Optimal Item Farming

Optimal Item Farming

We have developed a list of best practices for optimal item farming in Diablo 3 based on observations from the last beta patch.

This article in the new Diablo IncGamers Stategy Section analyses data collected from hundreds of full playthroughs of the beta test and helps us understand how to best structure our magic find runs.

Full analysis after the jump – the results may surprise you.

Disclaimer: Blizzard has repeatedly said that beta was NOT for testing balance and mechanics, it was for testing software and hardware. All analysis is speculation as of beta patch 18 and may change post-launch.

Farming: “Repetitive runs of mastered content to obtain items”

TL;DR for optimal farming:

  1. Seek out area exit aggressively, avoid siderooms.
  2. Kill everything except small packs of normal monsters
  3. Prioritize zones that provide more items-per-minute (our data logging tool can help you get that number)

While hard facts on monster drop rates and spawning mechanics are currently hard to come by, we’ve been able to glean some information from a mix of blue posts and beta observations.



A) Any item can drop off any monster (of a given monster level)

B) Champions and elites have attractive drops rates and can be killed very quickly compared to bosses
The efficiency of boss farming will worsen as encounters become more time consuming.
Blizzard Source

C) Any monster pack is equally likely to spawn as a champion or elite
Monster locations and groupings are random with elites randomly added.
Blizzard Source

D) There are no champions or elites with static spawn locations
Beta had static spawn quest monsters, but if they are a convenient source of loot this would violate blizzard’s stated design goal of ending the repeated shenk/pindle/eldritch or tele-baal 2 minute long games.

E) Area layout is mostly random, but sometimes you can infer the location of the exit

F) In D3, we are far less mobile and monsters have an increased ability to initiate fights


Best Practices

1) Kill priority: Elites > Champions > Bosses > Large normal monster packs
Stragglers are not worth killing – move on after clearing the tightly packed monsters.

2) Backtracking is bad
To get loot efficiently, kill Elites and Champions. Since their locations are unknown, the fastest way to spawn one is to aggressively explore new areas.

Should I explore side rooms that don’t have champions or elites?
No. Side rooms are dead-ends that require backtracking, and should be avoided.

Should you take the area exit as soon as you find it?
Yes. If you run past an exit, you ensure that you backtrack later on.


Does it matter where I farm?

We don’t know exactly how drop rate mechanics will work post-launch, but with the data we’ve collected from beta using DDD  some areas are clearly better than others.


You can find more information for beta 18 specific loot here 


In practice you can’t evaluate areas in isolation
You will need to plan farming runs as a chain of areas starting at a waypoint, and look at their averaged items per minute. You also need to be careful to take into account monster level, as that will affect the quality of the item dropped.

From the data, a Hidden Cellar run looks very attractive since it has high drop rates, but the items dropped are of lower quality compared to the Cathedral. Additionally, since the run can be done so quickly (under a minute potentially), the time it takes to logoff/login/get to the starting point of the run, becomes a significant problem – adding 20 seconds to a 60 second run drastically hurts efficiency.

Once the full game is released we should be able to zero in on the best farming runs through our loot tracker, and account for the effect monster level has on item quality.


Topics in need of investigation:

1) Would rushing a new toon to a boss kill be a faster source of loot? (taking advantage of the buffed drop rate for your first kill) Nevermind, found a blue post clarifying that boosted droprate on firsttime kill of a boss is ONLY in normal difficulty.
Blue Post

2) Are area events (like resplendent chests) worth the time investment?

3) Nephalem Valor Buff – whether or not this will be able to incentivize always ending MF runs at bosses

Blizzard Source

Related to this article
  • Opening 102 Horadric Cache Bounty Bags
  • Guide: Multishot – The Speedfarming Demon Hunter

  • You're not logged in. Register or login to post a comment.

    21 thoughts on “Optimal Item Farming

    1. This seems to ignore the nephalem valor system that will be in the full game as well.  We haven’t really gotten an idea on how that will work yet but in theory boss killing will be more attractive, but only after killing a certain amount of elites/champions.

      Though the information about ignoring side rooms here still seems relevant. 

