User talk:Kanda/Nicholas

Start
Nicholas has now visited enough areas to allow some speculation on what he will collect on remaining areas, hasn't he? Well, I compiled a first analysis on my user page (truth is, it did not seem wise to add this amount of data on this already long talk page). Feel free to discuss it, especially if it allows for more accuracy. Kanda 08:15, 8 November 2010 (UTC)
 * Your predictions assume that he collects trophies every time. But so far, 18 out of 82 times (22%) he has been collecting rare crafting materials, consumables or stackable salvage items. Furthermore, he doesn't always collect a trophy from the region where he is (From memory I can mention Gloom Seeds and Chromatic Scales). Finally, he has only asked for 64 of the 235 trophies in the game so far. --Manassas  [[File: User Manassas Mannysig.png]] 11:36, 8 November 2010 (UTC)
 * Hence the abundance of 'most likely' ^^. I've added a note about it, but I have to take 4 cases out of your numbers: those when he wanted an item that could be obtained by trading local trophies. The rare materials and off-zone items remain, though, but are not really predictable as far as my skill goes. Kanda 12:34, 8 November 2010 (UTC)
 * You forgot the 5 times he has asked for salvage items. --Manassas  [[File: User Manassas Mannysig.png]] 13:50, 8 November 2010 (UTC)
 * Such as Charr Hide, Luminous Stone, Roaring Ether Heart, Maguuma Spider Web and Pillaged Goods, indeed. Not exactly forgetting them, in fact I did include salvage items (except foe armors and half-digested things). For whatever reason I always considered them as trophies. Now that you say it I'll look at the list again, maybe I missed one or another. Kanda 14:33, 8 November 2010 (UTC)
 * Found 8 of them. Added them, along with a part about salvage items in the hypotheses.Kanda 15:07, 8 November 2010 (UTC)
 * There are a lot more, the list is here: http://wiki.guildwars.com/wiki/Category:Stackable_salvage_items --Manassas  [[File: User Manassas Mannysig.png]] 15:11, 8 November 2010 (UTC)
 * Thanks for the list. Added them all, except again for disgusting stuff (I assume he will never collect that, until he does)Kanda 17:09, 8 November 2010 (UTC)
 * I don't know if you want more references, but I compiled a list of all trophies, rare mat, etc. that exist and whether Nick asked for them. I didn't know that you had made a list too, otherwise I would have used your info to build my list.  I think that the work you've done can complement mine and I am ready to let somebody else do the heavy lifting for future modifications (and/or adding it to the wiki).  http://guildwars.incgamers.com/forums/showpost.php?p=5671536&postcount=2582 Zahra 14:47, 10 November 2010 (UTC)
 * A quite exhaustive work indeed. I'll see which locations I can add from my list. As for the translation in wiki-layout, I generated almost all this page automatically using the 'concatenate' function. Kanda 15:54, 15 November 2010 (UTC)
 * Kanda, you may want to make note of incidents where Nick has asked for 2 different trophies from the same foe. Rogueonion 16:12, 15 November 2010 (UTC)
 * I have added a line in the hypotheses. Kanda 18:21, 15 November 2010 (UTC)

Gz on Soul Stone
Soul stone is on your list as most likely. Well done. — Tennessee Ernie Ford ( TEF ) 18:08, 15 November 2010 (UTC)
 * Yes! This one was easy however. Kanda 18:21, 15 November 2010 (UTC)

Gz on Elemental Dust
Another successful prediction. (I probably would have missed it as likely, since Nick hasn't shown any inclination to travel to underground locations, such as Battle Depths.) — Tennessee Ernie Ford ( TEF ) 17:45, 29 November 2010 (UTC)
 * Well, he did visit the Undercity, so it would not be the first time.Kanda 20:02, 6 December 2010 (UTC)

Items that have appeared since this page is up
Have you considered sorting those based on likelihood? (Rather than the order that Nick requested them) — Tennessee Ernie Ford ( TEF ) 20:09, 14 March 2011 (UTC)
 * I could. Is there a particular benefit? Kanda 17:26, 21 March 2011 (UTC)


 * The cycle page already has the nick-request order. Sorting by likelihood might offer some insight into whether there are other patterns (and makes it easier to compare with what's left). On the how useful is it to people 1-10 scale, it's probably about a 2.5 (whereas this page is about 11.5). — Tennessee Ernie Ford ( TEF ) 18:01, 21 March 2011 (UTC)
 * Alright then. Better? Kanda 19:10, 22 March 2011 (UTC)


 * Awesome! (TEF moves the rating of this page from 11.5 [out of 10 possible] to 12.) — Tennessee Ernie Ford ( TEF ) 19:29, 22 March 2011 (UTC)


 * Still not enough. How about some chart of the remaining trophies? Kanda 20:27, 22 March 2011 (UTC)


 * Um, how about super awesome. That's got to be worth an extra .5 TEFmarks. Robdalf 14:33, 20 April 2011 (UTC)

Probabilities over 100%
How are these % probabilities calculated? Due to the fact that an event's probability cannot be greater than 100% (by definition of percent), I think that this page does not getting all the respect that it rightfully deserves. I hesitate to suggest a fix without knowing the calculation method. --War_Pig5 17:06, 2 May 2011 (UTC)


 * It's really a ratio of # of areas that contain the specific trophy to the number of available trophies. For example, if trophy A is the only trophy in two zones, then the ratio would be 2:1. He could call it 100% (sticking to the technical definition of probabilities as you describe it) or 200%, indicating that it's more likely that trophy A will show up than some trophy B that is the only item available for a single zone.


