Shipwright Archive
Thread: Firespray disk hunting tips
DVad
Sun Nov 07, 2004 6:16 am
#53
Now I no longer know what you are talking about. The discusion was about how to seed the random number generator. you said in response to that discussion "it could be tied to the interval the user takes inbetween clicking his mouse...". now you are saying that you weren't talking about the seed, but instead are talking about multiplying the random result by this time interval. why? thatserves no purpose. the results of rand() arefairly random. the time interval between consecutive mouse clicks isn't, at least not in the way that the computer can measure. the fact that I can't predict it doesn't make it random. but to tell you the truth, computers are not really verygood at measuring time at which an event from a mouse arrivesany more accurately than a millisecond or so. they tend to round off any attempt to measure intervals of time from an inputdevice to the nearest millisecond (roughly speaking).You may think in your head that the time interval you speak of is1.2343847464637363363, and it may beso. but when the computer attempts to calculate that time interval, because of round off error at the input device, it is going to be something like: 1.2344. A human being is actually capable of clicking with enough precision that the interval you want to calculate will come out exactly the same 100 times in only a few hundred mouse clicks because of that round off error on the input.
No. Real time user input is a horrible way to try and get a random number. For two reasons. 1) human beings are creatures of habit. and 2) the input deviceshumans use to communicate with the computer are only as accurate as they need to be to keep up with a human's reflexes.
This is what i am talking about. In computing terms rand() is not random, as was stated above (by the guys from MIT(sic))
Why do you say "fairly" random....lol
Come out exactly the same 100 times in only a few hundred mouse clicks------yes BUT it is then * by the rand number generated.
The point of where i came into this argument was that there was no way of getting a truly random number generated. I cant remember now when/which computer it was but when you switched it on and ran a certain program on it to generate a random number it always gave you the same one, as this is how computers work.
IF you can get truly random number then there is no way people can utilise threads like this one...lol
PetaByte32
Sun Nov 07, 2004 7:29 am
#54
Malitevv wrote:
PetaByte.
Marzuk has every right to have an opinion on this subject. constantly flaunting the fact that you've RE'd things and he hasn't is basically nothing more than baseless taunts and insults that have nothing to do with the question at hand. Yet whenever he makes a good point that you don't know how to refute.... that's what you start doing: flaunting the number of things you've RE'd, as though this fact is somehow your secret trump card that allows you to win every argument. Constantly bringing that up is not the way to form an argument.
You then accuse him of needing to invest more skill points in logic! Until you stop flaunting the number of things you've RE'd as a way of invalidating the words of people who disagree with you, you've got no place to be talking about who should and should not be working on their logic skills.
I have some food for thought on my anecdotal experience with REing. Two night ago I was grinding through some RE components when my cat jumped on my desk and swatted my mouse with her paw. Guess what happenned? The next thing I RE'd yielded a schematic piece! I didn't get another piece for a long while even though I RE'd a few hundred more components. then my cat came into the room and started bugging me. Sure enough, my next RE attempt yielded another schematic piece!
What does that tell us?
Doesn't mean squat if you ask me. But in all honesty, none of your"tips" are any different than my experience with my cat. The only thing that is different is that it's obvious to even the untrained, that my cat couldn't be affecting the RE result. But in truth the statistical correlations between my cat and my RE attempts are just as strong as the correlations you think you've observed when moving things around in your inventory while REing.
Everyone is entitled to their opinion. I have never once said I was right or correct. Yet he does. And he presents his "I am the computer god so burn in hell Petabyte, you moron" additude badly.
As I said if he wants to discuss this then fine. I am open to discussion. But when someone calls me a moron or talks down to me with each post then guess what? Its not a discussion anymore. Its one man who thinks he is socorrect it gives him the right to treat others against his arguement like dogs.
I think it was him who suggested I go play Diablo 2. This is about what I expect from one of the script kiddies over there. He should go play Counterstrike. He would love that community.
So I no longer care for his opinion. I will no longer acknowledge any of his posts ever. As far as I am concerned he doesnt exist. When he can offer intelligent thought filled banter without nasty verbal slashes then we can have discussions. Until then if I want something like what he offers I will go talk to my 13 year old nephew. They both sound the same.
