequential because what could happen and what does happen are two very different things. The goal of a RNG is randomness, it's the point of it. Reguardless of the possibilites of your system, if the results your system produces are not randomized then you have not made a proper RNG. Plus, most RNG systems link their seed to the internal clock of your computer, which can be changed, paused, or even reset by several factors.

The only way we could tell if the RNG on this game is truly working is to be able to run hundreds upon hundreds of rolls on the computers the game is running on, then measure the spread of the results. If the results of the RNG are not statistically even then there's something wrong with the system, period.

[DarthPhysicist]

Originally posted by Kolchak:

These ideas are true, but only in theory. No system in which a single variable or pattern of variables constantly repeats can be considered random. Possibilites are inconsequential because what could happen and what does happen are two very different things. The goal of a RNG is randomness, it's the point of it. Reguardless of the possibilites of your system, if the results your system produces are not randomized then you have not made a proper RNG. Plus, most RNG systems link their seed to the internal clock of your computer, which can be changed, paused, or even reset by several factors.

The only way we could tell if the RNG on this game is truly working is to be able to run hundreds upon hundreds of rolls on the computers the game is running on, then measure the spread of the results. If the results of the RNG are not statistically even then there's something wrong with the system, period.

Knock yourself out.

Edit: I wasn't going to ask this at first, but how pray tell do you pause a crystal oscillator? The ability to do this would have profound effects on frequency and timing. Steering/tuning is not pausing btw.

[RustyLinks]

I have to side with darthphysicist here... if you were suggesting that you could somehow replace the seed and try to predict the psuedorandom numbers then generated by the software in a meaningful way then, well I'd still think that would be ridiculous, but at least it would be actually possible.

Tampering with the random seed generated by the internal clock when that is affected by random environmental factors, though, would require Godlike abilities to manipulate the universe at a quantum level. Doing so would probably cause you to disappear in a puff of logic or something.

[DarthPhysicist]

Heisenberg and Einstein (and even Newton) always have their say.

[Dode]

Originally posted by crossby:

I have to side with darthphysicist here... if you were suggesting that you could somehow replace the seed and try to predict the psuedorandom numbers then generated by the software in a meaningful way then, well I'd still think that would be ridiculous, but at least it would be actually possible.

It was done with BB1 just after LE release.

Tampering with the random seed generated by the internal clock when that is affected by random environmental factors, though, would require Godlike abilities to manipulate the universe at a quantum level. Doing so would probably cause you to disappear in a puff of logic or something.

Reading that seed wouldn't require anything of the sort, though, and if you can read it and you have the algorithm then you can replicate the next roll.
Although if you are suggesting that it have a "random" external input into the seed each roll then you're wasting your time: there are already programmes around which can emulate, and therefore falsify and make predictable, such inputs, effectively nullifying them as "random" inputs.

Of course all that is moot: the RNG for online games is server-side. Best of luck reading that seed from memory...

[Cheesehead]

Also, consider this fact. Some people have posted that the odds of back to back double skull rolls are 1 in 1296 (6 x 6 x 6 x 6). This is wrong by any standard. You are not rolling all the dice at once. The arguement that states this wrong by a simple concept. What you are saying is that once you roll double skulls, it will not happen again for 1296 rolls. That obviously never happens.

Yes, these are simulated rolls server side. Any debate about the RNG generator is pretty much invalid because of that. The rolls therefore are unpredictable because you have no way of knowing them.

This also wasn't intended as a debate about a conspiracy theory about random number generation in programming. It was to explain to new players that bad rolls happen in the game despite your best wishes about a perfect world that favors one player.

[Someone Else]

Sadly, the "odds" are not only agaainst the player, but they are predictable as well.

Let it be known that I truly like this game and have played it much more than is apparent, but the RNG here is NOTHING compared to real, physical dice. I have played many video games for many years that use RNG and many, many more board, tabletop, and RPG games that use insane amounts of dice rolls as well as simply played with physical dice for character creation, fun, experimenting, etc. and have never seen such repeatable and predictable results as I see here.

