Thats my latest craft record. No reason why I should have so many fails. Farming seeds for weeks on end selling SS all wasted now. I can see why so many people have been quitting.
Thats my latest craft record. No reason why I should have so many fails. Farming seeds for weeks on end selling SS all wasted now. I can see why so many people have been quitting.
So despite the 'official' word that all enchants are supposed to be 66% (2/3) success rate, only weapon enchants seem to truly come close to this ratio. Granted the # of reported weapon enchants is greater than the number of armors and Jewels attempted, I'd say enough were reported to make the results pretty solid.
In my experience 60% has been pretty decent, but 70% crafts fail alot more for some weird reason.. I'm at about 50/50 overall now though.. I craft when I get a "gut feeling" that it is worth to try :P Only worked with weapons so far, jewelry crafting seems to fail alot, especially earrings!!!
In any game of probability, you need at least a couple thousands repetitions of the experiment, to even dare making any assumptions.
Imagine rolling a normal 6-side die. The chance to roll a 1 or a 2 is 33.33%.
In your example, you rolled the die 10 times and you got 7 times 1 or 2 and only 3 times 3-6.
Now would any of you say, that the die is flawed ? No. In an RPG, you would curse the gods, throw the die away and take another one, despite knowing, that it got the same probabilites.
It is funny, that with dices, noone makes such silly quick assumptions, but when it is a computer generating pseudo random numbers, then you need to have your percentage after only a few tries, to not cry havoc.
If you craft 10.000 items, then you can start making assumptions on the real percentage. But 10 tries are simply not enough for any message.
You could have the same outcome, even if the success rate were 80% for real.
Schmu you're right but the law of great numbers shouldn't have to apply here.
I've tried 4 A weap crafts so far... and I have missed them all.... 2 carnage bows = 56 mils of mats at the time of try and this week 2 elysians...
I have yet to succeed once. I also failed 1/2 attempt at majestic jewelry.
So you or NC could tell me how many thousands Elysian axes I need to craft in hope to approach the 60% success figure ?
It's just like the 'random spoil' and the droprates : VPC ... = Very Poorly Coded.
But it's nothing new and nothing that will change...
Schmu you're right but the law of great numbers shouldn't have to apply here.
Erm... So you want to have 10 tries at 60% and be guaranteed to have 4 fails and 6 successes ?
Heh, imagine, you succeed on the less important (expensive) crafts and then you KNOW, that you will fail the next craft at 60% no matter what :)
I wouldn't know a mathematical way to get rid of the law of great numbers. If I knew, I would be in Vegas, playing Roulette.
It is a pure game of luck. I am sure, that there are people out there, who got a better success rate than 60%. Maybe some even got a 100% streak with 5 crafts. I guess you wouldn't complain about the 60%, if you were one of those.
My craft record is been like that from C3. I have a 60% to 70% fail rate. Its seems like the big stuff always fails atleast twice. I have an idea what is going.
Schmu you're right but the law of great numbers shouldn't have to apply here.
I've tried 4 A weap crafts so far... and I have missed them all.... 2 carnage bows = 56 mils of mats at the time of try and this week 2 elysians...
I have yet to succeed once. I also failed 1/2 attempt at majestic jewelry.
So you or NC could tell me how many thousands Elysian axes I need to craft in hope to approach the 60% success figure ?
It's just like the 'random spoil' and the droprates : VPC ... = Very Poorly Coded.
But it's nothing new and nothing that will change...
You're making the assumption that all of your random numbers are coming from a "private", if you will, random number stream.
I only paid slight attention in math (just enough to get my bachelors) but this sample is way too small to be considered significant.
Why are the numbers such a secret for NC soft. Or the % of drops...
The percentages are posted on the Korean site. Problem is, no one believes them. Here's why:
I want to craft a lance (sorry Lothlian :p)
1. Fail
2. Fail
3. Success
Here I stop since I got what I wanted. This skews the results to look like it's only a 33% success rate.
Dont care about the mathmetician side of it tbh. This game puts a huge strain on the economy by using these random %'s to do everything.
Here is where I love this game:
1. Do a quest for 2 weeks to get enough TOI chapters to turn in.
2. Turn in TOI chapters for a CHANCE to get DC recs.
3. Now assemble the mats for a CHANCE to craft a DC helm.
Rinse and repeat.
Yesterday turned in two sets of chapters for 0 recs. Thank you. Back to TOI.
Dont care about the mathmetician side of it tbh. This game puts a huge strain on the economy by using these random %'s to do everything.
Here is where I love this game:
1. Do a quest for 2 weeks to get enough TOI chapters to turn in.
2. Turn in TOI chapters for a CHANCE to get DC recs.
3. Now assemble the mats for a CHANCE to craft a DC helm.
Rinse and repeat.
Yesterday turned in two sets of chapters for 0 recs. Thank you. Back to TOI.
On the other hand, my clan leader turned in the 13 BP last week and got a Nightmare Helm, Glove AND boots recipe...and all 3 crafts succeeded. It really sucks when the crafts don't work, but when they do...w00t!
Yup, that was pretty neato to hear all of it worked Meate. But then again, Death did 12 majestic earring crafts at once, only 2 worked, so its pretty easy to get discouraged.
On the other hand, my clan leader turned in the 13 BP last week and got a Nightmare Helm, Glove AND boots recipe...and all 3 crafts succeeded. It really sucks when the crafts don't work, but when they do...w00t!
I believe Ben Franklin said: "Even a blind squirel can find a nut."
So it worked for someone. Big deal. Its worked for me in the past as well.
It would be nice if there were a more-work quest in which you have to kill a ridiculously large number of mobs (about 50% more than you would need on average doing the BP quest and attempting your crafts, as a cost for your security) to acquire 100% recipes.
However, I still think A Grade should have stayed optional, then there would be no problem. :p
A grade is supposed to be hard to get. That part is a good idea.
But limiting it to a quest that is ******ed and then giving you a questionable 60% chance of it working is not very well thought out.
Over time each level of gear has become more and more main stream. Because of the design of this grade, it is main stream, but usually through questionable methods.
CURLY: "Hey, Moe, I got a great idea for this game -- we allow people to overenchant at a lousy percentage, and to try and craft expensive stuff that fails a lot, gnuk, gnuk, gnuk."
MOE: (after sticking fingers into Curly's eyes, and beating him over the head with a frying pan): "Moron. If we make people gather stuff for months, or create a really expensive weapon, and then they blow it, they will get frustrated and quit the game, and we lose their monthly money."
CURLY: "Hey, Moe, I got a great idea for this game -- we allow people to overenchant at a lousy percentage, and to try and craft expensive stuff that fails a lot, gnuk, gnuk, gnuk."
MOE: (after sticking fingers into Curly's eyes, and beating him over the head with a frying pan): "Moron. If we make people gather stuff for months, or create a really expensive weapon, and then they blow it, they will get frustrated and quit the game, and we lose their monthly money."
Where is Moe when you need him??????????
Except for duals, when are you forced to OE anything?
On the other hand, my clan leader turned in the 13 BP last week and got a Nightmare Helm, Glove AND boots recipe...and all 3 crafts succeeded. It really sucks when the crafts don't work, but when they do...w00t!
I believe Ben Franklin said: "Even a blind squirel can find a nut."
So it worked for someone. Big deal. Its worked for me in the past as well.
To bad that isnt the point of this post at all.
The point of the post was that the OP failed a bunch of crafts, and has apparently lost faith in the actual % of success rate, according to the recipes. I am simply pointing out that for as many as she failed, others have succeeded. This is perfectly in line for the topic of the thread.
In my experience 60% has been pretty decent, but 70% crafts fail alot more for some weird reason.. I'm at about 50/50 overall now though.. I craft when I get a "gut feeling" that it is worth to try :P Only worked with weapons so far, jewelry crafting seems to fail alot, especially earrings!!!
Woops, forgot about that word. >_<
Still if enchants are skewed this bad, what does this say about crafting %'s?
Computer 'random numbers' are indeed questionable in most cases.
