One more thing about the dice.
Moderator: Community Team
Forum rules
Please read the community guidelines before posting.
Please read the community guidelines before posting.
-
General Cornrow
- Posts: 20
- Joined: Tue Jan 24, 2006 1:49 pm
- Location: NYC
Three things:
First:
Spicy brings up an EXCELLENT point. We see less than 1/2 of the rolls. People should look into a game next time and see how their opponents fared. For example, in one of my games an opponent placed 3 armies in Western Australia on 3 turns and did not lost every battle! Good call Spicy!
NOTE to Lack: Would it be possible to keep a list of all rolls that occurred in a game so we can see what happened when we weren't playing (i.e. who attacked who, who was successful?) It would have the added effort of allowing people to see their defensive rolls, as Spicy pointed out.
Second:
You are correct, Lack. In theory this should mean that any given roll has an equal chance of arising. I made my suggestion of enumerating all possible outcomes and choosing one to try and add a more realistic "feel". Players may feel better knowing that all outcomes do exist and that each one can actually occur. Perhaps the strongest case I can make for my suggestion is that it ensures that every number generated is used. Although I agree that using the current method every number has an equal chance of being omitted, the key to using random numbers is to use as many as possible to ensure randomness. Given the current attack methods (usually 3 vs. 1) and the number of rolls we go through (30,000/day) we lose a lot of numbers. (NOTE: My suggestion may also cut down on the "sticky" feeling, but I have to give that some more thought before I post).
Maybe the easiest way to do this is to just generate a string of numbers (as per Troy and smloh) and choose those that are needed. That ensures all numbers are used.
Third:
I am glad we all agree the most important thing is to not recycle the file.
Let us know when the new file is up running, and how you plan to ensure that it is not recycled.
Thanks, Lack!
First:
The reason why the dice seem so sticky is because we don't see our own amazing defense rolls.
Spicy brings up an EXCELLENT point. We see less than 1/2 of the rolls. People should look into a game next time and see how their opponents fared. For example, in one of my games an opponent placed 3 armies in Western Australia on 3 turns and did not lost every battle! Good call Spicy!
NOTE to Lack: Would it be possible to keep a list of all rolls that occurred in a game so we can see what happened when we weren't playing (i.e. who attacked who, who was successful?) It would have the added effort of allowing people to see their defensive rolls, as Spicy pointed out.
Second:
When generating each line on the list, I randomly picked a number from 1 to 6 five times. That does not favour any combination, does it? Since each line is independent and is used once, each roll combination has an equal chance of arising the moment you roll the dice.
You are correct, Lack. In theory this should mean that any given roll has an equal chance of arising. I made my suggestion of enumerating all possible outcomes and choosing one to try and add a more realistic "feel". Players may feel better knowing that all outcomes do exist and that each one can actually occur. Perhaps the strongest case I can make for my suggestion is that it ensures that every number generated is used. Although I agree that using the current method every number has an equal chance of being omitted, the key to using random numbers is to use as many as possible to ensure randomness. Given the current attack methods (usually 3 vs. 1) and the number of rolls we go through (30,000/day) we lose a lot of numbers. (NOTE: My suggestion may also cut down on the "sticky" feeling, but I have to give that some more thought before I post).
Maybe the easiest way to do this is to just generate a string of numbers (as per Troy and smloh) and choose those that are needed. That ensures all numbers are used.
Third:
I am glad we all agree the most important thing is to not recycle the file.
Let us know when the new file is up running, and how you plan to ensure that it is not recycled.
Thanks, Lack!
CockAsian wrote:These rolls are f*ckin' retarded!!!
I totally agree! No matter how many times I stand on one leg while whistling the national anthem, or say 50 hail mary's while sacrificing a goat, none of the usual methods for influence fickle fate have any results!
It's really too bad that complete games, with every attempted attack and its result, is recorded and displayed, or we could test the unluckiness or luckiness of each particular player.
In the absense of this, I recommend pulling some strontium-90 or cobalt-60 from an old fire detector and hooking it up to a geiger counter via a USB cable, for real randomness. Or consult that rivetting best seller from the RAND corporation, One Million Random Numbers With Standard Deviations.
Ok, the people blaming the dice are being very silly indeed.
Consider the way the game works.
RNG is called for 1st attacker die
RNG is called for 2nd attacker die
RNG is called for 3rd attacker die
RNG is called for 1st defender die
RNG is called for 2nd defender die
If there is a problem with the RNG, it must be one that generates a sequence of low low low high high low low low. A mighty weird periodic sequence that just HAPPENS to match your playing pattern. If the sequence has a slightly different period, then this would drift out of sync and you won't get your alleged effect.
Now, consider something else. You aren't the only person playing. Rolls from other games will increment the internal state of the RNG. That's why it isn't the case that every game is the same. Other players will pour in a huge amount of additional randomness.
FInally, even if we recycle the random number buffer, what will be required for the rolls to repeat is for precisely 500k rolls to elapse between one attack and the next.
So in the end, blame the RNG or not, you are asking for a huge range of independent physical events to conspire against you, just so that you fail your attack. Isn't that a little too much to ask?
Erm, no. There is no reason at all why that should be true. In fact, there are good reasons that skipping ahead in a sequence in a consistent way like this can allow problems to appear which do not exhibit themselves when you just go from one number to the next.
That's insane. That's saying that there is a difference between calling out dice rolls as you roll them, and writing down dice rolls then reading them out.
Consider the way the game works.
RNG is called for 1st attacker die
RNG is called for 2nd attacker die
RNG is called for 3rd attacker die
RNG is called for 1st defender die
RNG is called for 2nd defender die
If there is a problem with the RNG, it must be one that generates a sequence of low low low high high low low low. A mighty weird periodic sequence that just HAPPENS to match your playing pattern. If the sequence has a slightly different period, then this would drift out of sync and you won't get your alleged effect.
