Wall of text incoming. I was a bit surprised that there is not that much information available from the wiki/forums/reddit/etc. regarding drop rates from kills. I’m a GW1 vet and still a bit stubborn in wanting to find ways to make “old school” mob farming possible and competitive in terms of gold/time. Unlike metas, mob farms are always up and you can hit them for 5 minutes or 30 minutes, depending on your playtime that day. In the past I have been able to occupy some very lucrative niches supplying things that are in demand, but not really produced by the “meta” farms. Along the way, a roller-coaster ride of the best open world mob farm ever! that got introduced with PoF, followed by the collapse of that market, followed by new market demand shifts got me interested in actually figuring out how mob drops work, how MF influences things, and how to look for new farm idea without having to go grind out kills to get data. So here I have some results from drop rate and Magic Find testing. As the numbers are based on statistical testing, I do not claim that the drop rates are “true” drop rates, but rather, what I have tried to do is to get an idea of how drops work and generate a useful model that fits the data. I don’t claim that this is how the drop tables are actually programmed (I’m not a programmer), only that the model is my best interpretation of how things seem to work. So, without further ado…

I Swung A Sword…I Swung A Sword Again…Hey, I Killed Something!

When you kill a foe it appears that there is first a roll to determine whether you get anything, and if you do, what general type of item you will get. The main foes used in this testing were Canyon Spiders (lvl 63, N = 2000), Jungle Spiders (lvl 80, N=2001), Dredge (lvl 31, N = 3011), Branded (lvl 38, N=2275). Based on my testing so far, this initial roll has a table that looks something like:

                       Material Table Drop – Approx. 40%                       Gear Table Drop – Approx. 5%                       No Drop – Remainder (Approx. 55%)

(There are some exceptions for a few foes like Marmox, Deer, and very low level foes, which all appear to have higher chances for No Drop.)

Assuming you get a successful materials roll on this initial table, there appears to be a second roll that determines whether you get a crafting material or lodestone or just a salvage item. Similarly for gear, I assume that there is a roll for exotic through white gear. (There is still a decent amount of uncertainty regarding true gear drop rates. It is too early for me to tell whether different foes have different gear drop rates. Further, I have suspicions that gear drops are where DR shows up first, but cannot be 100% certain that it is not simply a bit of bad luck combined with a relatively low drop rate. I have seen no effect of DR on materials table drops in my test farming or in my other usual farming that can be upwards of 600 kills per hour of mobs in a relatively small zone.)

Magic find DOES NOT appear to influence the first roll that determines if you get a drop. That is, whether if you kill a foe with 0% MF or 500% MF, about 40% of the time you will get an item from the Materials Table. I’m going to focus on the Materials Table instead of Gear from here on because the higher overall drop rates are more amenable to testing.

Magic Find Gets You Better Loot, Not More Loot

OK, lets say that you get a successful roll for a Materials Table Drop (YAY!). What can show up in this table? Junk, Salvage Items, Basic Food Crafting Materials (Eggs), Loot Bags (e.g., Light Miner’s Bag, Fine Crafting Materials (meat, t1-t5 blood, claw, etc.), Rare Crafting Materials (t6 blood, claw, etc.), and Exotic Materials (Lodestones). Actual contents vary depending on the foe type, and some foes have relatively unusual tables (such as Drakes not dropping junk). Magic Find DOES influence your chances to get the “better” items in this table (Note that what the game thinks is better is not always what you think is better.) This is part of what helped me figure out what items share a table and how drop chances in that table scale with Magic Find. That is, as your Magic Find goes up and you get more fine crafting materials, something else must drop less frequently. Now that we have a basic model and some hypotheses, we want to know what each of these percentages are and how they scale with MF. So off to the test arena!

Magic Find Test 1 – Canyon Spiders – lvl 63

Spiders made for a great subject to test my hypotheses because they have relatively simple Materials Table drops. If you successfully roll a Materials drop you either get junk (Prickly Spider Leg) or a fine material (Full Venom Sac). I ran tests in 100 kill increments (different characters, different days, doing something else in between) up to a total of 500 kills each at 164%, 339%, 389%, and 509% MF. Adding up junk and venom sac drops and dividing by total kills gave a total materials drop rate of 40.8%. Gear dropped at a rate of 5.8%.

DATA PLOT Here is the plot of the results, shown as the fraction of kills that give fine mats (venom) or junk. Each point is 500 kills.