In my experience 60% has been pretty decent, but 70% crafts fail alot more for some weird reason.. I'm at about 50/50 overall now though.. I craft when I get a "gut feeling" that it is worth to try :P Only worked with weapons so far, jewelry crafting seems to fail alot, especially earrings!!!
Woops, forgot about that word. >_<
Still if enchants are skewed this bad, what does this say about crafting %'s?
Computer 'random numbers' are indeed questionable in most cases.
Statistically speaking, they are random enough. We did lots of testing on various random number algorithms in my college simulations class.
Yeah.. It depends on the algorythm.. I don't know why, but my instincts say that the algorythm they use is time based, and if it is, you can forget about randomness..
I don't know why, but my instincts say that the algorythm they use is time based, and if it is, you can forget about randomness..
I don't know. I've been programming for, let's see, roughly 26 years now and one thing I've learned is that programmers are lazy. Therefore, I believe that the algorithm L2 uses would be the simplest. I could be wrong but everyone posting that it took them 3 tries to craft a 60% item are posting skewed results, statistically speaking.
<hijack>
Just out of curiousity, why are time based randomisation algorithms bad? The basic idea of using a real life rather-random event to simulate a computer one always sounded sensible to me. :confused:
<hijack>
Just out of curiousity, why are time based randomisation algorithms bad? The basic idea of using a real life rather-random event to simulate a computer one always sounded sensible to me. :confused:
my guess is if you know the code you can beat it. But that is just a random guess.
As we all know, RNG's aren't truly random. But they're random enough to create a recipe that has a ~60% success rate. Over a period of 10's of thousands of trials, you will end up with a success rate that isn't exactly 60%, but close enough to say that it is 60%. In fact, you would be extremely lucky to actually get a 60% average unless you have an infinite set of trials, which isn't possible.
Now, after say 100,000 crafts you have a 50% success rate on a 60% craft, you have a problem.
Computer 'random numbers' are indeed questionable in most cases.
LOL and why? They fixed the dice issue and as long as you can't predict the list, the number is indeed random by perspective. Failing an extra two crafts than statistics allows for? You have a 4 % chance of having success on 9 70% crafts. 70% isn't too far from 50 so the odds are moderately balanced on the other end. I can crunch all the numbers you want on SAS if you prefer, but I'll just approximate that means about equal odds for failing 5/7 crafts. I know people that never fail crafts, I know people that have failed a dozen. The % chance for failure IS good. And it's a part of the game to DECREASE adena on the server. All those crystals/gems came from somewhere. And that's a large chunk of crafting cost. There are many systems which remove adena/items from the sever to maintain BALANCE. I made a full set of majestic jewelry with no failures. I kept going, made 4 more pieces to sell. Made a BW hvy set at 60%...then failed 2 A weapons that accounted for the value of everything I had success on. I said "well crap" and went back out hunting with friends. You guys read 60 or 70 percent and you don't think. You just see a number greater than 50 and think "wow that's decent odds click click." Those odds SUCK. There's a huge space for error, for loss, and you will suffer it unless you are crafting items that are 1% of your net worth. Some people gamble the farm on 60-70%...that's ******ed, I hope you never do such a thing in RL. Learn from this game, 60 or 70 is FAR closer to 50 than it is 100. Your odds are not good, they're random.
neugf said sadley ... there is plentyof proof to suport it is def time based and null is the simplest i can think of i can dig up some old threads we had on this but i might as well dig up a server change thread ;P
neugf said sadley ... there is plentyof proof to suport it is def time based and null is the simplest i can think of i can dig up some old threads we had on this but i might as well dig up a server change thread ;P
Translation please? Not flaming, I just didn't understand a word.
Wouldn't you need to know a few factors, such as which number range is being operated in (you can make a random "60%" with a number from 1 to 5 or 1 to 1000000 after all), the exact time on the server, the exact algorithm and the number pool for that?
<hijack>
Just out of curiousity, why are time based randomisation algorithms bad? The basic idea of using a real life rather-random event to simulate a computer one always sounded sensible to me. :confused:
my guess is if you know the code you can beat it. But that is just a random guess.
Unless you know the actual algorithm and the seed (based on a clock broken down into very small time intervals and pushed through a [usually] complicated calculation), your not going to be able to beat it. Not only that, but the RNG is not client-side (at least, it better not be :p), so you're not going to be able to influence it, anyway. From the time you've clicked your mouse on the Craft button to the time the server is ready to decide if your lucky, the RNG has gone through potentially millions of numbers.
Yeah, you have to know a few things. If you use time as the seed to a random number gen, its not that bad, but if you use time to get every different "random" number, then it's bad, and you can pretty much figure the algorythm out by trying it.. Of course network lag plays a factor there too, so even knowing everything may not be enough to abuse it..
Some people gamble the farm on 60-70%...that's ******ed, I hope you never do such a thing in RL. Learn from this game, 60 or 70 is FAR closer to 50 than it is 100. Your odds are not good, they're random.
Please come play poker with me, bring your cash card. If I get 60-70% odds on a hand, I am thrilled to death.
Time based algorithms are random unless you can monitor a large number of results. Worked properly, if I crafted 10,000 items I could start to predict failure/success but who couldnt? And who crafts that many at once? You can't change the random number generator you can only work around it. It is important to denote what is necessary to do so. Hence why die were fixed.
just a default sead to call in random number generators
but the game is based off of time the proof i have is all verry old but back in the first olimpic event it was e4very one who killed a mob at a cerent seckond got a event item alof of the events are like that (granted nc has done some trickie things recentley to fix this) same in l2 day bouth of em same as necter same as every event that you collect for
and it is conceiveabul to figgure it out all you need is the sead and the outputs and yoou CAN figgure out the algorithim used with eneugf of bouth
Yeah.. I actually saw that on the L2day event.. Exactly when I got one of the rare letters 3 ppl in vent said they got that same letter that exact moment..
Yeah.. I actually saw that on the L2day event.. Exactly when I got one of the rare letters 3 ppl in vent said they got that same letter that exact moment..
qft ;) lineage 2 at its lazeyest the seckret to the events is to kill as fast and often as posibule just your luck that you killed it at that seckond witch brings me to my outher point why would nc bother writeing a seckond algorithim for drops or for even crafting or enchanting, theay woudent .
If I remember corerectly, the rule of thumb is that the error in your average measurement goes like the square root of the number of measurents. So if
N = 10, sqr(10)/10 = .31 --> 31% error
N = 20, sqr(20)/20 = .22 --> 22% error
N = 100, sqr(100)/100 = .10 --> 10% error
So if you succeed in 5 out of 10 crafts you could estimate a success rate of 50% with an error of
.31 * 50% = 15.5%
So you would guess the success rate to be in the range 34.5% to 65.5%.
It has been 30 years since I learned this method of estimating error, so I may not have it exactly right. Perhaps someone else remembers the details.
That's extremely ROUGH to rule of thumb, and you are not doing standard statistics on a mean. We are talking about an "on off" characteristic. 1 = success 0 = failure Error is much more difficult to understand in these cases.
Erm... So you want to have 10 tries at 60% and be guaranteed to have 4 fails and 6 successes ?
Heh, imagine, you succeed on the less important (expensive) crafts and then you KNOW, that you will fail the next craft at 60% no matter what :)
That's not what I want.
I want a way to get A grade that does not involve a stupid and almost impossible quest in ToI, combined with a way to make it hard to get that does not make bone prices skyrocket.
I wouldn't know a mathematical way to get rid of the law of great numbers. If I knew, I would be in Vegas, playing Roulette.
I wouldn't go in Vegas even if I knew. But Nice or Monte Carl... why not ;)
It is a pure game of luck. I am sure, that there are people out there, who got a better success rate than 60%. Maybe some even got a 100% streak with 5 crafts. I guess you wouldn't complain about the 60%, if you were one of those.
I'd still complain. Just because others fail means hundreds of thousands animal bones get blown everyday. And it's the reason why they're at 1300+ nowadays. Which is an ebay/farmer reason.
