Mike_B20
Joined: Nov 14, 2004
|
 Posted:
Dec 01, 2004 - 06:52 |
 
|
| Mnemon wrote: | Mike_B20 ... this random generator is NOT a JavaBB exclusive, but one used by all sorts of Java programs. Now honestly, do you think a failure in such a widely used RNG wouldn't be noticed long long time ago when applied to more "useful" applications? How much more evidence do you need - have a look at the links BMM supplied. Then go ahead and disprove them with collecting your own material, seriously. As long as you just cry about it, without presenting some conclusive evidence that goes beyond "I've seen things" - please don't bother.
-Mnemon |
I'm not crying about it...except perhaps as one in the wilderness.
"It is in the testing dear boy, the testing"...not sure who said that, but he was obviously a great man.
Really though, it IS in the testing that an application is proven.
You are probably right.....we..should...all...shut...up.
NO!
Screw that...the RNG is BUST, BROKE, SCREWED, TITs UP, ONE DECEASED PARROT !! |
|
|
BadMrMojo
Joined: Aug 02, 2003
|
| Posted:
Dec 01, 2004 - 07:31 |
|
| you wrote: | | "It is in the testing dear boy, the testing"...not sure who said that, but he was obviously a great man. |
| I wrote: | | Really. We do have numbers to prove it. Check out the log checker link on my site (link's below). |
Thanks!  |
_________________
Ta-Ouch! of BloodBowl
Condensed Guide for Newbies |
|
Mike_B20
Joined: Nov 14, 2004
|
| Posted:
Dec 01, 2004 - 07:47 |
|
The problem has nothing to do with the overall distribution as far as I can see, but an abnormal distribution of results.
Eg, two consecutive, three dice blocks are identical.
Put that in your new algorithms and smoke it. |
|
|
Mnemon
Joined: Aug 02, 2003
|
| Posted:
Dec 01, 2004 - 07:52 |
|
How about you do some work yourself and prove your point. BMM's sourcecode is up there, tweak it yourself. The rest of us who don't have a problem with it can do more worthwhile things meanwhile. Btw - I failed 5 attempts at picking up a ball on a 3+ (four of them with a re-roll) in a row recently. Then the opponent ttmed a player in and picked it up in the tz without a rr and dodged out. However, I do not complain, but thought it rather amusing - these things happen and are why BloodBowl is random at all, in the first place, and not entirely strategy based.
-Mnemon |
|
|
Mike_B20
Joined: Nov 14, 2004
|
| Posted:
Dec 01, 2004 - 08:17 |
|
My commiserations. A terrible tragedy. Perhaps if you had double zoned this travesty would not have happened.
Anyway...I am not in a position to prove the veracity of BadMr-Mojo's data concerning the accuracy of the Fumbble RNG.
Perhaps I will be shortly if we disbelievers are continually burnt at the stake...
Out of the ashes a Phoenix rises..great idea for a new BBowl unit , no? |
|
|
MrMojo
Joined: Apr 17, 2004
|
| Posted:
Dec 01, 2004 - 10:21 |
|
110 posts! Can we make it up to 1000? I think we can!
No seriously, mikeb20, if you dislike this so much and truly believe RNG is broken, why do stay here and play and then complain? Find a better or make one of your own. I'm sure you would have time to do extensive amount of calculation and testing, let's say for about 5 years or so. When you're ready come back and check if we're still alive.
Besides, if it was broken, so what? It would be broken for all of us, not just for you and for couple of others. It's like democracy, it might be broken, but there's really nothing to do about it (except vote, which I do :-] )
This is just ridicilous. This RNG thingy and fouling threads.... they go on forever....oh man.
And I haven't even been here more than 8 months and I have already seen enough threads about fouling and randomness,
Please people, stop. Please! |
_________________
My post count
Jesus loves me this I know, 'cos my Bible tells me so. |
|
Laviak
Joined: Jul 19, 2004
|
| Posted:
Dec 01, 2004 - 10:54 |
|
A suggestion:
If you really want to test out the java 'random()', I suggest that you look up some statistical theory - there are plenty of textbooks around. Next, learn some of the basics in java, and write some software to calculate the covariance matrices (i think that's what its called) using consecutively generated numbers... There should be PLENTY of information available, as these things have been around for quite some time.