      • I think you’re probably right, and I REALLY hope you’re right. progressing through an act killing champs and then killing a boss would be a nice dovetailing of efficiency and fun, and I know blizzard is aiming for that.

        Its complicated by the fact that the neph buff will boost MF everywhere, but boost it even more for Bosses.

        The longer it takes to kill bosses (due to phases, or whatever) the harder it will be for the neph buff to make me want to kill bosses, instead of just skipping around selectively running different area sequences.

        I’ll go ahead and add neph to the list of unexplored things though, it should be in there. 

      • Only thing is until you look in said side room you don’t know if there’s an elite/champions in there or not.

        • It isn’t too rough to figure out where spawn locations are. A lot of them will be set, even in randomized dungeons.

    2. The funny thing is, I tried to do this method in the beta and the dumb f—s just wanted to lap SK.  Now that would have been fine with a sweep and clear, but they would literally run past every pack and activate the king.  Some dude nerdraged me telling me the king had the best drops in the beta.  I spyed his char and he had no mf on.

    3. This is pretty much the way I would intend to play except that I will do quests and bosses as well for extra gold and xp.  A boss kill might not have that great of drops, but it sure is a lot of XP.

    4. ” … the time it takes to logoff/login/get to the starting point of the run, becomes a significant problem – adding 20 seconds to a 60 second run drastically hurts efficiency.
      Has Blizzard expressed anything with regards to some sort of game-creation-frequency-policy that might result in temp banning?  Because string together a few runs like the one described above in Diablo II and you might be banned for a few hours or a few weeks, it was anybody’s guess back then.

      • Far as I know they haven’t said anything – I did spam creating/leaving games in beta to try and test, but never got a temp ban. Not conclusive though.

        I’d be surprised if it lives on in D3 because it did such a great job at interfering with normal play. If you had a couple friends over at your house playing D2 and you didn’t stagger your ‘game joins’, one of them probably got a temp ban.

    5. Hmmm, I suspect there will be items in the game which are highly sort after and can be obtained mid to late game. The first an foremost thing that I’ll be doing is trying to identify why a specific modifier is valuable. Then work out where to get it from, then hopefully being able to build an effective character to farm it. What we don’t know yet is which character will lose less firepower by equipping more magic find %. This is something I really want to know foremost.

      Edit: There will also be area’s with more or less chest like items that have good drops that are easily accessible. This will be worth keeping an eye on once we get started. And don’t forget folks, magic find doesn’t bare diminishing returns like D2 did, so more IS more.

      • [quote]What we don’t know yet is which character will lose less firepower by equipping more magic find %.[/quote]
        Actually we are almost sure. The least item dependant is the WD (in general just 1 type of skill -mana spenders+huge self buff,he only needs decent + int and +dmg/ats/crit). Then its wizard (max dmg output and arcane rest on gear combined with 2 skills = you dont need generators; +MF means youll just swap 1 skill for a generator). Monk and barb – wait and see but similar, monk needs dex instead of str, will be very different if barb is tank spec. Which is not recommended because MF is the average of the party. DH is the most gear dependant, needs to split more stats to be efficient because of her 2 types of resoures which regenate much slower than arcane power.

    6. > Has Blizzard expressed anything with regards to some sort of game-creation-frequency-policy that might result in temp banning?

      I expect that this will be the case to prevent the event of an exploit related to this that they were not anticipating (ie. skeleton king before they moved the waypoint in beta). It could also potentially prevent excessive strain on the server’s during peak period’s.
      I would expect a limit to the number of games you can join within a time period .. ie. more than 2 games every 15 mins, more than 4 games and hour, more than 16 games a day, more than … etc. Just a guess .. no inside info here .. I don’t think they will tell us either 😛

      • Considering the game is ONLINE ONLY there is no way in hell they could get away with limiting the amount of games we can join / create…
        In Diablo 2 however the online part of the game was simply a “bonus”.   This has changed with Diablo 3 now being online only.  If they did implement it I would expect it to be removed asap otherwise Blizz will be facing some pretty crazy QQ, bad PR and perhaps going through the courts.