Also shown is a model of the drop rate scaling with Magic Find. I started with regressions and then manually tweaked the numbers slightly to achieve round numbers. Further, based on the Prior comment about Magic Find, the first model that I tried was simply Drop Chance of Fine Mat. = Base Drop Chance x (1+(Magic Find/100)). Which would give, say, a 1-in-10 chance at 0% Magic Find and 2-in-10 chance at 100%. This vastly overestimated the drop scaling with MF. After further inspection, the best fit round number model was:

Drop Chance of Fine Mat. = Base Drop Chance x (1+(Magic Find/1000))In the case of the spiders, the base chances (0% MF) of getting either junk or venom were both 20%. As you add MF the chance of venom goes up and the chance of junk goes down. Thus, this also is evidence that these items do share a drop table. As you get more venom sacs, those drops “crowd out” junk drops.

So now we can reconstruct a representative Materials Drop Table for these Spiders, expressed as the % chance to get a particular drop per kill:

                                              At 0% MF:     Junk - 20%,  Venom – 20%                                              At 500% MF:   Junk – 10%,  Venom – 30%                                           At 1000% MF: Junk – 0%,  Venom – 40%

Other % can be easily be found using the formula above.

It apparently takes 1000% Magic Find to double your chances of receiving a higher tier item. Not 100% as may be expected and implied by prior quotes on the subject. It also implies that at 1000% you should stop getting junk as drops.

Magic Find Test 2 – Dredge lvl 31

Dredge made for an interesting test case because I wanted to see if Magic Find also improves the chances to get loot bags. The Material Table for the Dredge contains junk (Dull Claw), a salvage item (Clump of Ore), and a loot bag (Light Miner’s Bag). N kills = 3011. Sum of Materials Table drops = 40.5%. Sum of Gear Drops = 5.07%.

DATA PLOT Here are the Dredge results plotted vs. MF. I tested at more MF levels, so I don’t have 500 kills at every level. Each point shown is typically 200 kills, but a couple are 100 and 150 due to a few tests being interrupted by real life.

The model was constructed in a similar manner to the spider model. What is interesting is that the salvage items appear to be lumped in with the junk category and drops do not increase with Magic Find. Increased loot bag drops crowd out salvage items. Further, the loot bag drop rate actually starts off at the same base drop % and has identical scaling with MF as the venom sacs for the spiders. Loot bags start at 20% drop rate with 0% MF and the remainder of the table is split between junk (15%) and salvage (5%). After a bit of analysis, we find the following Materials Drop Table for Dredge:

                                             At 0% MF:        Junk – 15%,   Salvage – 5%    Light Miner’s Bag – 20%                                            At 500% MF:   Junk – 7.5%   Salvage – 2.5%  Light Miner’s Bag – 30%

Other % can be easily be found using the model/formula above.Magic Find increases your chances of getting loot bags (moldy bags, miners bags, etc.) from foes. Salvage items do not increase with MF and the amount received actually decreases with higher MF due to the crowding out effect from getting more loot bags. There are apparent similarities in drop tables for very unrelated foes, which suggests that there may be a standard setup for the Materials Table and typical percentages.

More Spiders – AAAHHH! Level 80 Spiders and t6 Mats

Tier 6 crafting materials (powerful blood, powerful venom, etc.) were also an item of primary interest. Further, I wanted to compare the lvl 80 spiders to the lvl 63 spiders to see if there were similarities that would support the idea of “near universal” scaling or numbers in drop tables. I killed 2000 jungle spiders in Draconis Mons, 500 each at 339, 389, 409, and 509. My account MF is maxed and the prior 164 MF trials were done on a separate account that does not have this LS map unlocked, so I was unable to do the low MF point. Overall, 40.1% of drops gave an item from the Materials Table and 3.5% gave gear.

