Guest Article: Diablo III Loot Source/Type Breakdown


Site reader and contributor Varquynne sent in a very nice piece that must have taken a lot of work to put together. He watched a full beta play through and recorded every item that dropped, what type of item it was, including gold quantities, and what sort of monster or chest dropped it. Not content with just a list, he turned the data into an awesome infographic, with colorful pie charts and stuff.

If you’re wondering how many items dropped… a lot. Click through to see the full chart and a lot more text info, but just to give you an idea of the calculations, there were 107 white, 31 blue, and 2 rare items, nearly 9000 gold, and over 1300 total “loot sources” which includes chests, barrels, urns, monsters, weapon racks, hollow logs, and so on.

Here’s the start of the article; click through to read the whole thing and see the cool infographic. Also, if you’re as surprised as I was to learn that he counted all this from a movie, never fear, we’ve made sure that Varquynne has beta access of his own, now.

Where does all the loot come from?

I was recently inspired by a news post on the official Torchlight 2 website that took a look at the potential loot drops in the game. DiabloWikiErich Schaefer and one of the other team members, Adam, put together a great infographic that showed the breakdown of loot from a play session and the sources of that loot. Usually, those data tend to sit in tables on spreadsheets that, for most of us, aren’t really very interesting to sift through. However, data can be really cool, especially when it’s visualized in the proper way.

There exists a proportion of Diablo players that are data fiends. They strive to maximize the efficiency of their characters’ statistics and/or play style to increase their killing power. They pour over numbers and analyze the geometry of level layouts to give themselves the edge. Consequently, these tasks increase their chances at getting great loot. The net result of that data crunching? Generally, it leads to greater quantities of awesome loot, more character power, and usually a cooler looking hero. So, data gets transformed from numbers to something aesthetically pleasing and fun!

Following the team’s example from Torchlight 2, I decided to package up some loot data from Diablo III into something that I hope is at least somewhat nice to look at and fun for some of you to go through. Since I don’t have access to the beta, I compiled the data from a video of one playthrough with a freshly made character (beta, patch 15). The player takes a witch doctor from level 1 to level 9 in a session that is approximately 1 hour long. The data from the loot drops can be found in the accompanying infographic. It breaks down five main categories of loot: gold, potions, normal gear, magic gear, and rare gear.

Click through for the rest of the article…

Caveat: The comparisons and data analyses presented here are done with one trial comparison from each game. As such, it’s important to keep in mind that there will be substantial variability; not only in fluctuations due to random chance but also in the inherent conditions of each trial (e.g. variance in playstyle). These data should not be taken as rigid evidence for mathematical comparisons one way or the other. Simply stated, they are inconclusive. So, with that out of the way, I hope you enjoy.

Comparing loot variety – Torchlight 2 and Diablo III

Compared to Diablo III, Torchlight 2 seems to throw more loot variety at you (if not more overall loot). The Torchlight 2 loot breakdown graphic includes spells, gems, and scrolls. While gems are indeed in Diablo III, they don’t drop in the beta segment. In addition, Diablo III no longer includes scrolls (at least not in the beta), droppable runes, or presumably other consumables such as elixirs. Another aspect with Diablo III that we aren’t seeing here are the contribution of dyes or droppable crafting components which will be included in the final build. For some of those with the game developer inkling, it might be interesting to see how the two games compare in this regard and which one feels more satisfying loot-wise.

It brings to mind the question of whether more loot variety is necessarily good. I’m not sure. Certainly, I think that way too much can contribute to feeling overwhelmed and can make inventory management more of a chore.

Comparing loot quantity

Unfortunately, comparisons of loot quantity aren’t appropriate with these data seeing as the Torchlight 2 figures come from a character that has advanced fairly far into the game (as well as other factors). Let’s entertain the idea, though. From these figures, we can look at the overall frequency of loot which each game presents you. I’ve measured this as the number of loot items divided by the opportunities for loot. Torchlight 2 (2140 pieces of loot from 7543 sources) gives you some form of item-based loot 28.4% of the time and Diablo III (208 pieces of loot from 1293 sources) gives you some form of item-based loot 16.1% of the time. Overall, Torchlight II tends to throw loot at you more frequently. It’s a trend that I think will hold up for the release versions of each game, especially if the way loot was handled in Torchlight I serves as an accurate example.

Getting the biggest buck for your bang

So, from where does each category of loot tend to come?

Gold: It seems like most gold comes from normal monsters (~32.7%), quests, and clickables/destructibles*. Interestingly, despite showing up only three times, the gold from the treasure goblin account for ~7% of the gold that was found for the playthrough. It would seem to suggest that chasing that mischievous, little guy down will be worth it for the gold payout.

Potions and normal items: Potions (~31.4%)and normal items (~49.3%)come from normal monsters, mostly. Not surprising, I suppose. Runners up for potions include champions and rare/unique monsters*.

