Mr-Klipp
Joined: Aug 02, 2003
|
 Posted:
Jun 03, 2003 - 21:52 |
 
|
IIRC, this is why casino dice have the dots painted on, not drilled in. It could cause the dice to favor that side by a very very small margin, something detectable in thousands and thousands of rolls at a craps table, but not so much in the relatively few rolls made in a game of BB. |
_________________
Looking to get your minis painted? Look no further.
The Finishing Touch |
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jun 03, 2003 - 22:05 |
|
| Covertfun wrote: | chance of rolling double skulls twice in a row is
1:1296
.... |
Not true as you seem to mean it.
If you have already rolled a doubleskulls, the likelihood of your next roll being a doubleskulls is about 3% (1/36).
Only before a given 2-dice block you can say the probability for your next two rolls both being doubleskulls is 1:1296.
| Sinner wrote: | | As a medi skilled coach i do about 5 double dice blocks a turn in about 12 rounds a game, which makes 60 blocks. My opponents do the same, so we have about 120 double dice blocks a game (+-20). So it will happen every 10th game that you or your opponent rerolls a double skull into the same result. And for every game you don´t have those rerolled double skulls the probability rises. |
No, it doesn't. The probability stays the same.
However, i wouldn't dismiss Thagrot_Stunticrusha observations too easily. I still remember the thorough resistance I had to face when I discovered the MVP-bug, which was rejected almost unaminously until (and even after) skyjunkie confirmed it. |
|
|
cataphract
Joined: Aug 02, 2003
|
| Posted:
Jun 04, 2003 - 03:14 |
|
| Mirascael wrote: | | Covertfun wrote: | chance of rolling double skulls twice in a row is
1:1296
.... |
Not true as you seem to mean it.
If you have already rolled a doubleskulls, the likelihood of your next roll being a doubleskulls is about 3% (1/36).
Only before a given 2-dice block you can say the probability for your next two rolls both being doubleskulls is 1:1296.
|
the post was in responce to the observation that double skulls were being re-rolled as double skulls often. The probability of this event is 1:1296
Whilst obviously the probablity of an event is not in any way determined by the events that occured before it. We are looking at the event of double skull and double skull.
If we were looking at the situation "everytime I re-roll my block dice I get a double skull" then we are dealing with a probability of 1:36.
I have noticed that when re-rolling block dice I do ocassionally get the same roll (in the same order) this occurs about 1 time in twenty... which is 5% which seems fine (remember there are 2 push backs). I
do fondly remember the occasions a double skull has been re rolled as such. yet i don't remember all the ocassions i have re-rolled a double skull as a pow/ skull ,pow pushback/pow or whatever... a lot of block dice are rolled in a game of blood bowl... and you only really remeber the extreme ones "when my mino got double skulls and killed himself and the apoth failed" or "when I had a 2 dice block against a troll with my halfling and he SId a troll" or "when my Wardancer failed his GFI and got -1 AG"
A lot of dice are rolled in a blood bowl game and extremely favourable or unfavourable dice rolls are bound to happen to you a couple of time in your career... I don't think it's broken although that aformentioned minoutaur situation was bloody outrageous |
|
|
Frankenstein
Joined: Aug 02, 2003
|
| Posted:
Jun 04, 2003 - 03:53 |
|
| cataphract wrote: | | ..."when I had a 2 dice block against a troll with my halfling and he SId a troll" or "when my Wardancer failed his GFI and got -1 AG"... |
Or when the most successful fumbbl Wardancer (Farinrodorel, 255 SPPs) fails a GFI (1/6), fails the GFI RR(1/6), fails armour(15/36), is injured (1/6), killed (1/6), apoth fails(1/6) and turns into a zombie (1/2) - on his way to TD no. 4 (he never scored more than 2 in a single game before, but with rain, a niggling Thrower and being the only AG 5 Elf on the pitch). 
Overall probability: 0.0027%, happens approximately once in 37325 GFIs. Hmmm, the chance for the kill alone is 0.0054 per GFI. Not that low actually. Without apoth or RR 0.03%. Without both, 0.19% per GFI. Every 500th GFI with a Woodie will therefore result in a straight kill if no RRs or apoths are available. |
|
|
AlphaX
Joined: Aug 02, 2003
|
| Posted:
Jun 04, 2003 - 05:26 |
|
| Covertfun wrote: | chance of rolling double skulls twice in a row is
1:1296
.... |
This may be true - GIVEN you roll 2 dice and then roll again. BUT - given you just rolled double skulls, the chance you get double skulls on the next roll is (<i>should be</i>) 1:36. In general the chance that it will happen again is a statistical certainty... And the chance that it will happen again with 3-dice blocks is about the same...
| SkiJunkie wrote: |
So many people bring this up but so few bother to go through their log file and see, is it a bug...or is it just the way the numbers are coming off the random generator?
Info on the random number generator in java that is used in jbb.
The class uses a 48-bit seed, which is modified using a linear congruential formula. (See Donald Knuth, The Art of Computer Programming, Volume 2, Section 3.2.1.) Computer generated "random" number are not perfect, but from what I've read this is about as close as you can get.
|
Likely it is the way the numbers come off the generator. Most random generators have a correlation from number to number. (The correlation tends to be higher not for consecutive pairs but more for distant pairs. For example, if you make 32 rolls, and the first roll is a 1 there might be a 25% chance that the 16th is a 1... Which isn't exactly uniform, as the chance that the 16th is a 1 should be 1/6...)
It may also be the way you use the random result. That is, the rand result is 8, 16, 32 or 64 bits. How you use those bits to get a number from 1-6 can influence how "good" the generator is for emulation of a "uniform" distribution.
SkiJunkie, thanks for giving some information on the random function. I'm going to try to pull up my old notes and see if Knuth's is the best choice for uniform 1-6 ... (it may be that his is better for U[0-1]...) |
|
|
DMNT
Joined: Aug 02, 2003
|
| Posted:
Jun 04, 2003 - 12:04 |
|
| Covertfun wrote: | chance of rolling double skulls twice in a row is
1:1296
.... |
Chance of rolling double skulls after double skulls is 1:36.
Suppose you throw 50 blocks with 2 block dice in a game.
p(double skulls) = 1:36, probability of throwing at least once double skulls in one of them is
1-p(never double skulls) = 1-(1-p(double skulls))^50 = 100%-24,4% = 75,6%
So 75,6% probability to throw _at_least_once_ double skulls.
Therefore your probability to throw double skulls twice in a row is greater than 75,6% x 1/36 = 2,1%.
To be excact, the probability is SUM(n=0,k)[Pr(n,k)*(1/36)^n*(1-1/36)^(k-n)*(1-1/36)^n]
with k blocks.
I'll do some math here, and post a little bit later my calculations.
edit: missed one factor (1-1/36)^(k-n) |
|
|
DMNT
Joined: Aug 02, 2003
|
| Posted:
Jun 04, 2003 - 12:44 |
|
with 50 blocks with probability 0.0378598365994701
double skulls w/RR double skulls
60: 0.0452580330784191
70: 0.052599342511984
80: 0.0598842023224396
90: 0.0671130465685841
100: 0.0742863059716016
Download http://tukijaverkosto.org/calc.pl.txt a perl script that calculates also Expectational value of double skulls w/RR. (The average number of these quadruple skull rolls. Probability is the probability of throwing at least once quadruple skulls) |
|
|
|
|
| |
Making Assassins mor...
Search