When NC introduced the ONLY 60% in A and then S grade, they just forgot that they were backing the demand in mats without backing the offer ! (= more bones blown and burned and not more bones spoiled/farmed)
This is way too **** complicated. I am not nerd enough to even understand half this crap.
As far as I know it says 60%. Which means if I click that dog, it has a 6/10 chance of succeeding.
I wish I could've written that on my Applied Sadistics final oh so many years ago...
Lucky for us this isnt a college course. Its a farging game.
Which is supposed to be for entertainment. If you do applied statistics for fun, you need to take your computer in both hands right now, rip it off your desk and throw it through the nearest window.
Then rush out to the bar, order 4 of your favorite shooter (if you dont know one being that you havent left your basement in a year, ask the bartender to make something up, just make it strong.) Bring the shots to the nearest drunk woman sitting alone, give her 3 take one yourself and start getting yourself some.
For those of us that know this crap is boring as hell and just want to play a game, I say fix the A grade armor mess.
If you do applied statistics for fun, you need to take your computer in both hands right now, rip it off your desk and throw it through the nearest window.
To be honest, the only statistics I do for fun is black jack. Actually, fun and profit.
If I remember corerectly, the rule of thumb is that the error in your average measurement goes like the square root of the number of measurents. So if
N = 10, sqr(10)/10 = .31 --> 31% error
N = 20, sqr(20)/20 = .22 --> 22% error
N = 100, sqr(100)/100 = .10 --> 10% error
So if you succeed in 5 out of 10 crafts you could estimate a success rate of 50% with an error of
.31 * 50% = 15.5%
So you would guess the success rate to be in the range 34.5% to 65.5%.
It has been 30 years since I learned this method of estimating error, so I may not have it exactly right. Perhaps someone else remembers the details.
That's extremely ROUGH to rule of thumb, and you are not doing standard statistics on a mean. We are talking about an "on off" characteristic. 1 = success 0 = failure Error is much more difficult to understand in these cases.
I got out a statistics book, and found out I was very close. So like you said 1=success and 0=failure. Lets say half succeeded, so for N=10 tries we have
average = 0.5
variance = sum( (x-x')^2 )/(N-1) = 10*0.5^2/9 = 0.277
standard deviation(SD) = sqr(variance) = 0.527
standard error = SD/sqr(N) = 0.166
My original math was off because I used a divisor of N for the variance, instead of the more "proper" N-1. For large N the difference in small. Of course this range has a probability associated with it. Don't think I want to look that up.
Reguarding L2day Letter Drop Rates. Your right when you say that at one moment everyone gets letters. And some times muliples of the same letters if those mobs died at that same time.
What I found was that after killing alot of mobs over a 10 hr period that there were noticable intervals of letters dropping that changed every 60 mins. Durring one hr ever letters droped every 5 mins with a time window of 20seconds. That would be the time when you want to kill yourmobs.
Another 60mins later the letters would drop at 4mins &20seconds instead of on the 5 mins window. Key was to time your pulls and the last fatal blow for the mobs so that they died at the right time to get letters to drop. And right there you would have a 20 second window to kill the mobs and get another chance ad the letter drop server.
As far as I can tell the request to the server when you kill a mob is done at a time window interval that changes every hr. Worked Like Clock work for AOE Pole groups. :)
However the randomness of the letters that fall at that moment is random with the exception that if you kill those mobs at the same time you will get like 4 or 3 II's or 3 or 2 G's.
Now I dont claim to know if the Random enchant is bassed on said 5 min or even 15 second intervals. What I have noticed is that shortly after server reset the weekly reset you have a GREATER odds of success then later in the week say friday.
That applies to the O'e That I do. Perhaps with enchanting its linked with the "weekly" luck of the character in question. I would suggest that you lvl one characture for a week and then play on another characture for a week. Then try to OE on both charactures untill you start getting like 5 failures. Then the following week Play that characture. While the other character gets on Sleep mode not logging in tillt he following week.
At that time get both to overenchant and see what one has better odds.
Its almost like a Luck ratting on the characture for the week.
These are things I have noticed playing beta c1 c2 c3 c4. And having 2 l2day spellathons.
i am probably 2/12 lifetime on 60 and 70% crafts, but if i keep going it will eventaully work out in my favor........THATS THE LINE OF THINKING THAT WILL LEAVE YOU BROKE LIKE ME.
i am probably 2/12 lifetime on 60 and 70% crafts, but if i keep going it will eventaully work out in my favor........THATS THE LINE OF THINKING THAT WILL LEAVE YOU BROKE LIKE ME.
If it makes you feel better there is probably someone out there succeeding at around 90% because of your failures. :p
My whole craft career follows my OP. The system can and is being abused by a handful of people. Farmers are by no means stupid. I can think of a few people on Lionna that can OE to +10 or so often. I can also think of a few bots trains with really glowy weapons. Same with recipes. Fact is they aren't 60% to 70% for the person crafting. Its perdetermined string of 1's and 0's. Question is how long is the string and when the string is shuffled can that be perdicted?
Bad luck, VendettaForge! =(
In D&D online, people were complaining that certain "rolls" would occur more often that others, so the devs did a test and posted the results. Thousands of rolls, and no one result cropped up more often than any other. Players had submitted screenshots showing alarming streaks of low rolls, citing that as proof that the system used to generate random numbers was flawed.
All it proved in the end was this: God hates some people.
When it doubt about your luck remember this: Lineage2 is operated by Satan.
My whole craft career follows my OP. The system can and is being abused by a handful of people. Farmers are by no means stupid. I can think of a few people on Lionna that can OE to +10 or so often. I can also think of a few bots trains with really glowy weapons. Same with recipes. Fact is they aren't 60% to 70% for the person crafting. Its perdetermined string of 1's and 0's. Question is how long is the string and when the string is shuffled can that be perdicted?
There was a technique prior to the dice change which allowed one to manipulate odds. That is why dice were likely disabled for an extended period of time. I suppose it could still be done by crafting cheap B grade at 60%...but you'd need a lot a lot a lot of recs.
CURLY: "Hey, Moe, I got a great idea for this game -- we allow people to overenchant at a lousy percentage, and to try and craft expensive stuff that fails a lot, gnuk, gnuk, gnuk."
MOE: (after sticking fingers into Curly's eyes, and beating him over the head with a frying pan): "Moron. If we make people gather stuff for months, or create a really expensive weapon, and then they blow it, they will get frustrated and quit the game, and we lose their monthly money."
Where is Moe when you need him??????????
Except for duals, when are you forced to OE anything?
Those voices... they talk to me... tell me to make my weapons blue and shiny.... aarrrggg!!!
Actually that one is wrong. They had a problem with the random number generater in UO as well. It turned out after long testing that while over many many thousands of "rolls" you would get a somewhat even spread, but when looking closer at the number, sometimes the numbers would cluster together in alermingly high amounts, and also under certain conditions (because many things used the same generater) you could produce horrid or great results for a person far beyond what should normally happen with the statistical chance.
It is entirely possible that other MMORPG RNG's have similar flaws to them when investigated very closely.
Also another mistake I've seen many MMORPG programmers do, is to simply have it do the thousands of rolls in rapid succession, which doesn't show these kinds of problems. You need to do many different rolls over many different time intervals and watch what happen. You also need to look at the priority of rolls and see if it can affect the outcome as well (which often it can)
RNG's are not truly statistically random, they're simply made to mimic them.
Yes, there is theoretically a formula to predict the next number on the RNG table.
However, since there most likely is not a distinct RNG for each online player, the numbers are distributed over the whole playerbase all the time. As there is no way for an individual to influence the whole server population, it is impossible to predict, which of the next 1000 numbers on the RNG-table will be used for your personal next pseudo random event.
Directly after login, the server population is low however, especially, when the mob server is not yet loaded (mobs use Random numbers, even when there is noone near them). So if someone knew the formula, and knew a cluster of maybe 10-20 numbers in sequence, which would all be a success for the task they got at hand, they could risk to try to get into that interval of numbers with positive outcome.