DATA PLOT Results of Materials Table drops for the level 80 spiders are plotted HERE

As it turns out, the Materials Table for the lvl 80 spiders is similar to that for the lower level spiders shown above. Lo and behold, the model best-fit chance of getting junk at 0% MF is once again 20%. (Detailed method note: these spiders can give either 1 or 2 prickly legs as a junk drop. I did not differentiate and counted only the number of times junk dropped, not the quantity. So if I ended up with 10 legs, I then went back in the chat window saw that I got, for example, 4 drops of 2 prickly legs and 2 drops of 1 leg. The number of junk drops in this case is 6. I have not tried (nor do I care) to figure out drop rates of 1 junk vs. 2 junk.)Remaining drops are divided between t5 venom and t6 venom. The sum of these drop chances also fit 20% at 0% MF. Apparent base t5 drop chance was 16% and t6 was 4%. It is harder to pin down the actual t6 drop rate as a % of kills because they are rarer and there are two RNG steps involved in actual testing (first whether you get the material drop and second what that drop is). You can remove the first of these RNG steps by looking instead at the % of Materials Table Drops. The results are replotted in the right graph HERE and as you can see the 16%/4% t5 to t6 split appears to fit rather well. So we get our table:

                        At 0% MF:               Junk - 20%       t5 Venom - 16%     t6 venom 4%                        At 500% MF:          Junk - 10%        t5 Venom - 24%    t6 venom  6%

DATA PLOT Results of the lvl 80 spiders (with t5 + t6 drops comibined) compared to the lower level spiders are shown Here to illustrate the common scaling behaviours.

Using the Model

The model and scaling equations are also useful to predict mobs that might be worth farming or estimate how often you might get certain materials. I’m testing Branded to look for lodestone drop rates and see whether they take up space in the drop table in the same way as t6 mats, by replacing fines. My prediction was that I should get drop rates of 15% junk, 5% salvage, 20-X % dust, and X% lodestones. The junk through dust drops fit the model so far.With this model, you can make a first guess at what the drop rates should be for a wide variety of mob types. Take Bats. Bats drop junk, salvage (leather), and bloods. You might expect drop rates at 0% MF: Junk – 15%, Salvage 5%, Blood – 20% or for lvl 80 bats – t5 Blood 16%/t6 blood 4%.You can also estimate more complex tables. Take Devourers. They can drop junk, scales, or venoms. You could form a hypothesis that the table at 0%MF is: 20% Junk, 10% scales, 10% venoms. Then you do testing and adjust the % according to the model. You might expect that you should get roughly equal numbers of scales and venom. My initial limited results back this up.You can also use it to decide what mobs are not worth farming. Cooking materials (eggs, meat) appear to share some of the non-junk drop chance in the Materials Table. In other words, if you want bloods, Skales are probably not the best farming choice because they drop eggs too. Bats would be a better pick in terms of drop rate of bloods.

Other Interesting Factoids

Black lion chests appear to be on their own roll when you kill a foe. They can drop either alone or together with another material/gear drop. Drop rates appear to be in the vicinity of 0.5%. Similarly, a number map-specific event items such as the Dwarven Key in Dredgehaunt or the Gear Crank in Lornar’s also appear to be on independent tables and can either drop alone or with something else, suggesting that they are on a separate roll, independent of whether you get a material or gear drop.

Drakes do not drop junk.

Happy Farming!

tl;dr Magic find scaling may not be what you expect it to be. It appears to take 1000% MF to double drop chances, at least for materials that drop from foes.Different foe types share similar loot table % and scalings that can be useful for choosing the best farming spots.

and since I mentioned it above...

Best Farm Ever!

Path of Fire introduced perhaps the best mob farm in the history of the game but sadly, like many market-linked things, the newest two living story maps destroyed it. It was based on Evergreen Lodestones. Prices varied after PoF launch and in November, but the farm was usually a solid 40+gold/hr and at times broke 70 gold/hour if you were willing to do about 2 minutes of crafting after your farm. Unfortunately, Iboga farming in Istan coupled with the Olmakhan bags in the BL chests and a lack of new raid wings means that Evergreen Lodestones have dropped significantly in price. So I’ll use it now as an example. I have seen a few other folks farming this spot, but 99.9% of the time I had it all to myself aside from some poor sap who got knocked off of his mount by a stray harpy. In Desert Highlands, just a bit southwest of Springer Ranch, there is a flat open area with two small plateaus in the middle. It is just south/southwest of the S in Stampede Uplands on the map. There are groups of Iboga Sprouts with some regular Fanged Iboga and a couple of Veteran Ibogas. Overall, there are about 40-some enemies to kill if you run a complete loop and kill all. A loop takes 2.5-3 minutes with any decent killing build and by the time you make it back to where you started, the first mobs that you killed have respawned. You get 800-900 kills per hour. This was a very good kill rate, but what really put this over the top was the abnormally high Evergreen Lodestone drop rate. The Iboga were dropping 5-10x more lodestones than expected. It is tough to get firm numbers on the usual lodestone drop rates due to rarity and the need to run thousands of kills to generate stats, but suffice to say that my rough estimate for the base drop rate for regular lodestones from other mobs (Sparks, earth elementals, etc.) is probably about 0.5 to 1%. Killing these iboga with 500% MF was bringing in 35-45 lodestones per hour, or nearly 5% rate. An example 30 minute farm session yielded: 20 Evergreen stones, 2 t6 Venom, 5 t6 dust, 1 t5 venom, 7 t5 dust, 78 barbed thorns, and a bunch of gear, both regular and unid, plus 50.6 silver in raw coin drops. The lodestones themselves could be sold directly for 1-1.5 gold, or could be crafted into Renegade runes or Sigils of Transference to get a sales price equivalent to 2 gold per lodestone. New raid ensured high demand. Unfortunately, prices have plummeted. It is still ok, but not as outstanding as it once was.This farm is actually what got me quite interested in drop rates because I noticed that as I got more lodestones I got surprisingly few t6 mats.TL;DR Mob farming can produce competitive, or even superior gold income vs. other non-TP methods if you are able to find a good spot and market niche.

