I think the biggest thing shown between Ynnek and Slideshow is the excessive trendiness of the RNG, more then a definitive range of percentage for rares from commons.
Exactly! That is a significant difference with the RNG, and completely unacceptable. To go from 0.00017647% chance for one sample to 0.06% on another is mind boggling. Statistically speaking, that should never happen. Both of those were large enough sample sizes to factor out "normal" variances, and both should have fallen within the mean. No wonder so many people are p'ed off at the change, while a handful of others are saying "oh, it's not so bad". Yeah, it's not so bad if you happen to be the privileged owner of an account that has a better RNG seed than most everyone else. If it wasn't so pathetically boring, and would subject others to possible actual physical harm (carpal tunnel), I'd love to see results from others who claim to have a history of bad luck with the RNG, and samples from those who claim everything is hunkey-dorey as it is post-nerf. Unfortunately for really accurate statistical sampling, that is going to be highly unlikely to happen, and I can't say I blame anyone for not wanting to go through that turture in the name of "testing".
Ok, wait... It's been a couple of years since my last statistics class, so please anyone correct my math if I've got this wrong. But isn't the chance of rares for the above sample actually 0.00062696% chance? It's 23/36,685, right? According to the math above, with 36,685 harvestables and a 0.0627% chance to get a rare, the poster should have come away with 2,300 rares. (36,685 * 0.0627) Still about 5x better chance for someone on that account to get a rare than someone on my roomie's account... Edit: Yep, it HAS been a while since my last math class! I forgot to multiply the 0.0006 number by 100 to get the percentage. So their percent chance at a rare is 0.0627% and ours was 0.0177%. Carry on!
Insufficient sample size to draw such sweeping conclusions. More samples necessary, or more precisely, more samplers necessary. Sure, we have a boatload of individual refines amongst us. Although, annoyingly enough, not a large enough number to make too many conclusions with statistical significance. Given the extremely low success rate, an extremely large number of samples are required to reduce the error range to a fraction of the sampling rate. (Unless I completely borked the binomial probability calculation, please, someone else do one, I hated stats). BUT - I only know of three samplers. The two using similar sampling methodologies (pick a common, and grind it into the ground) - got similar results. The other, using a dramatically different sampling methodology, (a smaller number of each common from a wider range of types), got dramatically different results. Also, with only three samplers, we cannot rule out any bias's. Any of us could unknowingly fall prey to a confirmation bias, sampling bias, omissions of favorable circumstances, or a flat out misplaced comma or stack of results. No matter how careful one is, we need more samplers to accurately come up with any statistically meaningful results. * Post Nerf samplers. All the other samplers I know of were pre-nerf.
I was fine with the p. It just struck me as odd that the p and error range were the same. I don't see that very often. (Yes, I know you ballparked it, but I'm willing to trust your math.)
blah blah blah = Priceless. Before they shutdown online poker in the US you were one of my favorite types of people to have at a table. You can't honeslty expect to be taken very seriously when you say thing like this. It makes it as if you might possibly think that sacrificing chickens might improve your odds. Ynnek is at least being resonable and rational in this, and is absolutely correct the sample size is far too small to make any judgement on. Afterall you yourself believe that there are people that would run on the outside of that norm, and you blame it on the 'RNG seed'. I am more than certain that the developers are not using a static seed for each account, so that if it is in fact the case blows your 'golden RNG seed' out of the water. You have also clearly demonstrated a very strong motive, you hate the fact that people are charging a certain amount for something, you've made that clear and that makes a very strong case for someone to believe that you could potentially be in the omissions of favorable circumstances. Seriously though what amount of rares do you think should be seen on a percentage basis? (please don't say something silly like .5% I'm somewhat intrigued by this)
Aside from your snotty-ased comments (to which I just say wtf-ever dude, it's quite common knowledge that SoE does not employ a true RANDOM rng) you actually did ask a valid question: what do *I* think a reasonable percentage for rares is. PERSONALLY I think somewhere between 1:2000 is reasonable, but only if they up the rate at which you can refine, instead of one at a time. I'd even be happy with a cooldown on the refining if they up the quantity per click. I also stated that I would like to see more results from people testing this. To the rest of your smartased opinion, I won't say what you can do with it, but I'm sure you probably already know where you can shove it. Oh, and gratz, you've christened my ignore list! First person in almost 9 years!
There is no account based golden seed. There is nothing client side that influences the RNG. The only seed value exists for the first person do activate the RNG algorithm after a patch. After that it's all iterative. SoE's RNG has a long established history of trending. Each successive delivered psuedorandom number is more likely to be similar to the previous and next ones than it is likely to be dissimilar. So if you're having a run of back luck... stop. Let everyone else hit the RNG algorithm for a few hours, then jump back in. If you're having a run of good luck, mash those buttons and keep going.