This small chance to twink the system will likely go away within 5 mins after server up, as then there are enough players on, who do enough different and totally unpredictable (cause they are human!) things, that you cannot make any estimate on the next number.... unless the RNG would create 5000 numbers in sequece, that are all positive for your task at hand ;)
Example: Lets say, that I am a RNG. However my formular is extremely easy. The seed starts now and I am producing numbers starting at 1 and going up to 100. And cause of my flawed algorythm, each new number is the direct successor of the previous number.
My RN output would be "1,2,3,4,5,6,7,8,9,10,11,12..... 100,1,2,3....."
Now take a 20% chance. But not to easy, so getting a number of 51-70 would be a "winner".
And now we simulate the access to the RNG by PMs. The influx of PMs to me would be the sequence of getting RNGs.
Now you all would know the exact rules, but since from the momnet that I post this, noone knows, how many PMs I already got, noone knows, which number will be the next. And since it is not possible to send 2 PMs in such short sequence, that you can be sure, that there is no other in between, you cannot rig your positives, altough you know the seed, the formula and the details of the machine.
I'll get more indepth at a later time. But everyone keeps looking at a single pick of an RNG (MMORPG RNGs are very simple), and completely forget that there are hundred of systems all using the same RNG line of numbers, and all at various priorities, and sometimes a system can take multiple numbers based on the previous numbers. All have the possibility of generating "streaks" or "non-random" numbers for an individual system even though the overall random number generater is statistically random.
not nessaclary true kitten if the sead is time then there would be a verry formulaic pattern to the steps it takes ( same formula same sead awlays conna come out with the same numbr) to me that every moment in game only has one random number ( ie l2 day letter drop etc) becouse going by your theroy the odds of a train of mobs all droping the letters is next to inposibule
for the "random" %, for frop on my own, i noticed a thing at the time of pumpkin event : while i played with my warlord, i could collect about 250~300 nectar per 2 hours while the server was "empty" (morning for exemple), but i get about 50~75 nectar per 2 hours while the server was full (same mob, same area, same char, same lvl).
so, the % are altered by something and all i know is that more people you have less % you get.
maybe there a "maximum % per time allowed" ?
another word : if Majestic Helm is about 100-150Mil worth, it is the will of NCSoft, cause on japanese server some event give (at 0.01% ? lol) a 100 Mil worth item, and it is a majestic helm... (supposed to be quite cheaper).
I just failed another Elysian.... Im going to try another Basalt once I get the mats together if I fail it im going on a pk spree then quitting the game.
I just failed another Elysian.... Im going to try another Basalt once I get the mats together if I fail it im going on a pk spree then quitting the game.
People are overanalysing things. There is no "trick". There is no cover up. The chances are 60/70% just as the recipies suggest. But you need 10,000nds of attempts to even begin to start proving these figures are inaccurate. A friend of mine has succeeded 15 out of 18 60% crafts. Another has failed EVERYTHING. And me personally...we'll I'm 3 for 5 on 60% (which IS 60%) and 5 for 6 on 70% (which is actually 80%). So my stats are about right ... but there are always going to be people with good luck and bad luck streaks. It's simple as that.
Those with a good grounding in stats that have posted in this thread have got the right idea. The rest of you - I'm sorry if you have had bad luck but the system is not flawed...you are just unlucky :(
My personal bad luck is on overenchanting: 5 failures, 0 successes (all weapons, which are supposedly 66%). But I don't think there's anything wrong with that, the numbers just haven't fallen my way. I still believe that the rate IS 66%.
Or 0.06% 6/1000 chance for this to happen? That's really good odds. Some people in RL get struck by lightning. Some people in RL get struck by lightning and survive. Some people in RL get struck by lightening 10+ times in their lifetime.
Not that lightning is entirely random either, but given a billion people taking their odds, yeah somebody is gonna hit a bad streak...actually, numerous people will.
Some people in RL get struck by lightning. Some people in RL get struck by lightning and survive. Some people in RL get struck by lightening 10+ times in their lifetime.
Bartender: He's been struck by lightning... how many times has it been now, Reg?
Reg: S-s-s-s-s-s-s-six...
Chet: Six times?
Reg: S-s-s-s-s-s-s-s-s-six-sixty-sixty-six times. In-n-n-n-n-n-n-In-n-n-n-n-n-n-In-n-n-n-n-n-n-n-n the head!
People are overanalysing things. There is no "trick". There is no cover up. The chances are 60/70% just as the recipies suggest.
not true
without going into too much detail about random number generators there is technically a "trick"
problem lies in how many numbers are on the tumbler and if the tumbler is based per person or server wide
if the numbers are small enough you could actually work out the combination and increase your sucess rate to 100%
I'll have to admit to not having the will to read the whole thread.
I leave the a shields and jewels up in my craftshop overnight, and when I look at it the next day, I always look to see who succeeds and who fails. I usually have 3/4 successes, give or take (I don't see them all, the soulshots often push them up past viewable system messages.)
Personally, on Jewel crafts I'm 14/20 (give or take)
On A weapon crafts I'm 2/3.
People are overanalysing things. There is no "trick". There is no cover up. The chances are 60/70% just as the recipies suggest.
not true
without going into too much detail about random number generators there is technically a "trick"
problem lies in how many numbers are on the tumbler and if the tumbler is based per person or server wide
if the numbers are small enough you could actually work out the combination and increase your sucess rate to 100%
Yes very true. What I meant to say was more "there's no trick that can be easily or reliably worked out by players and exploited". The numbers involved I suspect are huge...and it would definitely be a server-wide tumbler as you say.
Yes very true. What I meant to say was more "there's no trick that can be easily or reliably worked out by players and exploited". The numbers involved I suspect are huge...and it would definitely be a server-wide tumbler as you say.
noone knows if the tumbler is server wide or not, thats why i say there may be an exploit in there somewhere
truth is anyone enchanting a bunch of d grade junk until it blows and then enchanting the item mathamatically has better odds than someone just randomly doing it
the advantage might be slight but technically its there
enchanting a bunch of d grade junk until it blows and then enchanting the item mathamatically has better odds than someone just randomly doing it
What?? Enchanting the item mathematically? You double-click the scroll and then click your item. Tell me, please, what are the mathematics behind that? I don't care if you blow up a thousand d-grade junk items. The next time you enchant you have a (insert actual odds here) % chance of success.
Like I've said before, unless you know the algorithm, what seed to use, and can ensure your requests gets to the server at the exact required moment, there is absolutely no exploit. If there was, people would be using it and it would be abused. Having +16 weapons on the server doesn't indicate an exploit. Having a 100% success rate on 10 items in a row doesn't either.
You double-click the scroll and then click your item. Tell me, please, what are the mathematics behind that? I don't care if you blow up a thousand d-grade junk items. The next time you enchant you have a (insert actual odds here) % chance of success.
not really
the numbers may be huge but with each successful or unsuccessful attempt they change
if the chance was say 70%, and it was useing a tumbler of only 10 numbers, and you failed 2 enchants in a row (possibly server wide) the next enchant would have a 90% success rate
those numbers are obviously moronic but like i said for each failed enchant the odds go up on the next attempt
the numbers may be huge but with each successful or unsuccessful attempt they change
That is a statistical fallacy and entirely incorrect. The RNG is client side. It was manipulable prior to the dice fix. Not due to statistics but simple due to (as previously stated) manipulating the WAY a computer pulls random numbers. It has nothing to do with manipulating the numbers themselves and you decieve yourself in thinking that way. Mathematics and statistics are FAR more complicated than most of you seem to imply. This is not a simple task, it requires extensive work.
Random number generation on a server doesn't happen in a vaccum, there are other systems that influence it too. There is no "trick", you can't really predict it, but its certainly possible for a live server to not give the desired %, give more streaks than should be statistically happening, etc.
which is what im saying
the numbers may be astronomical, but given enough attempts and infinate resources you could tip them in your favor
that being said i dont beleave there is an "exploit"
There was...dice were disabled related to the exploit. But even if you have a large large number of events *thousands* your ability to predict cumulative events doesn't change. You can predict odds on how many in a batch...etc etc. But with knowledge of part of a "population" you cannot judge the rest of the "population." It doesn't work that way...since events are continually generated.
