[outdated] Caustic Arrow Solo Map MFer (20/300+)
|
Serleth, Congratulations on finishing the full 40 challenges.
General question, do people up convert currency at the end of a play session, when you need a specific orb/amount, when you have a set number or value of a specific orb? I'm trying to find a good time to more constantly do this process, and wondering what/when others work on the orb exchange shuffle. |
|
|
Six-linking discussion. Contains math and a nice table listing probabilities for failing to 6-link after spending a certain number of fuses.
Spoiler
The chance to 6-link an item is likely just entirely random and is the same for any item which is capable of 6-linking. The Path of Exile program code just generates a random number which controls the outcome for whatever is being linked. Next fuse, another random number. The item doesn't affect this. Nightmare runs of over 6K fuses exist because that's the way probability works. 1300 or so fuses is the average cost for 6-linking - but let's say that you've just spent 1300 fuses and failed to 6-link an item. Is that item now more or less likely to 6-link than some other item? No. It's entirely the same. You have the same exact chance of succeeding to 6-link or not and previous linking history is irrelevant. So of course, just by small chance, some items will take 6K, 12K, 15K etc. fuses and still never link. The chance of that happening is unlikely, but it does exist. The exact way the probability curve works is going to be slightly different than the following, but it's probably not very far off. Use this math to help better understand fusing and probability of getting a 6-link. ============================= Let's say that the average cost of 6-linking is 1300 fuse. What's the chance of 6-linking in one fuse? We assume that the chance to 6-link is the same each time we fuse and is unaffected by previous attempts. Let p = probability to 6-link with 1 fuse we have .5 = (1-p)^1300 Solving for p we get roughly 1 in 1876 ------ So now for an odds table where we list the odds of you failing to 6-link after expending n fuses. Let F = chance of failing to 6-link after expending n fuses F = (1875/1876)^n Some values. Fuses spent (n) - Chance of failing to fuse (F) 1 - 99.94% 10 - 99.47% 100 - 94.81 1000 - 58.67% 1300 - 50% 1500 - 44.94% 5K - 2.67% 10K - 0.48% 15K - 0.03% ====== Use the table above to understand fusing probabilities - it won't be that far off the truth. So yes, there's a little over 2% chance that you'll be unlucky and expend 5K fuses without 6-linking. And as for failing after spending 15K fuses - yes, that's possible. It'll happen roughly 3 in 10,000 times. Are there more than 10,000 player-crafted 6-links in POE? Of course there are. So there's bound to be a few of those 6-links that cost 15K plus fuses to finally succeed. Last edited by hankinsohl on Sep 30, 2016, 8:15:37 AM
|
|
" Hear hear! Well done my friend. |
|
|
I absolutely love having a mathematician person in this thread.
Spoiler
I totally get I'm tinfoil hat about the whole thing sometimes lol
Thanks for the congrats! " Early league, I just keep an eye on the values and generally understand the trend of how currency *typically* moves. I.e., this league is different in that chaos has actually settled around 62:1 ex, whereas previous leagues it spiked to 80:1 before settling down to closer to 73ish, but based on that trend I would convert basically everything at the end of a session in order to maximize the benefit heading into the next day. But at roughly the 3-week mark, currency ratios tend to be pretty much fixed until the last couple weeks of the league. At which point, I (sorta) stop caring when I make the exchanges: It's worth noting however, that the time of day has an impact on whether you'll be able to trade at the upper or lower thresholds. This is mostly important for conversion to exalt, but I've noticed Europeans tend to be more patient and will demand higher amounts of orbs for exalts than North Americans, so if you log in during peak hours for the Western Hemisphere (5PM ET - 10PM ET weekdays, ish), you'll be able to exchange at the lower threshold. Eg, you will be able to get 60:1ex in WH peak time, but afterwards will have to trade at 62:1. In short, it comes down to desperation. If I need to move currency *now* in order to accomplish something, I do it whenever I need to. Otherwise, I tend to wait until late afternoon / early evening in my time zone in order to get the best deal. Jul 27, 2011 - Sept 30, 2018. Last edited by Serleth on Sep 30, 2016, 1:59:22 PM
| |
|
hi hank