Now, consider something else. You aren't the only person playing. Rolls from other games will increment the internal state of the RNG. That's why it isn't the case that every game is the same. Other players will pour in a huge amount of additional randomness.
FInally, even if we recycle the random number buffer, what will be required for the rolls to repeat is for precisely 500k rolls to elapse between one attack and the next.
So in the end, blame the RNG or not, you are asking for a huge range of independent physical events to conspire against you, just so that you fail your attack. Isn't that a little too much to ask?
the key to using random numbers is to use as many as possible to ensure randomness.
Erm, no. There is no reason at all why that should be true. In fact, there are good reasons that skipping ahead in a sequence in a consistent way like this can allow problems to appear which do not exhibit themselves when you just go from one number to the next.
This shows in a very basic way how using fixed numbers skews true randomness
That's insane. That's saying that there is a difference between calling out dice rolls as you roll them, and writing down dice rolls then reading them out.
-
johnnyrotten
- Posts: 2883
- Joined: Sat Mar 04, 2006 4:42 pm
- Gender: Male
- Location: Poole, England
-
max is gr8
- Posts: 3720
- Joined: Sat Jan 21, 2006 6:44 am
- Location: In a big ball of light sent from the future
For all of you mental heads I am also doing a test calculating how many times the roll is used overall and even to please you even more I will do the test and roll 500,000 to verify that they are rndom I'll be back in a month.
Ok I am doing the test but over time At first I'm just finding how often each number comes up
Ok I am doing the test but over time At first I'm just finding how often each number comes up ‹max is gr8› so you're a tee-total healthy-eating sex-addict?
‹New_rules› Everyone has some bad habits
(4th Jan 2010)
‹New_rules› Everyone has some bad habits
(4th Jan 2010)
-
max is gr8
- Posts: 3720
- Joined: Sat Jan 21, 2006 6:44 am
- Location: In a big ball of light sent from the future
fishfleas wrote:I have no idea where Fangz is coming from with my quote......
Just quoting one line from my huge explanation..... I think he is taking it out of context... or maybe he misunderstood me....
You said in your explanation that by rolling and then parsing the list, you influence the rolls. But that doesn't make any sense because each of the items on the list are independent of each other.
The 'problem' you identify with taking numbers off a predetermined list is identical to that of writing numbers down and reading them out in order. Each time you read out a number, yes, you would change the odds of generating the remaining from the list. But you need to be considering the entire event including the generation of the list in the first place, and thus the system retains randomness.
And the rest of the post is just really, really wrong and silly.
NYMEX wrote:sparticus, cry me a river. !
i have game 8 WHERE SOMEONE SURRENDERED BUT 7 TURNS LATER CAME BACK TO WIN THE GAME WITHOUT EVER ROLLING THE DICE
HE OWNED ALMOST THE WHOLE WORLD WITHOUT EVER LOGGING IN. AS WELL I AND ANOTHER PLAYER WERE ELIMINATED FROM THIS GAME FRO BEING DEADBEATS BUT WE BOTH PLAYED EVERY DAY. DELANY43 IS A HOMO AND A POTENTIAL MULTY,
CRACKATTACK, YOUR PROMPT REVIEW IS APPRECIATED
-
ZawBanjito
- Posts: 379
- Joined: Mon Jan 23, 2006 12:25 am
- Location: Somewhere
Re: come on now...just not frikkiin possible!!!
There have almost definately been over 7 million rolls by this point. It had to happen to someone and God must hate you more than anyone else on this site.Spartacus wrote:I am playing 6 games simultaneously and am extremely skeptical right now about the randomness of the die rolls!!!
3 times in a row, I attecked territorries which were occupied by 1 army with a minimum of 7 and in each case lost the battle!!!
I believe the actual odds of this ever happening are over 7 million to 1!!!
you people need to fine tune the randomness of the dice, cause after thsi happening I am not to sure we have a quality product here, this has completely changed my view!!!
FIX IT !!!
IT AIN't RIGHT !!!
SPARTACUS!!!
-
reverend_kyle
- Posts: 9250
- Joined: Tue Mar 21, 2006 4:08 pm
- Location: 1000 post club
- Contact:
The beauty of any completely random system is that anything that can happen, given enough time, will.
I have a cite for this & I can provide it if you want. One statistics professor, on the first day of class, tells his students to go home & flip a coin 200 times. Just glancing at the sheets he can tell immediately who cheated by writing down what they thought would happen, and who actually did the assignment.
How? Easy. The truly random ones frequently have streaks as long as 5 or 10 or more, but the faked ones don't. The student's dont expect it to happen, but the truth of the matter is, with 200 flips, the odds of NOT getting a streak like that are astronomically high.
I have a cite for this & I can provide it if you want. One statistics professor, on the first day of class, tells his students to go home & flip a coin 200 times. Just glancing at the sheets he can tell immediately who cheated by writing down what they thought would happen, and who actually did the assignment.
How? Easy. The truly random ones frequently have streaks as long as 5 or 10 or more, but the faked ones don't. The student's dont expect it to happen, but the truth of the matter is, with 200 flips, the odds of NOT getting a streak like that are astronomically high.
Here's the citation.
http://www.rexswain.com/benford.html
No specific mention of what kinds of streaks to expect.
http://www.rexswain.com/benford.html
No specific mention of what kinds of streaks to expect.
-
PaperPlunger
- Posts: 657
- Joined: Sun Mar 05, 2006 3:33 pm
- Location: Maine!
