57 coaches online • Server time: 22:27
Forum Chat
Log in
Recent Forum Topics goto Post FUMBBL HAIKU'Sgoto Post killing by fun?goto Post Pact/Renegades meta
SearchSearch 
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic
G_u_s



Joined: Jun 16, 2004

Post   Posted: Aug 01, 2004 - 16:48 Reply with quote Back to top

" Yes, you may see numbers that average out, but believe that you cannot truly generate a random number in a computer. This is why you see strings of 1s, and also strings of 2 dice skulls."

but come on ! you DO see those in real life BB FFS ! it's just that there are a ridiculous amount of games played IRL as compared to those played on FUMBBL and the other online leagues, thus the "formidable" results are more noticeable...

_________________
There's 10 different kinds of people on Earth: those who can understand binary, and those who cannot.
Renno



Joined: Jun 17, 2004

Post   Posted: Aug 01, 2004 - 16:57 Reply with quote Back to top

Gus

Do you honestly believe more games are played on the board then are played in Fumbbl?

I can push out 2 games on Fumbbl in the same time I can push out 1 game on the board. Smile~

In less then 2 months I've played more then 200 games on fumbbl and have only played 30+ on the board with 2 teams in 6 months. Wink

I'm just curious which random formula is being used.
G_u_s



Joined: Jun 16, 2004

Post   Posted: Aug 01, 2004 - 17:13 Reply with quote Back to top

"it's just that there are a ridiculous amount of games played IRL as compared to those played on FUMBBL"
that means that there are more games played on FUMBBL. this means there are more chances of a completely broken thing to happen, and there are more chances of people bringing it up on the forums.
and just because the computer doesn't "roll" random numbers doesn't mean they're not random to you.

_________________
There's 10 different kinds of people on Earth: those who can understand binary, and those who cannot.
Shinan



Joined: Aug 02, 2003

Post   Posted: Aug 01, 2004 - 17:19 Reply with quote Back to top

If the dice rolls are unfavourable you aren't doing enough to please the gods.

_________________
"Älä löi, en ole hurrit."
BadMrMojo



Joined: Aug 02, 2003

Post   Posted: Aug 01, 2004 - 17:49 Reply with quote Back to top

DaBoss wrote:
...I'm just curious which random formula is being used.

I'm not picking on you. Really. Check out this thread for a lengthy and thoroughly unproductive discussion on the industry-standard random number generation algorithm used by JavaBBowl and my personal favorite post ever.

_________________
Ta-Ouch! of BloodBowl
Condensed Guide for Newbies
H-Hund



Joined: Jan 19, 2004

Post   Posted: Aug 01, 2004 - 17:53 Reply with quote Back to top

Well, if you play many games you will get used to it!
The rolls are equal for all players 8|
Stupid



Joined: May 01, 2004

Post   Posted: Aug 01, 2004 - 18:11 Reply with quote Back to top

Sorry guys i with Mezir on this one, i been playing the board game a long time and clocked just as many games on the board as i have on FUMBBL, 200+. When playing the board my luck has always been as madly unpredictably as on fumbbl theres just as many double ones when you dont want them theres just as many bizaar casualtiess etc etc etc, so leeve it out and give us all a rest.

If you hate dice rolls that much real/computer generated go take up chess.

_________________
Remember the past never dwell on it.
Live the present to its fullest.
Plan for the future never expect from it.
Glomp



Joined: Jan 04, 2004

Post   Posted: Aug 01, 2004 - 18:11 Reply with quote Back to top

AAAAAAAAAAAAAAAAAAAARGH!
AAAAAAAAAAAAAAAAAAAARGH!

THE PAIN MAKE IT STOP!

Its RANDOM you have to cope woth these things. So you lose the game because of a few 1's and skulls. Who cares?
If you know you played well then you can come away satisfied.

These things happen, sometimes we forget how unrandom rolling real dice can be. Check out you opponent next time you play on the board - does he roll every dice roll properly? Or does he just throw them down a little bit so that it lands on the 6 again.

Most real life double ones i've seen and done have been caused by a frustrated player not rolling the dice properly.

The computerised dice roller WORKS if you didn't have the occasional game where you rolled badly it would be incredibly suspect.

Just deal with it folks Wink

_________________
Forum terrorist.
Jinxed



Joined: Jul 04, 2004

Post   Posted: Aug 01, 2004 - 18:21 Reply with quote Back to top

Yep. The only way to deal with it is accept the fact that EVERY dice roll will most likely fail.

That's how I cope. The dice are usually still not fair to me but given the fact that I rely on them to fail me I plan for disaster.
thecoyote



Joined: Jul 06, 2004

Post   Posted: Aug 01, 2004 - 18:29 Reply with quote Back to top

I do enjoy the game, and I in the beginning, I used to get mad about the "random" dice roller, but now I just laugh it off.

As for rolling dice being equally random as the generator: I think not. I have personnally counted 5 GFI rolls in one turn for each player that came up 1s. That's with rerolls too, folks. Think about the proability of that. That's 1 chance in 262,656 that will happen. Come on. Those of you who are proponents of the "javabbl is random" faction will undoubtedly say "well, yes, you just happened to be in the wrong place at the wrong time", but this is not the first time I've seen this. Though I have never noticed the same amount of failed GFIs in a row, I have frequently seen strings of 3 or 4. In a 3 string failed GFI, that's a 1 in 216 chance of that happening, and for a 4 string failed GFI, that's a 1 in 7,296 chance. You cannot convince me that the algorithm is completely random. It isn't. I'm certainly not happy about it, but I still play the game, and hope that those double skulls that frequently pop up on my players don't happen too much.

