Guild Wars 2 Wiki talk:Projects/Drop Rate Research

From Guild Wars 2 Wiki
Jump to: navigation, search

Level Scaling[edit]

One thing we need to figure out is if your level (and the amount you scale down by) affects the rate of drops. I've only just started collecting data, so it would be nice to have another player at a different level also collecting data from the same mob type. If the drop rates seem to be different, we must find a proper formula to calculate the rates for every level. JKaz 02:53, 1 September 2012 (UTC)

As found here, "A scaled down character continues to receive experience and loot that is 'good' for their real level, but it is somewhat less efficient for obtaining experience and items than playing level-appropriate content.", which suggests that a character that is a higher level than the content area receives lower drop rates. This does not say whether or not the rate is proportional to the level however. We will need get a baseline drop rate (ex: lv5 character in a lv5 zone), and then return to the same area killing the same mobs with higher lever characters (every 10 levels for example). Once we get all those statistics we can start creating a formula to adjust for the level scaling, if the above statement seems to be true. JKaz 06:33, 1 September 2012 (UTC)

After collecting data from the same mob at 3 different levels (22, 42, 80) with a sample size of 50 for each test, I've determined the drop rate differences between levels. Including the fact that an item dropped 0 times during my last test (Skelk Lurker, lv80 test, which may be removed the mob's drops), drop rates fluctuate less than 5% in both directions for different materials for each level difference. This also holds true for items that are not 'good' for my level (trophies, or lower crafting materials not exactly good for lv80); they don't seem to drop less despite a higher level. Preliminary evidence suggests that drop rates do not scale with level and the differing drop rates could just be because a low sample size was used. More tests will be conducted with other mobs at different levels before any conclusion can be made. JKaz 19:25, 9 September 2012 (UTC)

Party Scaling[edit]

While reading the GW1 wiki I found that the party size affects drop rates of items. I'm wondering if this should be something we need to track of. At this point I believe we should just focus on solo drop rates.

Drop rate decrease[edit]

From what I've seen unless a mob is assigned to you in a event or a renown heart, the drop rate seems to decrease as you get more kills. Maybe its a measure against farming? Or I’m just imagining things which is quite possible. --Dr.Mobius 08:22, 1 September 2012 (UTC)

I hope this isn't the case. This would mean a lot of extra experimentation and I'm not even sure we would be able to determine at what point the drop rates start to decrease. If we get multiple samples of the same mob from different people (ignoring any possible level scaling difference right now), wouldn't this reflect what a normal person might expect from the drops? Looking for further input on this. JKaz 15:28, 1 September 2012 (UTC)
Yes it would reflect on what a normal person might expect, but make sure you have at least around 4 people each killing the mob say 50 times? Also ensure the condition they do it is the same such as no quest attached or renown heart just to be on the safe side. --Dr.Mobius 08:06, 2 September 2012 (UTC)
Ok good, that's what I was assuming and was hoping to do, given enough time and people helping.JKaz 12:23, 2 September 2012 (UTC)

Mystic Forge[edit]

I figured I would post this here as it closely fits. Should Drop rate research be concerned about recipes in the Mystic Forge. I think that would be the most beneficial at the current moment. See Talk:Mystic_Forge#Fixed_recipes_research For what research has been done so far. Anzenketh

Not sure if it could apply, but this project is only for drops from mobs. It seems like that page has quite a bit of research already. JKaz 20:29, 11 September 2012 (UTC)

Problems with methodology[edit]

There are problems with the methodology used by the wiki:

  1. It makes it difficult to determine rates from more than one drop table.
  2. It treats "3 of an item" the same as "1 of an item" and there are two issues with that:
    • We can't determine the actual mechanics: is it a chance of item X and then a chance to drop 1 or 3 ? or are the chances of 1 or 3 part of the first table?
    • We can't help people determine the market value of containers: there's a big difference between a bag that drops 3 of something often and another that drops only one of that same thing more often.
    • It's also inherently less accurate to ask people to pay attention to 100s of results rather than compare before and after -- most of us aren't very good at collecting data.

It's probably beyond the scope of this wiki's mandate to collect data systematically enough to separate out results into multiple drop tables. But we can deal with the last two, by collecting total results.

  1. Current system: player opens 1,000 items and records each result: requires 1,000 notations. Results reported as chance an item drops; adds up to 100%. These cannot be used to determine expected values.
  2. Proposed: player opens 1,000 items and records what they end up with: requires only 1 notation per item. Results reported as expected number per drop; adds up to meaningless number. These can be used to determine expected values.

GWW doesn't have this issue for the most part because results in GW1 are typically binary: you get a set amount of an item or nothing. This was true until things like zaishen boxes were introduced. However, in GW2, many of the box/bag/containers, salvage kits, and foe drops can result in multiple amounts of unrelated items. Salvaging has one table for fine mats, another for salvaging an upgrade, and (it appears) another for ectos. (We could separate those rates easily for salvage kits, but less easily for other types of containers.)

tl;dr Changing our collection method would produce more accurate data that could be more easily used by other players to determine how much they think containers/kits are worth to them. 75.37.20.28 20:36, 6 January 2013 (UTC)

Item IDs[edit]

It is possible to get the IDs of the items with the /wiki command. The ID should be added to the list of drops from containers, as we don't know whether two champion loot bags with the same name but diferent IDs will have the same drop rates or not. MithUser MithranArkanere Star.pngTalk 19:25, 11 September 2014 (UTC)

We can create a separate table for each id to confirm if the drop rates are the same or not.--Relyk ~ talk < 19:31, 11 September 2014 (UTC)