It's not exceptionally complex to do., the maths isn't that complex, and java is a pretty simple programming language, so if you have the urge ... and really want to "prove" that the random number generator is broken (or that it is not broken), give it a go. That would be much more worthwhile than complaining. |
|
|
JanMattys
Joined: Feb 29, 2004
|
| Posted:
Dec 01, 2004 - 11:03 |
|
Mike, find me a team with 11 games played and with 11 players which has 1 MVP each.
I bet you can't...  |
_________________
 |
|
Yeti
Joined: Nov 09, 2004
|
| Posted:
Dec 01, 2004 - 11:25 |
|
Hmmm, about a 1/7148 chance, assuming that noone with an MVP dies before the 11th match, so there should be a couple. |
|
|
Azurus
Joined: Aug 02, 2003
|
| Posted:
Dec 01, 2004 - 13:31 |
|
| Mike_B20 wrote: | The problem has nothing to do with the overall distribution as far as I can see, but an abnormal distribution of results.
Eg, two consecutive, three dice blocks are identical.
Put that in your new algorithms and smoke it. |
Aaaargh!
Why can people not understand? The possibility that stuff like this can happen is one of the things that PROVE the sequence is acting in a (pseudo)random manner.
Something I learnt while building a chip-based PRNG is a lab class at uni: If you get a (large) random sequence of numbers between 1 and 6, and then get a human being to write a 'random' sequence of the same length, they will both appear to be random, with one big difference. The truly random sequence will have a lot more strings of the same number (and strings of patterns, like your 3 block dice, which is effectively just something like 1,5,4,1,5,4). The human-generated sequence will not have these strings, it will be more ordered.
This is because most humans (at least subconsciously-spelling?-) believe that random=fair. This is simply not true. As BMM said, random=unpredictable. Which is why I would be very surprised if these 'strange occurences' did not happen. In fact, if they didn't happen, I might start complaining that the RNG is broken myself 
Ok, enough serious for one day, I'm tired...but first, repeat after me...
Random is NOT the same as fair!
Random is NOT the same as fair!
Random is NOT the same as fair! |
_________________
*This is a public safety announcement. Azurus is a cynical, sarcastic idiot. Please ignore any and everything he may say. Thank you for your attention.* |
|
MrMojo
Joined: Apr 17, 2004
|
| Posted:
Dec 01, 2004 - 14:20 |
|
| Azurus wrote: | Random is NOT the same as fair!
Random is NOT the same as fair!
Random is NOT the same as fair! |
Fouling woodelves IS fair!
Fouling woodelves IS fair!
Fouling woodelves IS fair! |
_________________
My post count
Jesus loves me this I know, 'cos my Bible tells me so. |
|
JanMattys
Joined: Feb 29, 2004
|
| Posted:
Dec 01, 2004 - 15:25 |
|
| Quote: | | Quote: | Random is NOT the same as fair!
Random is NOT the same as fair!
Random is NOT the same as fair! |
Fouling woodelves IS fair!
Fouling woodelves IS fair!
Fouling woodelves IS fair! |
I like pie!
I like pie!
I like pie! |
_________________
|
|
JanMattys
Joined: Feb 29, 2004
|
| Posted:
Dec 01, 2004 - 15:26 |
|
BTW, I totally agree with the first part of the message by Azurus. |
_________________
|
|
mutescreamer
Joined: Apr 09, 2004
|
| Posted:
Dec 01, 2004 - 15:41 |
|
I totally agree with the pie related sentiment.
Basically, note everyone who is complaining is quoting statisitics like 1 in 5386 (number generated randomly by my head ) what that means is that on average 1 person in every 5386 will attain that situation. You should feel priveledged your luck is so bad  |
|
|
Eddy
Joined: Aug 04, 2004
|
| Posted:
Dec 01, 2004 - 15:46 |
|
@Mike
if the RNG isn't random, then it is predictable. if it is, then please tell us how to predict the forthcomming rolls. if you cannot predict them, then it is random by definition. until you have predicted the rolls, or even, let's be nice, when you have proven that it is possible to predict them, then we can consider your issues.
on the luck/skill/whatever.
- skill helps limiting the luck factor, but there is no way you can win a game when you get totally screwed by dice.
- luck or being lucky has NOTHING to do with "true" randomness. rolling 6 "1" in a row on a RL D6 is possible, the dice is random, and it is bad luck. |
_________________
'The generation of random numbers is too important to be left to chance.'
Robert R. Coveyou |
|
|
|
| |
Fumbbl on a Steamdec...
Search