Impressive research! And very interesting results. Thanks for your work!

Hats off for the dedication. This looks like a lot of work and very interesting

at least magic find helps for something but not much. Magic Find stat is a lie for things like reward chest and bags. also selling items for gems to boost magic find is pretty much strait up theft. i ran some statistics and either magic find is bugged after a set point or it actually has no effect on anything. i opened 2,000 trick or treat bags with 200% magic find. at bag number 1,243 i got a rare item that was worth a damn, a half ass infusion. i then opened 7,342 trick or treat bags with 572% magic find. not one rare or higher was received. so yeah how messed up is that? statistics show the reward pool gives better and more rare items at a lower magic find level than a higher one by a factor of about 239% give or take...... thanks anet

@"Razakin.5386" said:at least magic find helps for something but not much. Magic Find stat is a lie for things like reward chest and bags. also selling items for gems to boost magic find is pretty much strait up theft. i ran some statistics and either magic find is bugged after a set point or it actually has no effect on anything. i opened 2,000 trick or treat bags with 200% magic find. at bag number 1,243 i got a rare item that was worth a kitten, a half kitten infusion. i then opened 7,342 trick or treat bags with 572% magic find. not one rare or higher was received. so yeah how messed up is that? statistics show the reward pool gives better and more rare items at a lower magic find level than a higher one by a factor of about 239% give or take...... thanks anet

Opening Trick or Treat bags is not affected by magic find. They never have been. Your “statistics” are purely coincidental.

“Magic Find ~ increases the chance to receive higher-quality loot from slain foes. With a few exceptions, it does not affect containers (including champion loot bags), chests, or any other source of loot.These special items are affected by magic find:Containers received from reward tracks, generally called "Loot box" or "cache", which simulate loot drops from the named NPC.Chests and containers specifically mentioned to be affected by it, such as:Lost Bandit Chests in the Silverwastes.Caches and loot boxes from Heart of Thorns content.Piece of Unidentified Gear in Path of Fire.Divine Lucky Envelopes during Lunar New Year.

@"Razakin.5386" said:Magic Find stat is a lie for things like reward chest and bags.It was never meant to affect things like reward chests or bags. It only affects loot that drops directly from foes (or containers that simulate such loot, such as PvP reward track containers & unID gear). It also affects a small number of items that are marked as exceptions.

also selling items for gems to boost magic find is pretty much strait up theft.It's not in the least; it does exactly what it's advertised to do: increase the chance that you'll get better loot from foe drops.

For more, read:

• Magic Find article on the wiki
• Testifye's Research on 42000 pieces of unidentified gear at various different levels of magic find

Then i now know why everyone just sells the bags from all events instead of opening them up. even when killing mobs it does little to nothing actually worth noting. like said you need 1000% magic find to effectively double your chances. seriously? so yeah selling stuff for magic find is very much so pretty dang messed up. its meant for you to lose or spend money on what you need. either that or more time the game which has a change of you spending money. time is money. so yeah. the statistics of what you get kinda says everything. the amount of luck needed to get to the point where you could max boost yourself to 1000% magic find is in like 4% of gw2 players and even that might be an exaggeration

Which items are sold for Gems that boost Magic Find?