another thanks for the great explanation. But thats what i know why i wrote to serleth thats no matter you use one fuse one item or 1000 fuse one item. But what i try to say is, that computer rng isnt realy rng. AND in poe they must do so much rng events, that they have no chance to do it with a complex algorithm. Its for such a big number of rng events to slow. i put a little artikle in a quote who explain verry nice what i mean.And i think ggg dont use a really rnd number generation. And that would explain why some accounts have much more "luck" than others. " like serleth and i said bevore i like math diskussion on that lvl. And till we dont know how ggg generates the rng numbers for chaos an item or such things or for the generation wich item drops and with what pre or affixes(there must be some rules because items cannot have all pre or affixes). Or if there is a difference between rng for linking or rng for generating items, we cant say if its more or less random. But i am really sure that they dont use a rng generator but a pseudo rng generator |
|
" ok then i cant express what i mean. I am with you that its a per item generation. I thin all other things make no sense. But if there is a pseudo rng generation, then some accounts have more positve events then other, or some ping has a influence or really the server you use, or the gateway. As i wrote bevor since we dont know how the generation is done we dont know. But i am sure its pseudo rng. |
|
|
Here's the thing.
It's really just confirmation bias. Literally all the people I've ever seen who talk about the bad luck they've had linking have only spent 2000-5000 fuses in their career on 6S items. Is it bad luck? Maybe, but it's also only enough fuses to hit 1.6-4 6L's, presuming you hit the average likelihood. Which, across a maximum of four probable successes, you statistically have a decent chance of being so-called "unlucky." The "lucky" people are the ones like me, who've spent 65,000+ fuses, and thus have a much larger distribution and sample size of attempts in order to average out to "lucky." Jul 27, 2011 - Sept 30, 2018. Last edited by Serleth on Sep 30, 2016, 4:07:42 PM
| |
" 6-linking is totally random. outliers will occur in random populations - and the outliers invariably get all the discussion. "Hey I just 6-linked in 3 fuses! I'm so lucky!" - you hear this sort of think quite often. Or, "Damn! I just spent 3K fuses and I still cannot link my chest!" - this sort of comment appears with great frequency. You never hear, "Guess what guys! I just 6-linked after spending 1200 fuse. Boy oh boy I'm average!" As a result, there's plenty of bias out there towards extremely good and bad luck. If you happen to get lucky, all the good-luck comments are reinforced. Same thing happens if you are unlucky. All those bad luck comments begin to make complete sense. We see things around us based on our own experiences - the confirmation bias Serleth mentions. === As for per-item linking probabilities, or the probability of linking being affected by ping/lag/whatever, or by standing next to Merciless Brutus while linking... well... I think you know my views on this. Client side you use a fuse. It's sent to the server. Code on the server generates a random number. A lookup occurs to see what that random number means in terms of linking. The result is sent back to the client. This is how 6-linking works. What *might* realistically affect 6-linking probability? Well, it could be that the mechanism for generating random numbers is a pseudo-random number sequence generator - likely this is the case. What if they made a programming error and didn't seed the generator each time it's initialized? Well, then you get the same sequence of random numbers after each server restart. Maybe you could try to exploit this somehow. Or maybe there's a hardware-based random number generator at work. Might this hardware be subject to subtle external factors - solar flare activity, spikes in electric current, other things? Maybe. Maybe you could exploit this too somehow. But really, just stick with the "it's totally random" hypothesis - this is both almost certainly true and if not true it would be possible to collect a statistically meaningful sample to show that it's unlikely to be true. But that brings us to the mathematics of statistics... which is quite a bit more complicated. Last edited by hankinsohl on Sep 30, 2016, 5:01:17 PM
|
|
|
CA is an awfully dull skill I have to say. I'm sure this build works super well for what it was designed but it is very, very little fun. I'm considering to run another character just due to how unentertaining this build is.
|
|
|
To each their own. I have fun with it.
Jul 27, 2011 - Sept 30, 2018.
|