Tyranus
chyenfel
Sun Nov 07, 2004 9:38 am
#55
just my two cents:
the probability of a random event in a series of random events doesnt relate to its successors or predecessors as you have to see every event on its own. So getting 8 2/8 components in a row is exactly as probable as getting one of each component.
speaking of seed values, real time clock and user input you have to consider that the random number generator is running on the server. so for one it very likely doesnt have access to a delay in mouse clicks. furthermore you dont seed a random number generator every time you use it but just the first time. I wouldnt be surprised if they utilized some kind of central rnd routine that serves numbers to all kinds of functions. Even if it starts anew every use you cant easily hack a rtc seed since its using the servers clock.
You dont use values as inventory count, present items in seeding the rng but in calculating the success out of the random number.
finally observations as the ones tyranus makes are something we need. You shouldnt flame him but log into the game and try to approve or disapprove his assumptions. creating theories and proving/disapproving them is the only way to find a system. apart from that many people are just glad to find anything that could help.
DVad
Sun Nov 07, 2004 9:54 am
#56
finally observations as the ones tyranus makes are something we need. You shouldnt flame him but log into the game and try to approve or disapprove his assumptions. creating theories and proving/disapproving them is the only way to find a system. apart from that many people are just glad to find anything that could help.
I am not sure if this is aimed at me but if it is.
I did not flame anyone, I was just trying to sort a way out to make it random.
the probability of a random event in a series of random events doesnt relate to its successors or predecessors as you have to see every event on its own. So getting 8 2/8 components in a row is exactly as probable as getting one of each component.
yes thats right but much more likely to get some mixture
speaking of seed values, real time clock and user input you have to consider that the random number generator is running on the server. so for one it very likely doesnt have access to a delay in mouse clicks. furthermore you dont seed a random number generator every time you use it but just the first time. I wouldnt be surprised if they utilized some kind of central rnd routine that serves numbers to all kinds of functions. Even if it starts anew every use you cant easily hack a rtc seed since its using the servers clock.
Yes I am wrong here. I was getting mixed up with running a prog from new everytime.
So tbh it should really be random.....as in no way can anyone try to make a pattern out of it if the generator has been up and running for a while.
PetaByte32
Sun Nov 07, 2004 10:33 am
#57
A few new things for everyone to chew on.
1) Loading and logging might very well affect the outcome of getting a disk or not. When I go and collect around 900 loots off the bazaar (7 factories + 1 bank + 1 backpack + 1 personal inventory) and then I RE it all and not loading I find maybe 1 disk. Sometimes 2. However when I RE as I go I tend to find 6 or more disks. RE as I go means I go to a bazaar, pull off the parts I want to RE, RE what I can there, then move on to the next city and repeat. Not saying its fact. Just saying loading/logging might zero out your chances each time.
If this is true then it supports the idea that the chances are tracked and you have less of a chance the more disks you find. Which is quite possible and one way of controlling the amount of disks out there. Next time I feel like doing this I will go and collect loot then log a few times while staying in my city. If I get more disks then maybe there is something to this.
2) I now have around 60 disks. I need to count them again. But the thing is I have only ever found 2 disk 6/8. I have 5+ of each of the other disks except 6/8. This is really making me believe that disk 6/8 is the hard one. Whether its the hard one just for me or just for SWs on my server I dont know. It may well be that its for each person. Quite possible each shipwright has some sort of variable in the loot system that gives them one disk less common then others.
Of course the question is why do this? Well look at what happened with me. I had lost disk 6/8. My tool had ate it. So I was contacted by another shipwright who had disk 6/8 and wanted to make a deal. A deal that worked out for both of us. Thanks again Volket. This could be a system within a system to keep the shipwrights talking to each other and more social with one another. The DEVs love things like this. Look at AxP. Its only real use is to make us social and force us to talk to each other. You could play the whole game without ever talking another soul except for AxP.
Anyway just somethings to consider. Would love for someone to disprove or provethis by the way.