You double-click the scroll and then click your item. Tell me, please, what are the mathematics behind that? I don't care if you blow up a thousand d-grade junk items. The next time you enchant you have a (insert actual odds here) % chance of success.
not really
the numbers may be huge but with each successful or unsuccessful attempt they change
if the chance was say 70%, and it was useing a tumbler of only 10 numbers, and you failed 2 enchants in a row (possibly server wide) the next enchant would have a 90% success rate
those numbers are obviously moronic but like i said for each failed enchant the odds go up on the next attempt
Now wait a sec...are you assuming that once a number from the "tumbler" has been used, it can't be used again? We aren't playing the lottery or bingo here...
The RNG is going to pick a random number every time, without regard to previous random numbers that were picked. In your example of 70% success rate with 10 numbers in the "tumbler", if it picks a 1-7 it succeeds. The next time, if it picks a 1-7 it succeeds. And the next time, same thing.
Random number generation on a server doesn't happen in a vaccum, there are other systems that influence it too. There is no "trick", you can't really predict it, but its certainly possible for a live server to not give the desired %, give more streaks than should be statistically happening, etc.
incorrect the last l2 day letter event we stumbled upon a time interval and were able to kill off trains at that exezact moment and theay all dropped the event items we did this relieabuley for several hours
as for you saying that blowing up some thing then enchanting another doesent increase your odds your wrong it dopes increase your odds. think about it for a moment its like fliping a coin 2o times then you get 1o heads in a row your going to say that thoes 10 heads in a row dont increase your odds of getting a tails next flip?
its a dumbed down form of counting cards no where near as ackurate but it does the same thing it increases your odds
I am not sure how anything you're saying shows me to be "incorrect".
RNGs with hundreds and even thousands of subsystems using the numbers in varying priority and speed are nowhere near anything like "tossing coins" at all. If it was, then the 8x8 system of for instance UO would never have been able to happen, yet it did.
as for you saying that blowing up some thing then enchanting another doesent increase your odds your wrong it dopes increase your odds. think about it for a moment its like fliping a coin 2o times then you get 1o heads in a row your going to say that thoes 10 heads in a row dont increase your odds of getting a tails next flip?
This is the most inaccurate statement in this entire thread! Does the coin remember its history? Of course not. Then what is it that determines what the next coin toss will result it? It's random! Does the coin, while being flipped, alter the air around it and change physics for a sec so that it will be biased towards landing on one side or another?
What you're saying is that on the first toss, it's 50/50. And if a heads came up, the odds are now, what? 45/55 favoring the tails? And if heads comes up again the odds are 40/60 favoring the tails again? Of course you know this isn't the case.
But what you're going to say now is that getting 10 heads in a row is so rare, and that tails HAS to come up soon, because over time the number of heads has to be roughly equal to the number of tails. And that's why you think the odds of getting tails increases the more heads you have in a row. This is totally wrong! The odds of getting any combination of heads/tails is EXACTLY the same. The odds of getting HHHHHHH is the same as getting THTTHTH, which is the same as getting HTTHHTH, or any other combination. Getting a bunch of heads in a row is not a special case of probability, and it certainly does not affect the future.
But what you're going to say now is that getting 10 heads in a row is so rare, and that tails HAS to come up soon, because over time the number of heads has to be roughly equal to the number of tails. And that's why you think the odds of getting tails increases the more heads you have in a row. This is totally wrong! The odds of getting any combination of heads/tails is EXACTLY the same. The odds of getting HHHHHHH is the same as getting THTTHTH, which is the same as getting HTTHHTH, or any other combination. Getting a bunch of heads in a row is not a special case of probability, and it certainly does not affect the future.
I'm not a math or statistics person, but even I can see that this is in fact true.
While the actual toss for each result is going to always be 50%, once you get a sequence of events, the probability of getting a specific sequence is not the same as each toss.
(I'm sure there is a formula out there somewhere!)
Basically however, the chance of getting a mixed set of results IS much greater than getting an all heads or all tails result.
What happens in computers is that random number generation isn't truly 'random'. Oh it approximates it pretty well, but it's still not as good as a truly random generation. Thus you often get a bit of a 'skew' towards a certain set of results depending on what you are randomizing and how big your sample sets are.
Random number generation on a server doesn't happen in a vaccum, there are other systems that influence it too. There is no "trick", you can't really predict it, but its certainly possible for a live server to not give the desired %, give more streaks than should be statistically happening, etc.
incorrect the last l2 day letter event we stumbled upon a time interval and were able to kill off trains at that exezact moment and theay all dropped the event items we did this relieabuley for several hours
as for you saying that blowing up some thing then enchanting another doesent increase your odds your wrong it dopes increase your odds. think about it for a moment its like fliping a coin 2o times then you get 1o heads in a row your going to say that thoes 10 heads in a row dont increase your odds of getting a tails next flip?
its a dumbed down form of counting cards no where near as ackurate but it does the same thing it increases your odds
No...absolutely not. Flip a coin 10 times and get ten heads, the next 10 flips have exactly the same odds to be 10 heads as the first 10. It is not like card counting.
Card counting relies on statistics of the remainder. It takes knowledge of removed cards from a defined population of 52. Probability shifts as the population is exactly 52. Thus the first prediction is 1/52, but the next is 1/51. A comparison is not a dice or coin, but a 52 sided dice...and you cut off a side every time you roll.
Instead a random number generator works like shuffling the card back into the deck. So you can still pull that ace of spades and blow up your weapon.
Do it yourself! Pull out a coin! Toss the coin until you get two heads in a row. Now predict the next coin toss. Record the result. Toss until two heads again, predict, record. Continue. Don't like this? Set up a computer program to generate 1 or 0. Only record values that are occur after a 1 1 is generated. Then record the next number and count until there are 5000 numbers. You will quickly see the odds are the exact 50/50 you saw before. You are assuming a previous sample influences the next toss. It does not. You have not removed anything from the population. It is still heads and tails and you are going to get one.
Arguing that previous instances influence future random number generation is simply flawed. I will quote a '07 Journal of Statistics Article on teaching an intro course on the subject.
Two misconceptions I try to correct are larger samples always produce a better estimate, and a long streak of one outcome must be "balanced" with a similar streak of the other.
What happens in computers is that random number generation isn't truly 'random'. Oh it approximates it pretty well, but it's still not as good as a truly random generation. Thus you often get a bit of a 'skew' towards a certain set of results depending on what you are randomizing and how big your sample sets are.
No, random is an entirely theoretical term to describe the universe. A small sample skews YOUR approximations on probability. It does NOT change the ACTUAL probability. This is the intrinsic difference between probability and "sample probability." In this instance it is just a fallacy of hasty generalization.
Again, you're talking about statistics, this isn't statistics, but a computer program trying to approximate it using the least amount of processing cycles.
Arguing that previous instances influence future random number generation is simply flawed.
They won't in real life, in large complex computer programs they can, hence why we got things like 8x8 in UO and other systems in other MMORPG.
Again, you're talking about statistics, this isn't statistics, but a computer program trying to approximate it using the least amount of processing cycles.
Arguing that previous instances influence future random number generation is simply flawed.
They won't in real life, in large complex computer programs they can, hence why we got things like 8x8 in UO and other systems in other MMORPG.
If you are suggesting anti-frustration code, yes. But again, you would need code knowledge, just like with a RNG *manipulation*, to be able to work the probability. It has NOTHING to do with RNG.
In this case you would need to know how many "fails" triggers limitations on frustration and then achieve this maximum prior to attempts. Any code worth it's salt doesn't have this issue.
In this instance it is just a fallacy of hasty generalization.
Actually, it is the fallacy of converse accident.