        • > … there is no way in hell they could get away with limiting the amount of games we can join / create…

          Of course they can, they own the game .. just as they limit you to 10 active characters unless you own another ‘license’ 😉 Also they limit joining creating games/instances in D2 and WOW (and possibly SC2?).

          a blue post somewhere about temp bans on battle.net:

          – Rapidly connecting and disconnecting to Battle.net
          – Creating or joining numerous games within a short time period such as “Magic finding” or “Baal runs”
          – By the use of “add-on programs”
          – Chat channel abuse
          – Failed login attempts
          – In game latency caused by excessive spell effects

          Does it really matter .. sounds like it will be best to hunt down champion packs quickly for the MF bonus and higher chance of rare’s ..

          Then again .. you could be right 🙂

    7. This isn’t WoW, don’t call characters “Toons”. Everytime I read that word I have to stop myself from vomiting.

    8. Very similar to item runs, how about experience/level runs.  What is the best way to quickly level a character?  This is one guide I’m sure many people would be interested in at game launch.  Though obviously this guide/article might best wait until after release so there is more data.  Still, some of same principles of item farming apply.  For experience, what is most efficient: events, bosses, monster packs?  To date I haven’t seen much information or stats regarding experience.  

      For example, generally during experience runs in the beta, I don’t bother picking up anything but rare/unique equipment as it wastes time.  Similar to item farming, if you have a good set of ‘experience bonus’ gear, killing large groups of basic monsters is good, but stragglers aren’t worth it.  I followed the same ‘don’t backtrack’ rule.  I would either go through the crypts, looking for the soul jar event and uniques, or do Skeleton King runs as they seem to provide pretty good exp.  But I don’t know if this was really the best way as I don’t have any numbers?  What is the average experience per minute in the different areas?  What provides the most xp, events, uniques, bosses?  And what provides the best time vs exp gain benefit?  Is it better to do level runs with groups or by yourself, does it matter if it is a multiplayer game?

      Questions abound.  


      • There just isn’t enough information out there yet for efficient leveling unfortunately.

        In D2 a big part of leveling efficiently was ensuring that you were fighting the proper monster levels – if you were a lot higher or lower than what you killed, you’d earn much less than you would have if you were the same level. If you were playing in a /players 1 game you could easily push too quickly, and then be getting 70% of the exp you should against mobs that were difficult.

        but I have no idea what to expect in D3 for leveling efficiency.

        for group play,

        we did a test in beta along the following:
        killed a skeleton solo when we were lvl 9 – he got 45 exp
        brought a level one into the game but left him in town, and had the lvl 9 kill another skeleton – he got 45 exp
        had the level one teleport to the level 9 so they were on the same screen, had the level 9 killed a skeleton – both players got 45 exp
        regardless of whether alone, solo, or actually grouped, it was always 45 exp.

        the monsters took longer to kill with others along, but I’m not sure if it was a straight health multiplier (x2, x3, etc) like in d2. Blizzard seems to be aiming to make the exp rate gains the same no matter how many people you bring along.

        grouping with others can make things faster and more efficient when you synergize, but the cost of failure is higher with stronger mobs. 

    9. I would think that the Resplendent Chest event would be good if you need the XP, otherwise it might be questionable. But in a practical sense, unless you’re farming lowbie equipment, you’ll probably be interested in the XP. I think clearing out the Crypts makes sense as you level.

    10. Please, it? a game, not high school mathematic problem. We played it for fun, 10s is not an issue. If you are not a Chinese in full time job on RMAH, of course .-))

      • Heh, Diablo 2 was the best math class I ever had, it was basically an introductory stats class all in itself 😛 

        Games like this are great fun to explore and play in, the flashy combat explosions, the lore and storytelling, the excitement of new loot. But for theorycrafters, we’re perhaps equally interested in exploring the giant mathematical systems that drive that experience.
        We’ll chase efficiency as often as we’ll try to take silly builds and make them viable (I’m hoping for a armadillo thorns based barbarian). Its the same thrill of discovery that drives the ‘normal’ gamer, just a different angle.

    Comments are closed.