Magic items: Together, champions and rare/unique monsters are the greatest source of magic items. As a note, Blizzard denotes both these monster groups as “elites”. While they may represent the source for the greatest quantity of magic items, it’s pretty much a sure bet you’ll get a magic item whenever you do encounter a special chest*, treasure goblin (if you kill it), rare/unique monster, or boss. Something else that is pretty striking is how seldom blue items dropped from normal monsters for this particular playthrough – approximately 0.2% of the time. Was this the monster group the Diablo III team tweaked the most to curtail magic drops? Comparisons with earlier patch versions could help identify where the development team made the biggest adjustments.

Rare items: Two rares were found during this playthrough. One came from a normal monster and one from Leoric (odds boosted since it was a first-time playthrough). The sampling (as is the case for many of these categories) is quite small, but it’s expected from Blizzard’s comments that elite monsters will be the principal source of rare items (statistically speaking, of course).

No legendary or set items are found in the Diablo 3 Beta.

I hope you guys and gals enjoyed the little analysis. If you notice anything else that seems interesting, let us know! It won’t be too much longer until we’re all cranking out data like these (regardless of whether it sees the light of day or not)! At the very least, we’ll see it manifest in the progression of our characters.

*Notes on loot sources:

Clickable/destructible includes everything not classified as a barrel, urn, weapon rack, armor rack, or chest. This would include things like lecterns, logs, loose stones, etc.

Champions include monsters whose names appear blue as well as quest-related monsters whose names also appear in blue (with the exception of King Leoric). For example, The Chancellor is classified as a champion.

The figure for the number of normal monsters slain is likely an underestimate. Monsters that were slain off screen by pets or Kormac the Templar may have been missed. And, of course, counts may have been missed in the chaos of combat… though, I tried to be accurate.

Rare/unique monsters include enemies whose names appear in yellow.

–Varquynne

Opinions expressed in guest articles are those of their authors, and are not necessarily the views of Diablo.IncGamers.com. If you’d like to contribute your own article or other original material, we’d be happy to see it. Use the Send News button atop the main page.

Comments

You're not logged in. Register or login to post a comment.
  1. “Torchlight 2 (2140 pieces of loot from 7543 sources) gives you some form of item-based loot 28.4% of the time and Diablo III (208 pieces of loot from 1293 sources) gives you some form of item-based loot 16.1% of the time. Overall, Torchlight II tends to throw loot at you more frequently.”
     
    I wonder if the TL2 char had a skill or any mods that increased item find or magic find. 
    And I wonder how D3’s increased magic find will affect loot drops.
     
    Really awesome job you did. You obviously put a lot of work and time into it.  What software did you use to make the chart ?

    • I want that chart on a T-Shirt.

    • Thanks!  I’m happy to see that people liked the breakdown (or at least found it interesting!)

      Tableau Desktop was used to make the charts.  Then, I used Photoshop CS4 to arrange the data visually and input the captions/data annotations/etc.

    • I bought Torchlight 1 due to the recommendations from this site, I only played it for a couple of days. The loot drops were really way too frequent and it was on of the things that made the game very tedious

      • Happily, the devs agree and refer to the items in TL1 as “item soup.”  Very rushed implementation due to the extremely compressed development time. They’ve put a huge amount more work into them this time around, and Erich says he’s happiest with the items in TL2 than in any game he’s designed them for. A list that includes D1 and D2.

  2. A really interesting read and the graph is nice! It really seems like the elite mobs will be the best bet for good quality loot just as Blizzard has stated.

  3. Keep in mind, the Skeleton King always drops a rare your first time though. After that, you hardly ever see him drop more.

  4. 1 in 3 drops something? That is way too much. You’ll be sorting loot more than you are killing stuff…

    • Not really since most of that are white/grey items, and you’ll not pick those up anyway.

      • Then why drop them so much if you are going to ignore them anyway?

        • loot explosions TM.

        • like lacdanan says, loot explosions.   I for one do think it is important to have crap drops.  The ability to quickly see that its something not worth picking up is something that takes a very brief second.   It keeps you active in the loot hunt which keeps a person engaged.  If there weren’t any crap loot, we’d all get fairly bored I think.   And as others have noted about T1, too many drops worth picking up is equally as disengaging.  If I’m constantly working my inventory, I’m not killing things.   Another side of this isn’t just items worth picking up, but items that replace something you are using.  If you are constantly able to replace gear, it gives you no attachment to items you found.  Hellgate London was terrible at that when I played it.

  5. What would be interesting to know is how much gold average drops per clickable/destructable & barrels compared to how much gold average drops per monster.
     
    Because I’m always wondering if it’s worth to bash all the barrels and loose stones or not…

    • Me too!  Especially since some barrels just explode and do damage or spawn another monster.  Would be great to have a succinct way of deciding if it is worth it to break stuff.
       