+++A in logic class. ^^
Well, obviously not in this part. Converse accident is deductive logic. This is the result of statistical syllogism (words like most, many, etc). Statistics with percentiles and samples are called inductive logic. While Converse Accident may be a cousin to hasty generalization, they are different areas of reasoning. It's nice you know the subject, but there is a distinction. Nice try SWE ;) But you'll have to do better :)
Back to the topic...I knewwit! >< All 5 SoM's failed, just as I predicted :/. I bet a GM is no doubt enjoying this (I developed a theory that each craft/enchant has to be confirmed by a GM with either YES or NO :p )
how about you do a bit of research in to sequential probabilities sith then tell me how your words taste
it is a fact that the odds of pulling the same number out of the deck multibule times in a row get progressivley worse
The odds remain at 1/52, and I don't care HOW many times in a row you've pulled that same card out. Ok, ok...I know what you're saying. The odds of pulling out one card is 1/52. The odds of pulling out that card two times in a row is 1/2704. So it appears that the odds are getting smaller. But when this happens, the odds of pulling out that same card again on the third try is STILL 1/52. And on the fourth try, STILL 1/52. Sequential probabilities are different than event-independant probabilities, and I would recommend you learn the difference.
Pseudo-random generators
Most computer "random number generators" are not hardware devices, but are software routines implementing algorithms. Often they are supplied with such system software as language compilers (e.g., as one or more library routines) or operating systems (e.g., as system calls). These are more properly called pseudo-random number generators, and cannot produce truly random outputs. Algorithmic information theory defines a sequence of bits as non-random if it can be produced by some computer program that is shorter than that sequence (Chaitin-Kolmogorov randomness). Pseudo-random number generators fail that test dramatically. They can usually be programmed in a few thousand bits, but can produce far larger sequences.
There are several informal definitions of randomness, usually based on either a lack of discernible patterns in a sequence, or their unpredictability by, generally, an adversary. Output from well-designed pseudo-random number generators may pass assorted statistical tests probing for non-randomness (see NIST Special Publication 800-22, Knuth, The Art of Computer Programming, vol. 2, and RFC 1750 for details of such tests). The sequences such generators produce always have patterns, in one sense, since the algorithm that generates them has a finite state, and must eventually repeat one of those states. Given the original state of the generator, and the implementation of the algorithm, a pseudo-random number generator is totally predictable. Given even partial knowledge of that state, they are often insecure for many purposes. On the other hand, it is easy to produce pseudo-random number generators that are guaranteed not to repeat on any conceivable computer within a time-frame that is millions of times longer than the age of the universe. It is an open question whether it is always possible in some practical way to distinguish the output of such a well designed pseudo-random number generator from a perfectly random source without knowledge of the generator's internal state. Much of modern cryptography rests on the assumption that ciphers can be constructed whose output is indistinguishable from random noise without knowledge of a secret key used in the algorithm.
I am by NO means a mathmatics wizard, it was my worst subject in school, yet all math, from most copmplex to basic, boils down to nothing more then basic logic ( or common sense ), something Im actually rather good at ( go fig :P ).
If NCSoft REALLY wanted a more random RNG then a set algorithm based, why dont they look at purchising a hardware RNG for each of thier server / server subsets that run RNG? An example of a hardware RGE is
Yes its abit pricey, wadda you want for a set of random events a program on a computer cant find any set pattern for?
Meh I felt like contributing. Ive also just started the game not to long ago, and dont really have a dwarf character yet, or have crafted anything higher lvl, so I have no clue. Just my thoughts.
how about you do a bit of research in to sequential probabilities sith then tell me how your words taste
it is a fact that the odds of pulling the same number out of the deck multibule times in a row get progressivley worse
I spent my life since high school in math competitions. The odds of pulling 10 in a row are bad. The odds, each time you click, are the same. It's a hard concept to wrap your brain around, so I understand your frustrations. But my point, and that of many others, is absolutely correct assuming a pure RNG scenario.
Every time you pull a card from a deck, the odds are 1/52. The deck does not change, therefore, the odds CANNOT change. Card counting involves nonreplacement. In other words, in Vegas playing blackjack, before the deck is shuffled again, I have seen 3 Ace out of say 8 cards. The odds on that last Ace are poor. However, if we shuffle the deck each hand, the odds on those Ace are the same continuously.
I suggest before you continue to talk about something you are simply uneducated about, you try some at home experiments or...take your own advice and grab a text book on statistics and probability. I own a copy of SAS for a reason...it isn't cheap.
how about you do a bit of research in to sequential probabilities sith then tell me how your words taste
it is a fact that the odds of pulling the same number out of the deck multibule times in a row get progressivley worse
I spent my life since high school in math competitions. The odds of pulling 10 in a row are bad. The odds, each time you click, are the same. It's a hard concept to wrap your brain around, so I understand your frustrations. But my point, and that of many others, is absolutely correct assuming a pure RNG scenario.
Every time you pull a card from a deck, the odds are 1/52. The deck does not change, therefore, the odds CANNOT change. Card counting involves nonreplacement. In other words, in Vegas playing blackjack, before the deck is shuffled again, I have seen 3 Ace out of say 8 cards. The odds on that last Ace are poor. However, if we shuffle the deck each hand, the odds on those Ace are the same continuously.
I suggest before you continue to talk about something you are simply uneducated about, you try some at home experiments or...take your own advice and grab a text book on statistics and probability. I own a copy of SAS for a reason...it isn't cheap.
your still mixing hardcore mathematical chance with Real Life results.
Although every time you yank a card outa the deck, the chance of say yanking out a single specific card is 1 in 52... every time.
But exponetial chances of pulling out that exact same card 2 times in a row.... 3 times... 4 times..... you should know that your chance of pulling the card out every time is still 1 in 52... but all at the same time or over a period of time counting each card, it's still exponentiallly more unlikely and your chances get slimmer every time.
If you for example, pull a 2 of hearts out .. you mark that card down, there was a 1 in 52 chance of you yanking that card out..
ok, you mark that down,
now what are your chances of pulling that card out 2 times in a row after putting it back and shuffling the deck again? first off, it's a 1 in 52 chance, but then you have to consider that it's still a chance of 2 times in a row, so that HAS to be added into the equation.
You can't deny it. You can say your chances of going from getting 1 card out of 52, to getting that same card outa of the deck 16 times in a row is going to be still 1 in 52.... your math is either utterly terrible, or your stupid.
Do you seriously believe that the odds of pulling out a specific card could ever be different from 1/52??? You say you have to mathematically consider the history of other cards that were pulled out too in order to calculate the correct odds??? If there are 52 cards, the odds EVERYTIME of pulling out a specific card WILL ALWAYS be 1/52. They aren't 1/60, or 1/100, because there are only 52 cards! How could the odds ever be any different?
Ok, so you draw the Queen of Hearts, 1/52 odds. Now, put her back in the deck and reshuffle. There are still 52 cards, you have a 1/52 chance of getting ANY card, and one of those is the Queen of Hearts. Do you honestly believe that the Queen of Hearts has a lesser chance than any of the other cards?
So, you happen to draw her again. You put her back and reshuffle. Now, there's 52 cards again, one of which is the Queen of Hearts. Guess what? You have a 1/52 chance of drawing her again. Not 1/53, or 1/100, or 1/1000.
I think what you people are confused about is that you're saying the odds of getting 2 Queen of Hearts in a row has a low probability, and you are absolutely right. But that is not what we're talking about. We are talking about the fact that Mathematics, statistics, probability don't care about the history when it comes to independant events, such as drawing cards or flipping coins. I can safely say that neither cards nor coins have any memory and are incapable of affecting any future selections.
Nobody ever said that getting the same card 16 times in a row has a 1/52 chance. Nobody even hinted at it; I don't know how you came up with that. But even if you DID get the same card 16 times in a row, the odds of getting that card again the next time is STILL 1/52. There are still only 52 cards; you have an equal chance of drawing any one of them, one of which is that same card.
If you are incapable of grasping such a simple math concept, one that is taught in any stats/probability class, then there is no point in arguing further. Either you get it, or you don't.
Of course you're right. But the odds of pulling one specific card, and pulling the same card in a row is calculated differently.
