Recently I have been putting alot of time into farming the treasure monsters in act II inferno in hope of finding something to sell on the auction house. In this process it came to my interest that the four different kind of "treasure goblins" - so to speak - seem to have slightly differentiated loot tables. I therefore decided to note the loot that I found from farming in hope of finding some expectactions of "what to most likely get" when you kill this li'l son of a **** with a bag of gold!
I have searched a little around here and there for some elaborate research into this and generally could not find something definitive, so I decided to post my results here. IF there should be some people who have the actual loot distribution tables or have performed far more research than I have collected in the following thread, please bring this to my knowledge.
Anyway, now that the formal stuff is over, let's get down to business.
- - -
The goal of my data is to find:
1) If there is some sort of differentiated loot tables for the different treasure monsters, aka if bandits generally drop more blues compared to the goblins.
2) if there is an even or uneven chance of getting each kind of monster; hence expanding the scope slightly.
SO... What I have collected after a day of farming is a collection of data from 200 treasure monsters killed (these treasure monsters are either Treasure Goblin, Treasure Bandit, Treasure Seeker or Treasure Pygmy). The data is actually from a total of 208 monsters, but 8 of those pesky bastards either managed to run into a group of monsters or some random creature killed me and allowed the treasure monster to portal away...
I have farmed in a gear set having 165% MF and 0% GF. The % MF may inflict some of these results in the shown percentage of a given quality gear, however, the overall picture should be the same. The reason why I ran with +MF gear was simple: I wanted to farm items, and these data are merely something I came up with along the way.
It should be mentioned that the data set is quite small, so the errors are more likely to have an influence in the charts. I tried to collect as much data as possible without losing too much sanity and I found that 100 treasure monsters killed should do the trick.
Generally of the 208 creatures I found, 59 were treasure goblins (28%), 62 were treasure bandits (30%), 43 were treasure seekers (21%) and 44 were treasure pygmys (21%). It is possible that the actual distribution is 25% of each; this should be clear if a larger data set becomes available.
The following chart illustrates the average amount of gold found for each monster type.
It does seem that there is a difference in these distributions, however, if one could look at the data sheet, the amount of gold picked up ranged widely between 500gold and 4000 gold. It should be fair to say that each treasure monster on average drops approximately 1500 gold with zero gold find gear (at least in inferno act II).
The following chart describes the average loot collected for each treasure monster type.
Here something becomes quite clear:
The Treasure Bandit dropped vastly more Tomes of Secret than the other three types.
The Treasure Seeker dropped vastly more gems than the other tree types (in fact, 2-4 gems per kill).
The Treasure Seeker seems to be slightly behind on the amount of gear dropping compared to the rest of the types; but this could be due to the small sample size.
Overall, the Treasure Pygmy seemed to be the type dropping the most gear with the highest quality (again, low sample size, but also by looking at the ilvls of the gear from the pygmys - they just seemed better than the rest).
The Treasure Goblin may drop more items overall, but not necessarily the best quality. They do, however, drop nice amounts of gold.
Last, I should also mention that one pygmy had me dropping a legendary (ilvl 62 neck). On the second farming batch, one bandit dropped a set item and another pygmy did as well. While this is very interesting, the sample size is way too small to include this drop into the stastitics.
- - -
So to summarize:
Treasure Goblins are generally all-around great to kill, dropping mostly gold and magic items.
Treasure Bandits have the by far best drop chance for Tome of Secret (I had one bandit dropping 5).
Treasure Seekers drop gems primarily, usually 2-4.
Treasure Pygmys may be ahead in gear quality, but this is not concluded - the sample size needs to be improved for this
I hope you found this useful. I also know that some of you may find these statistics very rough and utterly useless, as no conclusions can be made with proper assurance of their validity. The point of this minor research - and it really is minor - is to point out that it actually make a difference to kill one treasure monster compared to one of the others. There is nothing in this research that can by any certainty assure you that e.g. the average amount of dropped rare items from a treasure pygmy is approximately 1.25 when you have 165% MF - it only shows that the different kind of treasure monsters does not seem to share the same "loot table" (I simply can not find a better word for it). It is not rocket science nor is it something that will have a huge impact on your gameplay, but it might be relevant info for some people. Some may say "I don't care about this, I just kill all of the goblins I may find" and honestly, I understand that completely. But after these statistics were made I personally skipped a lot of treasure seekers solely due to the fact that I know they drop primarily gems and therefore I could not be arsed bothering with them.
I may increase the sample pool if there should be more interest into this subject.
- Happy Hunting!
Edit2: Increased the sample size from 100 to 200. Changed the image URL and updated.
Edit3: updated with a link for the raw data, corrected minor typos and added a short statement at the end.
I'd like to see more with a bigger sample and preferably no GF or MF also, so the sample is not contaminated.
Yes I would like to increase the sample as well, but finding these creatures can be time consuming when data are to be collected as well. One thing is to farm them, another thing is to keep track of their drops while you farm them.
However, seeing the positive feedback so far I may continually increase the sample size (with MF) to get a better idea of the distributions.
I would also like to run without the MF gear to get a clean sample pool, but I will most likely only do this if I find myself the time for it. Like I said in the OP, I only carried MF gear 'cause I was already farming for items and doing this work just crossed my mind in between.
Wow, I never saw this much appretiation coming... very glad to hear from you folks and the inputs!
And I never thought this info would end up on the front page
Seeing the increased interest of this I should probably add my raw data to the original post, should people question my data - I just need to get to my own PC first.
Also, please note that the data are affected by magic find - but the overall trends that are summarized in the OP should be consistent (this means that the shown average values of higher quality gear might be lower with no MF, but the relative drop rates for each creature should be unaffected).
great job and great info BUT what is the point? if i see any of them im gonna go and kill it no matter what name he has
I added the "point" of my "research" at the end of my post.
My only concern was really the following: if there are four kinds of different treasure monsters (or treasure pixys if ya wanna please Daish), then why have four different types if there is no difference to them at all? I found a difference, primarily to the bandits and the seekers compared to the goblin and the pygmy. It's nothing more than that.
So, what can a person take this info into use for? Well, it depends. Personally I cba hunting seekers anymore since they drop something I am not interesting in getting. That's about it.
I didn't make this post to claim that I have found some extraordinary relations relevant for farming or something alike. Hence why I titled the thread as a minor insight, not something major. I just posted my findings since it seemed that no one had made such a research and the info could be relevant for some people.