Results of 500 Sand Hawk farms (Updated to 1000)


(Gamechanger97) #1

I would like to begin by stating that I do not use community patch or any other mods and that this farm is the result of pure vanilla read only farming.

A little bit of a preface, I was about to do some op8 sandhawk farming to try to get some flying elemental ones and was in the chat of a twitch stream where I mentioned that it seems sandhawks without an element and prefix are more likely than those with. Someone in the chat told me that no element was a 75% chance and that slag, corrosive, fire and shock split the other 25% equally. This person also claimed that all attachments, including not having any attachment, were all equally likely. So I decided to make a record of all of my sandhawk drops and calculate some stats based on the results. This is some of the more interesting stats at my current stopping point of 500 runs.

Number and percentage of each element:
No element: 344 68.8%
Slag: 34 6.8%
Fire: 42 8.4%
Corrosive: 39 7.8%
Shock: 38 7.6%
Of course, after only 500 runs there is some margin of error but I update the percentages every 10 runs and the no element hasn’t peaked above 71% since very early on. One interesting thing is that slag was actually my most common elemental drop up until somewhere in the 350-380 runs range when it was passed by fire. I also had a stretch of around 80 runs without a single shock and other times I will get 2 back to back. It is hard to say if after enough runs the percentages will line up with 75%, 6.25%, 6.25%, 6.25%, 6.25% but I do believe the actual chance of no element is around 70%.

Next up is the prefix. I would like to start off by explaining a few things about the prefixes. There are 7 options when the game picks a prefix. 6 of which correspond to an attachment and the other is for having no attachment. If the sandhawk has one of the 6 possible attachments its prefix will be either: Bladed, Stoic, Skirmish, Deft, Stopping, or Flying. If the sandhawk has an element and no attachment, its prefix will be as follows: Slag = Amp, Fire = Incendiary, Corrosive = Corrosive, Shock = Sapping. If the sandhawk has no element and no attachment, its prefix will depend on the manufacturer of the grip as follows: Dahl = Smooth, Hyperion = Light, Maliwan = Captain Scarrlet’s, Tediore = React, Bandit = Reserve.

The result: 396 or 79.2% had no attachment. Just the default prefixes on nearly 80% of the 500 sandhawks.

So as I mentioned above there are 5 grips. There are also 5 stocks. If we multiply 5 stocks * 5 grips * 5 element options * 7 attachment options. We get a total of 875 different sandhawks (note this isn’t counting sights). During my 500 runs I found 187 different combinations.

Lastly, I sorted the sandhawks into 3 categories: Common- contains any sandhawk that has no element and no attachment as this is the most probable option. Uncommon- these sandhawks had either an element or an attachment but not both. Rare- both an element and a prefix.
I didn’t keep track for each of these the number out of 500 that dropped but instead how uniques in each category were found. All 25 common sandhawks were found. 128 of the 250 different uncommon sandhawks spawned. And only 34 of the 600 possible rare sandhawks were found in my 187 unique drops. I did keep track of how many common sandhawks spawned total since it was easy enough to sum up 25 of them: 275 of my 500 sandhawks were no element no prefix, that’s 55%! Finally: based on my thoughts that you have a 30% chance of an element and a 20% chance of a prefix: You have a 6% chance of getting an element and a prefix. Then if you want say a perfect shock sandhawk with dahl grip and dahl stock you have to take that .06* the 1 in 4 chance of shock * the 1 in 5 chance of dahl stock * the 1 in 5 chance of dahl grip* the 1 in 6 chance of flying and you get .0001. That is a 1 in 10,000 chance for a perfect sandhawk.

Thoughts and or comments?


(I’m Nat and just a Number.) #2

500 :scream::scream:

How long did it take you?


(Focused for Success in a Distracted World) #3

This looks relatively consistent with some other tests, particularly the quickie Fibber test done here.
Oddly @Dr_Do-Little’s Veruc/Lyuda test was not consistent with that.

I don’t think we’re at a point of being able to infer the exact percentage. Knowing GBX the NE chance is probably some bizarre and seemingly arbitrary number like 71.4% - but 500 is a pretty good sample.

Well done! Having to chart multiple parts per drop is a big undertaking.


Edit : small suggestion - a bit of formatting would be helpful. Even something as simple as turning this :

into this :

Number and percentage of each element:

  • No element (344) : 68.8%
  • Slag (34) : 6.8%
  • Fire : (42) : 8.4%
  • Corrosive (39) : 7.8%
  • Shock (38) : 7.6%

or even this :

Number and percentage of each element:

Element Number Percentage
None 344 68.8
Slag 34 6.8
Fire 42 8.4
Corrosive 39 7.8
Shock 38 7.6

Here’s how to set up a simple table.