I know there's nothing client-side that influences RNG results. If there were, someone would have figured out how to exploit it, and it would have been fixed asap. That being said, I'd be more inclined to believe that there is absolutely nothing going on with some kind of weird rng stuff (server-side) if I hadn't, in nearly 9 years playing this game, seen first-hand how some folks truly do have the absolute worst luck, consistently over months and months, and some folks have nothing but good luck with very small spatters of not as good of luck. I've been in a lot of guilds in this game over the years, as we all know the average life expectancy of an EQ2 guild, lol. From super casual guilds with more crafters than adventurers, to some of the top raiding guilds. I've also talked to a lot of people over the course of nearly nine years. One thing that has come up throughout that history, is the bad luck vs. good luck accounts. Ive known people who couldn't win a loot roll to save their life. Over the course of months, not just hours... And others who consistently roll high, or the people who can't walk five feet without harvesting a rare.. The vast, vast majority of people I have come across in this game have been middle-of-the-road, though, and have "streaks" like you mention. But what do you think accounts for those with consistently bad luck, and why do you think some folks have consistently good luck? And to the OP, my roomie is going to work on another refining sameple for you. She's got one 44-slot bag with nothing but rhenium in it. No other commons to dilute the sample. After she's done with those, she'll get another 44-slot bag's worth. So another 17,600 sample. Not sure when she'll be done, but she's got it down to a science with all her stuff lined up just so, and then just clicks away while watching tv, lol, so it doesn't bother her too much.
Selective memory. No really. I've been at institutions that conducted studies into this kind of thing. The basic takeaways are two flavors of the same thing. 1) People tend to remember bad events and forget good events. Your fantasy football team losing has a much stronger aversive reaction than your fantasy football team winning has an appetitive reaction. 2) Self imposed perception. People decide that they are lucky or unlucky. They remember the events that match their belief and forget the ones that dont. They think they are unlucky, so they keep a mental tally of the rolls that they lost. When they win a roll, they think to themselves that they were finally 'due', and they mentally let that event fall out of their tally, because they were 'due', so it doesnt count. Add up months and years of this and probability aggregations within normal amounts of standard deviations "feel" like they're consistently one way or the other.
I would love to help you Meirril, but as this post is now 7 pages long with out a response from a red I have given up that they are even listening to their players on this subject. I have specked out of the ability and refuse to spec back into it till it is changed to at least do stacks at a time. But I am not going to spend countless hours of my game time standing around clicking. What's worse? I have a G510 keyboard, I could macro the entire process to do all the work for me while I watch a movie. But even that is not worth the ability the way it stands.
200 Iridium - 195 LQS, 5 HQS 200 Roots - 196 LQS, 4 HQS 200 Rhenium - 194 LQS, 6 HQS, 200 Swamp Ash Lumber - 196 LQS, 4 HQS 200 Thick Bear Pelt - 198 LQS, 2 HQS 10 Tungsten Ore - 8 Pure Tungsten, 3 HQS. (One refine gave 2 Pure Tungsten)
I'll just repost the feedback I made on Test before this went live: this change is stupid. Going from 0.5% chance at a rare to 0.05% is ridiculous. 0.1% could have been acceptable, but I don't see the need of changing it in the first place. It was some "free rares". ok. If you think so. You still needed to get enough raws and sit down and break it. Now it's worthless. Note: for those who screemed at Niami, we mostly did stats on rares while in beta. It was still posted on her boards and in the beta boards that one could get rares out of commons without any stats to back it up.
I really don't expect any devs to reply to this thread. Or for that matter the tradeskill dev to respond to any post in this forum. Communication from this dev has been very one-sided which is typical for all of the devs on the EQ2 team. While I'm sure he has looked at the thread, this was done to give US information, not to influence the dev. The better we understand the process, the better decisions we can make. Or if you like, the better argument you can make for a change. If the percentage is really in the 0.0X% range we'll start getting statistically meaningful numbers with about a sample size of 100,000 commons. Even then it won't be statistically accurate numbers till we start reaching the million mark. Thus the appeal to the community for inputs so no one person attempts the insaine and tries to do all the refines by himself. Even small sample sizes like mine and Cisteros are helpful since we can just pool them all together. One thing that will throw off the accuracy is that we arn't taking into account double-rare events but I think we can safely ignore that for now in the interest of getting a general idea. Though...its possible that the rate of double events could be very close to the rate of rare events. Like 1 in 10 rare events being a double? Ugh...this could be a little fustrating but I didn't take that into account earlier.
Again I would value Niami Denmother's and Whilhelmina's opinions over any of you any day of the week...
last night I refined rares for mastercrafting: 12 tungston = 5 purified, 7 HQ 10 bubunga = 4 purified, 6 HQ 22 banyan = 13 purified, 8 HQ, 1 LQ 10 blk saph = 3 purified, 7 HQ 8 metallic = 5 purified, 3 HQ Its bad enough I had a really bad RNG this set, but did they have to insult me further?!? They just had to give me a low quality, so I burned a rare for a whopping 12 copper! I will head out to gather later today and see how well I do with the next set of rares. But on average I am getting about a 40% purified rate, I guess that is what is expected. Only thing is I dont get rares from the goblin so I have to add to my frustration with spending hours gathering them to refine
I have come to the conclusion that the chances of a rare while harvesting are character based not account based. I have a character who gets a rares about 1 in 2 to 1 in 3 pullts. I have other characters on the same account who can harvest all day and never see a rare.