There is a 1/52 chance of getting queen of hearts, every time. But getting it twice in a row is 1/52^2, 3 times in a row is 1/52^3 etc, so the chance of getting the same card in a row gets exponentially worse.
Before this devolves any further than it has already, I just want to clarify a few things that people are trying to explain about their experiences with random numbers in the game. (based on my experiences and talking with others.)
1) The game states that there is a X% chance for a result.
2) Observed results for all %'s are close to their stated %'s when ALL attempts are combined together. (Mostly...the results for some enchants seems to contradict this a bit. >_< )
3) Experienced results per person seem wildly skewed.
Now to explain #3.
Assuming a 60% craft, we have 10 people attempt 10 crafts. (100 attempts) Most people would expect an average of 60 successful crafts. In practice the actual number of successes could vary anywhere from 45-75 easily enough. Most people would NOT however expect to get a 30 (or lower) result nor a 90 (or higher) result. This isn't to say it can't happen in any specific set of 100 crafts, but the given success rate pretty much assures that this occurance will be rare. If you did however get a set of results that was on the extreme edges of the probability you would expect later results to move your overall results closer to the stated %'s. (AKA Increasing sample size.)
What people are seem to be seeing however in game is:
5 People getting 3 Successes on 10 crafts
5 People getting 9 Successes on 10 crafts
In short, the win/loss streaks are skewed to the extreme edges of the dispersion at an unnatural ratio.
So instead of seeing:
#1 4 of 10
#2 5 of 10
#3 7 of 10
#4 8 of 10
People are seeing:
#1 3 of 10
#2 4 of 10
#3 8 of 10
#4 9 of 10
If you were to graph the results dispersion you would see something that would look more like a two humped camel (results concentrated at the extremes) instead of a more classic single 'hill' with most results clustered closer to the ideal.
THIS is what people are complaining about. NOT the total %'s, but rather their chance to get a 'streak' (win or loss, although it's usually the losses that stand out.) People become more noticably perturbed when it seems that the game likes to 'assign' them to be on the 'losing' side of the results and someone else to be on the 'winning' side.
the larger the sample... the more accurate the statistical results are.
consider this, as far as i can see, the entire server is a 60% of success, it's not specific to any one person.... which is what's causing issues.
i even know some people that have never had a +4 enchantment on any weapon fail for them, i know some other that have never had a 60-70% recipe craft fail for them... and some of these people have managed to craft or enchant a dozen times or more. Meanwhile there are those on the complete oposite end of the spectrum that have never managed a 60-70% craft, never have made a +4 weapon without it blowing up.
over the entire server, i would say that it's quite likely a 60% chance of sucess, where as per person, it ranges from 0% to 100% ... and it doesn't seem to sit anywhere really in between to often. IMO, it's reversed to a realistic result.
as for the 1 in 52 card deck again, if your mind cannot comprehend deteriorateing chance of success in drawing that same queen from a deck and reshuffling it every time after which several times in a row, know that each time is a 1 in 52 BUT combined is a much worse 1 in 52 to the power for 2/3/4/5 etc. then i'm sorry, you either failed that class you took, or you managed to make the bell curve, or know your books good enough to just "copy paste" the information you read to the exam paper and not actually put it to full use.
I wish people would stop talking about mathematical statistics, that has as little to do with a MMORPG RNG, as The Sims Online has to do with Real Life.
Well, point is people are complainging after 5-10 crafts that boohoo I have 40% success rate etc.. they're not seeing the complete picture. Statistics aren't facts, it's just a theory of how it looks to work, or a measurement.
I wish people would stop talking about mathematical statistics, that has as little to do with a MMORPG RNG, as The Sims Online has to do with Real Life.
Well sure it does, frustration code merely prevents statistical anomolies (ie perpetual failure). With L2 around for two years, as many accounts as there are....really the odds that somebody will swing a sword and miss for an hour aren't that terrible. Frustration based code may selectively eliminate chain based events, but perhaps the attempt here is to discuss "do crafts follow the same event calander as everything else" If so, perhaps they should indeed be unique. That is the only valid arguement to make.
The players who seem to think they understand statistics are very uneducated in the field, which makes me wonder if it is worth attempting to teach.
Simply put, my odds at 6 successful 60% crafts in a row are (.6)^6
After 6 ACTUAL successful events, my odds of a 7th successful craft is not (.6)^7, it is 0.6
Why? People keep saying its stupid to deny but they don't think. Odds are multiplicative.
We know the odds of 7 crafts = (.6)^7
And the odds of 6 crafts = (.6)^6
Think carefully, if we divide out those first 6 crafts, we still end up back at 0.6.
Try this concept..
You are doing a coin toss, but somebody has a weighted coin that favors heads. How can you get a fair coin toss with this coin?
Flip the coin twice.
If the two tosses are the same, ignore.
If the two tosses are different, the first one counts.
Why?
p(Heads)*p(Tails) has to equal p(Tails)*p(Heads)
So if the prob are even, there is no way to tell which will more likely occur, heads tails or tails heads, even if the coin flips heads 10x more than it flips tails. That probability rule is important...the odds are multiplicative in a 1/0 like this...justl ike success/fail. Thus the odds of one event occuring can't change from two events...just like two events can't change one...
Simply put, my odds at 6 successful 60% crafts in a row are (.6)^6
After 6 ACTUAL successful events, my odds of a 7th successful craft is not (.6)^7, it is 0.6
Do you realize that you have just contradicted yourself with this statement?
IF (.6)^6 is TRUE, then (.6)^7 MUST BE TRUE ALSO!
If (.6)^7 is FALSE, then (.6)^6 MUST BE FALSE ALSO!
Given any particular %, getting a 'streak' of results is a function of that %. YES each single 'toss' has a (.6) chance of success or failure. We do NOT argue this fact. The argument here is on how that % is spread out over any single sequence of 'tosses'. In short, the losing and winning streaks seem to be all out of wack compared to the given % for the results.
An event, after 6 successful events, has the same odds as any individual event.
7 events in series have a particular chance, but in order to achieve this, each event must maintain the same odds. I am, again, correct. You simply don't understand statistics.
6 prior events, cannot change the odds on the 7th. Flip a coin 6 times. After 6 heads, record the 7th toss. You will get 50% heads, 50% tails. If you didn't, if you got almost always tails because it is "due" then the odds wouldnt be (.5)^7, they would be (.5)^6 * p(new event)
Again...odds have to be multiplicative..
p(event)*p(event2)*p(event3)
So the probability of each event combines for the sequential event.
So by what you are saying, you have like a 5% (lets say) chance after 10 successful crafts
We know the actual probability of 11 straight crafts is (.6)^11
If the odds on craft 11 had changed, they would be (.6)^10 * (.05)
This will be my last attempt to explain statistics and probability. Some of you simply don't understand math or probability; if you've never even taken a collegiate level course in the subject, do you really think you can argue with those who have? Regardless, I suggest you take my word, and trust the rocket scientist in the room :P
In this instance it is just a fallacy of hasty generalization.
Actually, it is the fallacy of converse accident.
+++A in logic class. ^^
Well, obviously not in this part. Converse accident is deductive logic. This is the result of statistical syllogism (words like most, many, etc). Statistics with percentiles and samples are called inductive logic. While Converse Accident may be a cousin to hasty generalization, they are different areas of reasoning. It's nice you know the subject, but there is a distinction. Nice try SWE ;) But you'll have to do better :)
Do you look good dressed as a girl? Because if so, I'm inviting you down to take my physics tests for me.
I'll take the Chem tests. Those are fun. You can keep the physics tho. Take your Simple Harmonic Motion and shove it.
I have never mentioned any frustration code. I am talking about the RNG itself.
Again, you're using statistics from RL to explain an RNG, which simply doesn't work, RNG's are never truly random, there are many systems that use the RNG with different priorities according to either/both what numbers come up and what the system uses, RNG's can "streak" more than a truely random system would, RNG's can have many issues, which is why it has to be looked at from not only what you're trying to achieve but also how it works on a production server both short and long term. Several dev teams failed to do that and it lead to systems like 8x8 (Which is a completely and utterly IMPOSSIBLE system if things were really independantly random)
Thus the odds of one event occuring can't change from two events...just like two events can't change one...