      Also curious if we can find out what ‘treasure level’ breakable stuff has vs. monsters.  I don’t remember all the details, but there was an item level or treasure class associated with monsters and breakables in D2.  e.g. if a barrel is going to give similar item quality to a monster from 3 areas earlier – not as worth it.

  6. My wife is a statatician and she was impressed. 

    Not so much in the stats but the presentation.

    Any one know what packages were used for this? 

  7. Nice to know this kind of hard work does not go unrewarded 🙂
    Great work and hope he enjoys the beta … 

  8. I had the 3 days of access from this site, so I did many a play through in those few days.  The entire time I played I never found a single rare item EXCEPT for the first kill with a new character of Leoric.  And each time I played I hunted down every elite monster I could find on my back to Leoric.

    • Yeah when a rare drops in the beta it is something of a special occasion, outside of the one you get from the first play through of the Skeleton King.  I have had a few runs where I didn’t get any rare items from the Skeleton King even on my first run.  It happened to me 3 times I think.   Those where back in Mid Jan so not sure what patch that was and if that can even happen now.   I have just recently had 2 rares drop from the Skeleton King just as a side note.

  9. There are two things that bothers me. First is that there is not much data to go from. And the second thing is, that rare items will most likely drop more often at later levels.

    • Great points.  If I find some more free time, I’d absolutely like to expand the data (and keep a number of elements constant, namely gold find/magic find).

  10. Give this man a beta key so he can make more beautiful charts!

  11. Very good job…

    …and wel deserved bet access. Way to go Flux & Co! 

  12. I am in awe. Varquynne, are you a consultant?

  13. I seriously love how hopelessly devoted all Diablo fans are

  14. Varquynne, I have a question.   What kind of play style did they player have?  Did he OCD and break every thing he saw and look in every corner?   Did he just kind of make a B-line to the end?   Or, did he just kind of play it normally (meaning a little bit of exploring but not checking every corner)

    • It was a somewhat quick playthrough.  The player didn’t explore all areas.  They were pretty much objective/quest driven, but not to the point of skipping monsters in their path.

  15. Damnit I love statistics, I wish one of these times I would log onto battlenet and see that I have beta access. At this point I don’t think it will happen. I am just curious if the machine I have now will even play the beta. I think it would, but probably not on the highest settings.

  16. Wow, you counted this all manually from a movie? O_O Do you know of a way to mine the data directly from the game? That would be amazing. 

    • Could use image segmentation using coding in a program like Matlab.
       
      Have to watch out for duplicates, like if you press alt, then wait, then press alt again without moving. Avoiding duplicates might take some more effort to avoid.

      Also figuring out monster deaths would be a bit tricky. If you’re trying to figure out what monsters drop what, this would take a lot of contextual coding. If you just want overall statistics, this is a lot simpler, since you can just manually count the number of barrels, monsters, etc while not having to write all the data down.

      Come to think of it, manually choosing representative images from a given movie, then running those through an image segmentation algorithm could be pretty rapid and relatively high accuracy.
       
      I am now thinking of doing this…

      Before anyone else jumps in, of course there is probably a way to use something like maphack (to figure out what is lying on the ground based on either data in memory or data being sent to the user from the server), but that is riskier than image segmentation. Definitely a lot easier. Still have a problem of duplicates, but it may be easier to avoid this way.

  17. Good post, and a great wall of text at that. 🙂 People are trying to amuse themselves while waiting for the release date. Can’t blame them.

  18. Ok, Im never gonna smash a barrel ever again. According to the chart, almost everything yields more gold or better loot than barrel/urns, and yet there is more barrels/urns than anything else but normal monsters. 

    • You forgot something. Barrels and urns dont kill you … It makes them more sexy for loot especially in hardcore  😛  

      • Well the Skellies that come out of the Barrels might kill you.   When I do naked runs I tend to shy away from them (barrels) when I am already in combat with another.  Last thing I want is extra baddies to deal with when I am trying to survive a rare Unburied attack.

  19. the bottomright chart is wrong, though accurate.
    ofc there’s only going to be 1 boss, since you can name and identify him.
    but the barrels, and normal monsters – well obviously theyll be different,. but exact same spawn location etc?

  20. Good stuff! Thanks for sharing.

  21. Great post.  Anecdotally, all of my rare drops have been either from normal monsters or Leoric.  3 dropped from normal monsters and the rest from new chars killing Leoric

    I’ve played the beta for about 14 or 15 hours, so one rare per normal monster every five hours at a solid killing rate (23+ dps, but it goes up to 50+ with my WD).

  22. I will sadly be forever doomed to smashing barrels no matter how rare the good drops from them are.  I was (un)lucky enough to find a rare item in a barrel.  It has only happened once out of the 5 months I’ve played beta, but that one time was enough.  I must smash all!

  23. Awesome job Varquynneand and big up DiabloIncgamers Beta key for the guy, he really deserve one!

Comments are closed.