You can simply not compare the two nor can you defend this game against anyone who has played two score years of RPG's. Never in all my years have I experienced the physical dice "act" the way the Blood Bowl dice "behave". I cringe often in Blood Bowl before certain plays because of the play that is about to unfold (eg: amazing "luck" for the ai and outright "hopelessness" for the player).

You state the odds and randomness in Blood Bowl RNG but these "odds" simply do not happen with physical "odds" as they happen here. You see, it is about the timeframe and predictability.

I, as well as numerous other players, have experienced multiple double skulls back to back to back to back but have NEVER, not even once, experienced back to back attacker down. The thing is this happens at the start of every new game I play when a big play is about to occur that would be in my favor but when the ai is in the same predicament (or worse) they amazingly pull through. It simply occurs too much too often. If you don't believe then compare with physical dice rolls or random number drawings, you simply do not see this behavior because what we have here is not truly random. The dice "roll" will always be the same even if you exit and re roll where in physical dice the roll is not preprogrammed every time you roll.

I wish this game had a single player mode where you can physically roll dice and input the results to then be carried out by the game. That would be a genuine experience.

Anyway, the point is not to fight as one side will likely not be able to persuade the other side, but that does not change what is being experienced. One side likes this game and experiences randomness and another side likes this game and experiences "randomness". It is not the same.

[McMacky]

I don't understand the insistence of people who claim the RNG is broken. If you don't like it don't play, but I've had enough of people consistently blaming when things go bad on RNG. Sometimes you get awful games. It happens. One game as high elves I had 40% of my 6 sided die be ones. Know what happened? I lost. Know what else happened? I moved on.
If you don't enjoy the curveballs the game throws at you I don't think the game is for you, simple as that.

[Dode]

the RNG here is NOTHING compared to real, physical dice

Real, physical dice aren't all that: http://www.dakkadakka.com/wiki/en/That%27s_How_I_Roll_-_A_Scientific_Analysis_of_Dice

[DarthPhysicist]

I would like an explanation then how I can beat the ai 90% of the time.

Edit: My Bret record against the AI (what I consider the worst team implemented so far) is 28-2-2 for a 87.5% win rate.

[Someone Else]

Again, one website or one players ability to beat the ai has NOTHING to do with the RNG not being random as physical dice.

I can beat the ai often as well but that is not the arguement here.

Also, I already said I like the game - that does not mean I have to be blind to things either or like for it to be improved.

I also am not blaming anything for when I win or lose.

The point is that the RNG...

[Dode]

You need to show that the RNG is not random, then. Because every test which has been done so far that I am aware of shows the distribution falls within expected ranges. Do you have any REAL evidence (i.e. numbers) to show this is the case?

Even if you do get a single instance of the RNG falling outside those ranges you have to expect there to be a margin of error on your measurements, a margin which can be measured based on the sample size and the confidence desired.

[the_ghostboy]

In my multiplayer matches I have experienced double attacker down results against me heaps of times. It's hilarious, especially if it's a chaos warrior against a wood elf lineman.

A turn later, my wardancer will roll double 1s on his dodge. Then the silly bugger breaks his armour and ends up missing the next game. What are the chances eh? But it happened.

My point is, the RNG is fine. You win some rolls, you lose some.

As one of the posts above said, how often do you notice the double defender down rolls? They happen a lot. In one match, I watched a whole line of Saurus without block skill make every single 2D block. Not one push or defender down. It was frustrating as hell, but it happens.

Hey, but I've only been playing RPGs for thirty years, so what would I know ;)

[Someone Else]

I have been playing RPG's for over forty, so what do I know either ;)-

Maybe someone can show that the RNG is random? Also, if the RNG is truly random then how can you say "very test which has been done so far that I am aware of shows the distribution falls within expected ranges.". Expected ranges and random are two different things...

Predictable is not random...