Tyranus
Marzuk147
Sun Nov 07, 2004 12:36 pm
#58
"I think it was him who suggested I go play Diablo 2. This is about what I expect from one of the script kiddies over there."
No, I did not suggest that you go play Diablo 2. I only stated that I saw this EXACT kind of discussion for 2 years, in Diablo 2, and even when code was examined and massive statistics compiled against it, it never went away. It was so prevalent (and likely still is) that there are entire pages on those myths.
No, I did not suggest that you go play Diablo 2. I only stated that I saw this EXACT kind of discussion for 2 years, in Diablo 2, and even when code was examined and massive statistics compiled against it, it never went away. It was so prevalent (and likely still is) that there are entire pages on those myths.
RobertPalpatine
Sun Nov 07, 2004 12:56 pm
#59
so far i have done close to 500 components of all levels 1-7 and i have only gotten disk 6 off a level 1 incom yes very random.
DVad
Sun Nov 07, 2004 2:09 pm
#60
Still don't know what your point is though. You still haven't explained why multiplying a result from rand() by the time interval between mouse clicks serves any useful purpose. We could just drop it. It doesn't really have anything to do with this discussion either.
I said this from above. I have already said it will serve no purpose because my program knowledge is zero.
I was just trying to introduce another rand number. Never mind lets drop it....
But just 1 niggly bit (lol) I meant the time from starting to RE kit to the time to actually clicking to RE it....
But hey your right lets drop it sorry for all this stuff on this thread maybepeta should try starting it again.
Scythe
Sun Nov 07, 2004 2:14 pm
#61
I would tend to think that the server random # gens go on something of a cycle, aka not truly random. Consider this: Back in the day before there was a master Smuggler to make Mol clamps, a friend and I grinding up containers discovered that by rolling a chance cube IMMEDIATELY before each slice, whatever color came up would be the color wire to cut. This worked over 90% of the time. So it seems that the random gens do follow some sort of a pattern, granted a slightly offbeat example.
PetaByte32
Sun Nov 07, 2004 2:18 pm
#62
Malitevv wrote:
I agree that he is being a jerk about it. But his science is right.
PetaByte32 wrote:
Everyone is entitled to their opinion. I have never once said I was right or correct. Yet he does. And he presents his "I am the computer god so burn in hell Petabyte, you moron" additude badly.
As I said if he wants to discuss this then fine. I am open to discussion. But when someone calls me a moron or talks down to me with each post then guess what? Its not a discussion anymore. Its one man who thinks he is socorrect it gives him the right to treat others against his arguement like dogs.
I am not saying if he is right or wrong. I feel he is wrong because of several things that have happened since REing loot. And I dare say if he made a shipwright and REd as much loot he might even come back and say the samething. I doubt that though as he seemed the type of person who would rather lie then admit if he was wrong.
As for me if I am wrong then so be it. No pride lost for me.
The main reason I consider this not as random as we might think is this. I got 4 disks in a row before. 4 RE attempts and 4 disks. This is what really has me thinking there is something not completely random about this system. Yes it maybe random to a point. But I believe there might be hidden variables added in.
The big one I am thinking about now and want to test soon is the logging/loading theory. Crafters have told me if you start getting too many crit fails that you should log to reset your generator. And as I stated above it seems when I have to load between cities I get more disks. Here is what I know so far on this.
Sitting in one area and not loading or logging I RE 900 loots and get 1 disk in the first 100 loots and nothing after. I have done this several times.However going from city to city I get a disk inthe first 100 loots then nothing. But when I load to another city I get another disk in the first 100. This is starting to be the MO for me. No more grabbing a ton of stuff and bringing it back to RE it in the safety of my city. Instead I go from city to city and sit near a bazaar, usually sampling something for survery xp, while pulling loot off the vendors to RE. This would mean that loading or logging zeros out or resets your generator. Which in turns means there is a way to "force" a disk. Hence not as random as we think. If there is a way to set up a disk find situation then the system isnt as random.
Now this also brings to light another question. If this is in fact the case, then is it possible for the opposite to happen? Where in your generator is always bringing in successes. Possible and maybe why I had the feeling of things coming in waves. But there I would have to try not to get a disk and I cant see how that is possible. Plus if I ever got into a good set of rolls I surely wouldnt want to log and risk losing that streak.