I have to agree with Chaos on this one, RNGs are never truly random in the programming world. Given the right information a person with sufficient skill could accurately predict an RNG's output at any given time.
The real question is whether or not that kind of information is available for L2, and whether or not NC's programming is complex enough to ward off attempts at predicting the output.
Back in my early roleplaying days I played IRC games that relied on a dice bot to determine the actions of players. I was surprised one day to find that someone not affiliated with this specific game could predict what the bot would output every single time. While NC's code is infinitely more complicated than a simple dicebot (one would hope), the same thing could be done.
Yeah, though in MMORPGs its more of a question of finding where anomalities come up and exploit them, not really predicting (other than within the anomalities).
Yeah, though in MMORPGs its more of a question of finding where anomalities come up and exploit them, not really predicting (other than within the anomalities).
Alright let's jump to RNG topic, I'll discuss that and ignore future posts on statistics.
TRUE RNG is indeed, hard. Pseudo RNG is still applied today in cryptography even. Secure data is still sent via secure RNG techniques. Essentially, as long as you aren't using OS rand(), your RNG seed is based on far more than the server's clock. Still, if you're suggesting RNG is client side...which I could see, I could see it as possible. Considering NC already addressed an issue on the subject, I think they are trying to ensure security.
At work, I often use software like Matlab which runs a Lagged Fibonacci based RNG with some register shift. The period on it absurd...2^1000+ thats over 1...followed by 300 zero's. This RNG subject became popular (at least in the professional science world) years back and was quickly addressed by many. While it is possible that NCsoft programmers were lazy (and don't get me wrong...I can DEFINATELY see that with what they call "pvp balance") but it really isn't hard to get pseudo random numbers. Mind you, pseudo random is HUGE. Unpredictable huge. If the period on something is as long as Matlab's RNG...
well..let's say 100,000 events happen each second in L2 (these would be seeds generated)
100,000*60*60*24*365*2...
In 2 years we've seen all of 1*10^12 events. That's a long way to wait for that period to loop back. There would need to be literally TRILLIONS of TRILLIONS of events per second (which would of course completely stop the server) to start hitting the period.
Now, I'm trying to keep this VERY basic, but if NC so much as used a simple RNG system, the code would still be near IMPOSSIBLE to break because you have to gain a string of events to really be able to learn much. A number here and a number there get you nothing. And there is no way to know much else about this. If RNG is entirely client side, well, I can throw a little work on my PC and never miss, always crit, always craft forever once I finish some experiments and back calc the info.
Think about it, if this were possible, a bot mage could ensure a crit every single time he nuked or that you drop a specific item red...almost anything. Some RNG is weak, I agree, but even then you'd need intel on server boot time, clock...well just...lots.
This will be my last attempt to explain statistics and probability. Some of you simply don't understand math or probability; if you've never even taken a collegiate level course in the subject, do you really think you can argue with those who have? Regardless, I suggest you take my word, and trust the rocket scientist in the room :P
Rocket scientist eh? Two words for you: Hubble. Telescope.
Your own equation works against you.
odds = p(event)^tries
Every time you increment [Tries] the [odds] get smaller. You cannot do anything else or the equation is no longer mathmatically true.
THUS, if you have (.6)^10 = .006 odds
(.6)^11 (HAS TO BE!) = .0036 odds
Again. We are not arguing about the probability of any single event! We are arguing about the *distribution* of a sample of events! As soon as you add (^Tries) into the equation the chance of getting successive results gets smaller! Thus the longer your streak is, the less likely it is that the streak will continue.
What we are saying is that although the total results *may* match the given Probability (60%) of a craft succeeding, the distribution of failures and successes does NOT appear to agree with the probability of their distribution. (Most likely due to the improper use of a poor RNG.)
Simply put, my odds at 6 successful 60% crafts in a row are (.6)^6
After 6 ACTUAL successful events, my odds of a 7th successful craft is not (.6)^7, it is 0.6
Do you realize that you have just contradicted yourself with this statement?
IF (.6)^6 is TRUE, then (.6)^7 MUST BE TRUE ALSO!
If (.6)^7 is FALSE, then (.6)^6 MUST BE FALSE ALSO!
Given any particular %, getting a 'streak' of results is a function of that %. YES each single 'toss' has a (.6) chance of success or failure. We do NOT argue this fact. The argument here is on how that % is spread out over any single sequence of 'tosses'. In short, the losing and winning streaks seem to be all out of wack compared to the given % for the results.
actually he didn't contradict himself.
he was saying that going from 0-6 successful crafts is to the power of 6... but the event of moving from 6 to 7 alone.. is still a single event chance.. not a combined chance.
This although rings true, but if you consider all previous events it is still indeed a combined to the power of 7 chance.
Heres where we are all having a hard time figureing out.
A single event is still based on the single event chance, a combined event is based across several events..
going from +15 to +16 on a weapon is still saposed to be a 60% chance of success. However going from +0 to +16 is horridly unlikely.
And the chances of making it from +15 to +16 although 60%, it still falls into the catagory of being a very low chance of that 60% chance actually ringing true for that person.
So, to just try and rehash.
We are dealing with multiple RNGs if you will. Where a specific action is saposed to be 60% chance, but that 60% chance also relies on previous system that measures your chance of success as well.
I mean pull a queen of hearts from the dech for the 7th time is still 1 in 52... but then it also falls back on the previous attempts and is then 1 in 52 (to the power of 7) there.. It's combined, Will you fall into the 1 in 52 chance of the one in 1 trillion chance, who knows. Thos are mathematical factual numbers that in the end, may never actually ring true.
Thus the longer your streak is, the less likely it is that the streak will continue.
Entirely incorrect. We are no longer discussing probability however. If you wish to learn more about your errors I suggest you pickup a text or enroll in a course at your local university. You simply do not understand, and it is merely annoying people in the thread with the fact that you are uneducated on the subject.
PS - The Hubble is perhaps one of the greatest exploration successes in history. Move a mirror on the ground and see how easy it is. The Hubble is absolutely a legend in our scientific history and the men and women working on the project captured images that have only begun to be fully studied and understood. Success is 90% persperation. 700 failures to create something nobody else had ever done before is absolutely acceptable. Thank you for continuing to demonstrate why your lack of intelligence prevents you from understanding. Now leave your ignorance out of the thread.
I crafted my Basalt hammer and S jewlrey with out fails.
Basalt Battle Hammer
Tateo ring
Tateo earrings 2 of them
Tateo necklace
All in a row without fails WTH?
Maybe it was the PK spree thing. I did craft these at a different time of day than I normally craft.
what's funny, is the fact that the huble was made incorrectly.. the lense was made to the oposite of it's specs.
So it made things look smaller? :eek: :p
well so small that you couldn't tell wth anything was .. then yes.
no it made one hell of a distorted image... huge problem.
they corrected it by using a second lense to reverse the error. A much smaller lense fit pretty much into the camera that was taking the pictures. Although the results after are great.. they could probably have been better
Nope. Doesn't bother me at all. One of my rules of L2 is "someone will always have more money than me." So, gratz to you.
However, what part of "this is about me" don't you get?? :p
Oh...it's about you alright...you see life is like a box of chocolates...except for you SirE...your life is a bit more like a box full of ACTIVE GRENADES!
OK I didn't wanna read 16 pages of this so I just read one. There is a 60% chance that EACH craft succeeds. After you fail/succeed a craft it is still 60% chance on the next one. There is no OVERALL 60% chance of XX items. Each item can roll let's say 100 numbers and 60 numbers are winners 40 numbers are losers. You can end up on those 4 or those 60 in each craft.
that seems to be exactly what people fail to understand. i suggest they all watch the first few minutes of "rosencrantz and guildenstern are dead". it has pretty much the best display of how odds work that i think ive ever seen. :)
this is exactly why i dont try to overplus anything. my luck sucks!