@"Razakin.5386" said:Then i now know why everyone just sells the bags from all events instead of opening them up.

Clearly, some people are buying the bags, otherwise you couldn't sell them for love or money. Those that sell are more interested in coin; those that buy are more interested in playing the GW2 lotto: hoping for the lucky drop. (Some people gamble because they don't understand the math; some just because it's kind of fun.)

even when killing mobs it does little to nothing actually worth noting.No, that is incorrect. It does quite a lot. Each bit of magic find means you get fewer high-chance drops and more low-chance drops than if you didn't have any MF.

like said you need 1000% magic find to effectively double your chances. seriously?From where are you drawing that conclusion? You can't "double your chances" of anything because you're talking about different loot tables. With blue unID gear, you can eliminate any chance of getting blue drops with ~850% MF. That doesn't double your chances of getting rares per se; it means that the loot you do get is more likely to be rare, because it can only be green|rare, not blue|green|rare. Every bit of MF helps.

so yeah selling stuff for magic find is very much so pretty dang messed up. its meant for you to lose or spend money on what you need.I'm not quite sure what you mean by this. It's fairly trivial to increase your MF: salvage all the blues & greens you get (which is worth doing anyhow, because the value of the salvage is nearly always worth more than the gear and it's much, much faster to process). You can also salvage ecto with mystic|master|silver-fed kits; the amount of crystalline dust you get is worth nearly as much as the ecto and you end up with a lot of luck.

the amount of luck needed to get to the point where you could max boost yourself to 1000% magic find is in like 4% of gw2 players and even that might be an exaggerationEvery 100% you boost your MF makes a notable difference in the value of the loot you receive over the long run. It's not going to mean you get fancy stuff; it means that all the detritus you get will be worth more. How much more depends a lot on how you play, how often, where, and whether you focus on loot or not. If you spend all your time doing stories and the occasional labyrinth, it won't help much. If you spend all your time in the lab and the occasional story, it makes a huge difference.

You can run your own test. Join a "no doors" lab farm and don't boost your MF. Count your bags before & after. Then use every boost you can manage, again counting before and after. Be careful to spend the same amount of time and the same commander (so the route is about the same). 30 minutes is probably ok; 60 is better.

@"Inculpatus cedo.9234" said:Which items are sold for Gems that boost Magic Find?

I suspect they are referring to the Heroic & Item Boosters.

Ahh. I didn't even know those were still available in the Gem Store. Lol. I get so many Boosters just from playing, I never look in the Gem Store for Boosters.

@Illconceived Was Na.9781 said:

@"Inculpatus cedo.9234" said:Which items are sold for Gems that boost Magic Find?

They’re not Gemstore exclusive however. I’ve got a ton of boosters in my bank and didn’t pay gems/statuette for a single one. /shrug. I use them methodically.

@Haleydawn.3764 said:

@Illconceived Was Na.9781 said:

@"Inculpatus cedo.9234" said:Which items are sold for Gems that boost Magic Find?

Sure. But newer players aren't likely to have access to a lot of birthday presents, BL keys that drop in game, or other sources of the boosters. They'd see the scrolling advert in the BL Shop's interface, one that sometimes suggests spending on boosters. If one of those players believed (mistakenly) that MF affected results from bags, you can imagine they'd be upset enough to post about it.

That's why it's suggested to consult the Wiki whenever new players have questions of the sort. :)

my bad on my part that i thought magic find affected that? its not clear to regular players as to what it affects fully. even so the patterns when using for example the trick or treat bags in itself is messed up. one tiny pool it gives and some to make it seem like "rare" drops and then im guessing the system runs it at .78% of super rare and rare items outside of that specified normalized pool. thats what it seemed when all 9342 trick or treat bags. so yeah

@"Razakin.5386" said:my bad on my part that i thought magic find affected that? its not clear to regular players as to what it affects fully. even so the patterns when using for example the trick or treat bags in itself is messed up. one tiny pool it gives and some to make it seem like "rare" drops and then im guessing the system runs it at .78% of super rare and rare items outside of that specified normalized pool. thats what it seemed when all 9342 trick or treat bags. so yeah

I'm not clear what you're trying to say here. The drop rates for each tier in the ToT preview gets geometrically lower. The "super rare" category seems to drop items about 12 times in every 100 000 ToTs. There's about a 11% chance of getting one if you open 1000 bags (although of course, one of them might be a relatively worthless Gift).