Also another thing I thought of recently is what if force luck has something to do with it? I really would like some feedback on this from other SW with and without force luck. I have luck IV and listening to other SW on my server and on here I think I am finding more disks then they are. Of course I am quite possibly REing alot more thenmost too. Which is why I would like some feedback on this. If force luck does have something to do with it, it doesnt disprove if its random or not. Just proves you can increase your chances. I have definate feelings about this though as I would prefer all FS and jedi stuff to stay as far away from JTL professions and systems as possible. But who knows.
Tyranus
Malitevv
Sun Nov 07, 2004 2:25 pm
#63
PetaByte32 wrote:
I think here is the general problem. I can feel a pattern. I dont know what but I can feel it. I think the situation is that most people RE loot randomly so they see randomness. But when you spend 3+ hours straight REing loot everyday. Well thats another matter. Its almost like a pulse. I can start to feel its peak.
You forget something very important with this comment: if the random number generator is server side, there is surely a single random number generation algorithm that many different unrelated random events are all using. That means people RE'ing and people slicing (for example) are all getting their random numbers from the same generator. This means the only way you could ever directly observe the actual random sequence is if you were the only person logged into the server. And even then, you have to remember that every time a new spawn appears, it's exact location and exact type are both randomly selected. So even if nobody logs in, you will not be the only thing querying the servers random number generation algorithm. Since the nature of the other queries is not patterned and will be random, there is no way that the servers use of rand() to produce random numbers will ever produce a pattern you can observe even if you RE a million items in rapid succession. If you want to look for patterns that is fine. But you should drop any pretense that the use of rand() or any other psuedo-random number generator could ever be the explanation for them. In this case, it can't.
But the biggest problem of all is that there is so much. So many different variables that can and I believe do play a part in getting a disk or not. Whether or not you log,
Whether or not you log can only play a role if the developers were stupid enough to use a client side random number generator that reseed with client side information when the game is turned on. OR if they were stupid enough to give each player his own random number generation algorithm that is run server side but uses client side information to seed. Neither makes any sense. Unless the programmers are very silly, each server will have 1 random number generator that gets seeded when the server resets and no amount of logging or relogging is going to have any affect on it because it is being used by the server for all random results for all players and zone spawns that it ever needs to calculate.
if you keep your disks in your inventory, company the loot comes from, and so on. And this may well be what the DEVs are counting on for their randomness. Keep the system tied up in all sorts of variables so that no one can stick even one part of the system to a table.
If they do anything other than have the server call a psuedo-random number generator to produce the random results then they are fools. Because nothing else that they could come up with will ever be as random as what a server-side pseudo-random number generator could produce. Why would they waste time relying on anything other than a pseudo-random number generator for their randomness? It just doesn't make sense. Because psuedo-random number generators arethe best possible way for a computer to produce randomness.
I am not saying that they arent using a random number generator. I have never said that. I am saying there might be variables that affect the outcome of each RE. A simple number gen that could be tracked except having 50 variables added in would give a look of randomness just as having a complex multi tier number generator. It would look random until you figured out the variables. If there is 50 variables, what are the odds someone will figure it out? Not very likely.
I'm not saying that there might not be variables that affect the RE, but if there are it cannot be to help make it more random. Every variable that they might use to affect the RE result will only serve to make it less random because variables like you are discussing can be known by the client. They can be tracked and the results of those variables can be studied, whereas the calls that the server makes to a pseudo-random number generator cannot be known tracked or studied by the client.
No. The simple fact is that if they used any variables that are controlled by the client to affect the RE result, the only possible reason they would have done this is to make the result less random. Because that is what using those variables will do.
The more I RE the closer I feel to nailing something down in this system. Look at the original FS path. Before anyone knew what it was. People had theories. I remember a few people had the theory it was based on professions. No one had any real evidence. But these guys strongly felt mastering professions was the key. But they got flamed by everyone else because the DEVs had said the system was random and certain checks had been put in place. But we all found out those few were completely right.