There really is no solution to this problem, as I stated in my last post. You simply CANNOT "randomize" rolls on a computer. Anything that works on mathematical principles cannot be random. Agree with it or not, you have only one choice: live with it and play, or don't. I choose to play because I love Bloodbowl.

thecoyote
walkertullaris



Joined: May 18, 2004

Post 18 Posted: Aug 01, 2004 - 18:33 Reply with quote Back to top

According to discret mathematics (finally a useful area, never thought that when i read it at the university) you will fail 1/9.

You will fail 1/3 if u have 3+ that makes 1/3*1/3 = 1/9. The move will then succeeds 9/9 -1/9 = 8/9.

Of course the "Muprhy's law" has not been included in the equa
thecoyote



Joined: Jul 06, 2004

Post   Posted: Aug 01, 2004 - 18:34 Reply with quote Back to top

Forgot one thing:

It never bothers me to lose a game because of failed rolls. It bothers me that somehow both of my werewolves on a brand new Necro team get RIPped in their first game. It bothers me that both dwarf blitzer star players get RIPped in the same game. I couldn't care less about losing the game because of failed GFIs, or uncompleted passes, but losing costly positional players is annoying.

thecoyote
BadMrMojo



Joined: Aug 02, 2003

Post   Posted: Aug 01, 2004 - 19:17 Reply with quote Back to top

thecoyote wrote:
...As for rolling dice being equally random as the generator: I think not. I have personnally counted 5 GFI rolls in one turn for each player that came up 1s. That's with rerolls too, folks. Think about the proability of that. That's 1 chance in 262,656 that will happen.

I'm not sure what you are talking about here. You saw someone make 5 GFIs and then fail one? That's exactly 1 out of 6 that failed. That can't be the point you're trying to make, so what's this all about again?

thecoyote (my emphasis) wrote:
Come on. Those of you who are proponents of the "javabbl is random" faction will undoubtedly say "well, yes, you just happened to be in the wrong place at the wrong time", but this is not the first time I've seen this. Though I have never noticed the same amount of failed GFIs in a row, I havevs. frequently seen strings of 3 or 4....

This is what we like to call anecdotal evidence. If you want numbers, take a log at your log files. Pick any particular match you wish from your "results" folder. Then open the .zip file and find "Xxxxvs.Yyyy.log". Open that file in a text editor. Notepad, simpletext, text edit, whatever.

Scroll down to the end and you'll see something like this:
some log file wrote:
D6: 1(176) 2(172) 3(180) 4(196) 5(178) 6(190) Total(1092)
D8: 1(6) 2(3) 3(2) 4(1) 5(9) 6(4) 7(2) 8(2) Total(29)

It's about 10 lines up from the bottom. That's the number of times that you rolled each number on each die. In the case I posted here, my opponent and I rolled a combined 1092d6. See how the numbers are all close, but not necessarily exactly the same? That's what you want to see.

Now here's the tricky part. Any idiot could look at that log file and say, "AH-HAH! I told you they weren't random! You rolled more 6s than 1s and therefore it's broken!" Yes, the idiot can say this and no, it doesn't make them any less of an idiot.

If you had an infinite number of die rolls, you would end up with an exact distribution of numbers. Since there is no way to roll an infinite number of dice in our finite universe, we have to settle for getting closer to infinity. This means that the more dice you compare, the closer you will get to an exactly even distribution. If you were to take 100 log files and add them up, the percentage difference between each number would be tiny. A thousand log files and it's microscopic. Ten thousand log files and it would be infinitessimal. That means really, really small, by the way.

In the example above, the average would be 182 of each number. the 172 twos is the farthest from the average and that's just over 5% off. This is from just one game. Now you can feel free to run through the 100/1000/10,000 log files as I described in the last paragraph and watch how that percentage shrinks. Maybe I'll write a script for people to upload their log files and get a percentage... hmm...

It's random enough, your perceptions are just biased. Check your own log files and quit bitching. Really. It's wasted effort to argue and then hide behind ignorance rather than listening to reason. Sorry to be so short with you, but this has been done a million times before and every time it's new faces saying the same old flawed arguments. It wears a man down.

_________________
Ta-Ouch! of BloodBowl
Condensed Guide for Newbies


Last edited by BadMrMojo on %b %01, %2004 - %19:%Aug; edited 1 time in total
BadMrMojo



Joined: Aug 02, 2003

Post   Posted: Aug 01, 2004 - 19:25 Reply with quote Back to top

Doh. Just noticed that in my example log file, I didn't pick the number with the greatest deviation. There are 14 more 4s than the average, which would be a 7.6.... % deviation. My mistake.

Also, since Wikipedia rocks... here's your homework assignment.

_________________
Ta-Ouch! of BloodBowl
Condensed Guide for Newbies
odi



Joined: Aug 02, 2003

Post   Posted: Aug 01, 2004 - 19:36 Reply with quote Back to top

Just check the messages.log and see what you've rolled.
Sometimes there are a little more 1's around, but then again, sometimes there are lot's of more 6's turning up. The thing is that you've just used your 6's on 2+ rolls.

Edit: Guess I should have read the whole thread first, as this was already pointed out.


Last edited by odi on %b %01, %2004 - %19:%Aug; edited 1 time in total
Display posts from previous:     
 Jump to:   
All times are GMT + 1 Hour
Post new topic   Reply to topic
View previous topic Log in to check your private messages View next topic