 * Given that Nick sometimes likes to request non-trophy items, it's not a crazy way to go. Kanda could call it, likelihood or Kanda's predictive ratio or some such, but probability has the right connotation, too. — Tennessee Ernie Ford ( TEF ) 17:26, 2 May 2011 (UTC)


 * Okay, I found it: "adding these scores." Combining these probabilities by simple addition is not mathematically correct, but based on the other things I read there, I infer that you don't want to fix it. Also, using your method, I calculate that the probability that you already knew this was 101%. --War_Pig5 18:55, 2 May 2011 (UTC)


 * LoL, War Pig. — Tennessee Ernie Ford ( TEF ) 19:04, 2 May 2011 (UTC)


 * Oh, I am willing to consider any equation that would make the probabilities more accurate. I just did not find the time nor the brain power to find the right equation myself. Kanda 11:03, 4 May 2011 (UTC)


 * As above: you could call it likelihood instead of probability (since the former doesn't have a technical definition). I'm partial to Kanda's Predictive Ratio though. You (and I) believe that Nick will definitely request items with a KPR >=1 unless he doesn't collect a trophy. (I suppose you could reduce the probability by 12% to get a truer KPR, based on the fact that Nick has collected local trophies 88% of the time.) — Tennessee Ernie Ford ( TEF ) 11:11, 4 May 2011 (UTC)


 * Nice update (However, what do I have to do to convince you to go with KPR? ;-) — Tennessee Ernie Ford ( TEF ) 20:15, 4 May 2011 (UTC)


 * First I'm trying to figure out how to teach Excel to tell me what the actual probabilities are. So far... not easy. The only way I can think of is the base equation of probability.
 * Also : Likelihood/KPR>=1 does not mean that probability is 100%. The best example (for now) is Mossy Mandibles: 100% L/KPR and only 71% probability. Without even taking 88% into account. Kanda 20:28, 4 May 2011 (UTC)


 * In reality, nothing is 100% probable, except in the past tense. Before an event, the best that you can say is that you have no reason to expect any other result. For example, in theory, a die will land displaying a number between 1 and 6; in reality, it could land on an edge, a corner...or fail to land at all — we don't account for that in the mathematical models.


 * Monte Carlo (gory detailed version)
 * Your current model is simple and accurate enough for another 8 months; we're not trying to send a rocketship to the moon. However, if you want to get more advanced, then you could use excel to create a model that took all the remaining trophies/zones into account in more detail: assume that...
 * Nick will appear in the zones you have listed 1x and will select a trophy from that zone with equal likelihood.
 * When a trophy is picked, you will remove it and the zone from the list of possibles.
 * You will also decrement the remaining count for any other trophies (in that zone); if the count is zero, they are also removed.
 * You can also take into account the non-local request by creating a 12% chance of something showing up not on the list (which slightly changes the remaining likelihoods).
 * Then, you can run a monte carlo simulation (in excel) for the remaining set of visits and repeat that scenario an arbitrary number of times. Count the number of times each trophy shows up across all the runs and divide by the number of runs...and voila.


 * Simplified probability
 * An easier way of getting close to the same idea is to add up all of the likelihood numbers, divide by the sum of all the likelihoods. That can be safely called a probability: the ratios will (by definition) add up to 100%. The monte carlo above will give you somewhat different numbers; this method over-emphasizes items that appear in multiple zones and under-emphasizes those that appear in only a single zone.


 * But neither method gives your readers any more information, really. The only thing people really need to know is: try to keep everything on the list. If you can't, merch only the items with the lower KPR numbers. Everything else is just fun for those of us who like discussing Big Theories. The order of items isn't likely to change (noticeably). — Tennessee Ernie Ford ( TEF ) 21:02, 4 May 2011 (UTC)

I Love You
This theorycrafting is amazing 24.78.131.249 21:27, 9 May 2011 (UTC)

issues
Feathered Crest is still on the main list, but I think it should be removed because it is now being collected. Any thoughts on moving this list out of user-space or otherwise inviting more help from collaborators?--War_Pig5 20:33, 27 June 2011 (UTC)
 * oops. I confused a Scalp with a Crest.--War_Pig5 20:38, 27 June 2011 (UTC)
 * Indeed ; Sensali give both crests and scalps depending on the area.
 * As for moving the list out of user space. The issue is that the numerical data -likelihood and all that- is a personal estimation and not facts (they are not even real probabilities). So I cannot simply put them on an official page. As per the rules, however, anyone is free to take whatever they find useful; here is where I will be updating things, though.
 * It is not that I am against help from other contributors; however, what I display here is the front-end (visible) part of a more complex work that revolves around Excel sheets and other unreadable stuff. So either I have to explain the whole thing, or I cannot copy/paste anymore due to (probably useful) contribution. Either way this is more than I am willing to invest, unfortunately (actually playing the game and fetching the trophies takes time, too). Kanda 21:55, 27 June 2011 (UTC)

Sage Lands
I notice that you don't list Ebon Spider Legs in the Sage Lands though there are a significant number of spiders there. --Valshia 17:54, 4 July 2011 (UTC)


 * Good catch. According to Sage Lands, Maguuma Spiders do drop them. I've added it back to my lists. Thanks! — Tennessee Ernie Ford ( TEF ) 18:11, 4 July 2011 (UTC)
 * Oops. Added back, I don't know why I missed them (and more importantly, now I don't know what else I have missed); thank you very much. Now I just wish I had read this last week and not sold my stock. Kanda 23:42, 11 July 2011 (UTC)