Funny you bring that up. Everybody knew it was about mastering professions. The flame wars on the old Jedi forums were notabout whether or not mastering professions was the key. The flame wars were about whether or not you could predict the next profession from the previous professions once your holocrons went silent. And what we found out was that all the crazy theories that people were coming up with about how you coud affect or predict what your silent profession was going to be by doing things on the client side were completely wrong.
Nonetheless, the arguments that those people gave were the same arguments you have provided in this thread.I don't mean that as an insult. It's a fact. That is what drew me into this thread. I read it and said, PetaByte andTyrannusareusing the same arguments that all thefalse prophets that claimed:"I can predict your silent profession" on the jedi forums were using back in January/February of this year.
I'm not saying you are one of those people, but youmight want tobe careful. You are using the same scientifically faulty logic that replaces "gut feeling" with cold hard facts, that those people used. And they were just as wrong as all the people that flamed them said they were.
My biggest point in the arguement for whether or not this is truely random, is the prize itself. The firespray. I honestly feel the devs are starting to learn from past mistakes, IE holocrons. Holocrons were a completely random loot item and with in a month of them dropping, people that wanted themhad literally 5 or more of them. They were so prevelant you could go to coronet and see 10 or more players shouting to sell them. If the firespray disk system is completely random as some think then it wont be long before we see the samething with firesprays. With the amount of loot out there its inevitable. If firesprays get that common then it defeats the purpose and will hurt the economy once again. An economy, I might add, that cant take too many more hits like holos, crystals, and pearls did. This is one reason I feel the disks are truly 100% random. There has to be something in there that the DEVs can control to make sure things dont get out of hand.
That is a faulty comparison. The frequency of holocrons was never a problem with the old jedi system. Shoot. Were you here in december of last year? They gave every single player a free holocron as a christmas present, and for a very large number of players, that was the only holocron they ever owned. The holocrons were never the hard part of the old jedi grind anyway. The hard part was that once you did your first 4 holocrons, the holocrons went silent. It didn't matter how many you had. They wouldn't tell you anything.
In any case, as I've already explained, all of the variables you are hoping are there can only serve the purpose of making it easier for players to get their schematic. It will not make it harder. If you think they made some mistake with the holocrons that they need to "learn" from that affected the frequency of the holocron then there is a very simple way to do that: make the frequency with which the schematics randomly appear smaller. Adding client controlled variables as a means ofreducing the frequency of loot makes no rational sense at all. It would be monumentally foolish for them to do what you are proposing if the reasons you give are the actual reasons.
Tyranus
Message Edited by Malitevv on 11-07-2004 01:28 PM
DVad
Sun Nov 07, 2004 2:37 pm
#64
In any case, as I've already explained, all of the variables you are hoping are there can only serve the purpose of making it easier for players to get their schematic. It will not make it harder. If you think they made some mistake with the holocrons that they need to "learn" from that affected the frequency of the holocron then there is a very simple way to do that: make the frequency with which the schematics randomly appear smaller. Adding client controlled variables as a means ofreducing the frequency of loot makes no rational sense at all. It would be monumentally foolish for them to do what you are proposing if the reasons you give are the actual reasons.
Roger that
Malitevv
Sun Nov 07, 2004 2:39 pm
#65
PetaByte32 wrote:
2) I now have around 60 disks. I need to count them again. But the thing is I have only ever found 2 disk 6/8. I have 5+ of each of the other disks except 6/8. This is really making me believe that disk 6/8 is the hard one.
this reasoning is horribly flawed. If you have 60 disks, and they were all randomly chosen, the probability that you would have more than 5 of every single of the eight disks isnot significantly different fromthe probability that you would have only 2 of one of them. So the probability of the thing that you think you should have observed is not significantly different from the probability of what youdid observe, yet you believe that your observation means something.
The fact that you believe this demonstrates that you do not understand statistics or probability.
Your observation is a perfectly normal and expected result of a completely random process. It doesn't provide any evidence at all that any one schematic is more rare than another. Quite the opposite.
Message Edited by Malitevv on 11-07-2004 01:41 PM