(The doc is in.) #4

I think the major difference with the Lyuda/Veruc is that they are not reward items. Beside the number of runs.
I wasn’t home for the last week and was very busy prior to that but I should have more free time over the next 2 weeks.

Good work.


(Gamechanger97) #5

Thanks to everyone for the advice and thoughts. I plan on doing many more of these with many other farms I do. @Jefe Thanks for showing off the formatting and providing links, that will definitely be useful in the future! @nat_zero_six about a week.


(Is this thing on?) #6

I also worry with drop items that some may be missed if they glitch through the floor or slide into a wall. I had issues with that working up the stats for the Assassins and Henry the Stalker - definitely going to skew the stats if some random number of actual drops go missing and can’t be counted!

Hopefully sometime tomorrow I’ll be able to grab those Sandhawk stats and calculate the confidence intervals.


(Gamechanger97) #7

I would love to see how that turns out!


(The doc is in.) #8

I’m a bit worried with “disappearing loot” too but it’s not much an issue with the Veruc/Lyuda. Unless you pin Mobley by the church. Which i don’t do. I did had Death Trap chase Gettle underground once thou. So i have no way to know if here was a Lyuda drop on that occasion Not spawning him anymore.

I’m a lot more concerned about forgetting to log in a run. Especially in long empty streaks.

Overall I think players experiment like that, with the lost loot, provide an “effective” drop rate. Which is more useful than the theoric/real rate. Farming the Assassins you will lose some eventually…


#9

so i did my own sandhawk farms
in total i did 1000 runs all on op 8.
i noted the element of the weapon,
if it had the flying prefix,
if it had a dahl(or bandit which is ok too) grip
and if it had a dahl stock.
my results were:
fire: 79
shock: 87
corrosion 59
slag 70

flying 34

dahl 204
bandit 205

stock 198
sadly no perfect but i got a flying fire one with dahl grip but no stock :frowning:


(Focused for Success in a Distracted World) #10

Well based on your numbers, I’m out by 0.9% :laughing:


#11


32 tries later i get this thing not quite perfect but damn


#12

next 500 tries:

slag: 29
fire: 45
corrosion: 33
shock:39

flying: 15
didn`t really bother anymore to check for grips or stocks


#13

A little more data to add to this thread. No where near as detailed as the great work by gamechanger97. I expected a long farm when the first 30 came out non-elemental and i wanted a Flying or Stopping Fire or Shock. I accidently closed out the mission when a shock stopping showed up that I could not resist trying to save. :stuck_out_tongue_winking_eye: I never saw the BL2 auto-save flash on the screen so I suspect the XB1 cloud save may have gotten me, if it does auto-saves.

Sandhawk Offers
Runs Non-Elemental Corrossive Fire Slag Shock
1-25 25 0 0 0 0
26-50 17 3 3 2 0
51-75 20 0 2 2 1
76-100 21 2 1 0 1
101-125 17 2 2 3 1
126-145 11 1 3 1 4
Totals 111 8 11 8 7
Percentages 76.55% 5.52% 7.59% 5.52% 4.83%
Flying 1 1 0 0 0 1.38%
Stopping 3 0 0 0 1 2.76%

(Is this thing on?) #14

Cloud save just backs up local saves. It’s possible you triggered it without noticing - you’re awful close to the FT when you turn that quest in, so it wouldn’t take much. I’m not 100% sure but, since this is a main mission turn in, I think accepting the item triggers an auto-save when the script advances to the next mission in sequence. (It’s different with side quest rewards, since there’s nothing to actually advance to when you turn those in.)


#15

I saw a video or read a thread about someone successfully handing off a sandhawk to a mule if they are very careful to not move around and trigger the BL2 auto-save. I was very careful and quick. In the future I think i’ll do these offline as an additional precaution.


(jelmerjt) #16

Getting a perfect Sand Hawk or Fibber is like winning a pretty decent prize in the lottery. You’re looking at a 1 in XXXXX chance if you want a specific version. They’re blue guns and so they tend to not spawn with prefixes, elements and sights.


(B Pdbou Xoq7x Al5 Wh) #17

i am farming this right now and simply looking for a flying shock. it is getting tedious. i have a fying amp, while usefull, does not meet my specific needs for shield damage. i can only hope…


(Gamechanger97) #18

I am back finally with the results updated to 1000 farms! Still no perfect sandhawks yet but I did manage to get double dahl flying with no element and a dahl grip shock flying one. Those are the best I’ve gotten so far. First of all, some basic stats that don’t fit any of the tables too well and some thoughts on those:

  • 283 or 32.34% of the 875 unique sandhawks

  • 777 or 77.7% had no attachment

  • 537 or 53.7% were sandhawks with no attachment or element

So as far as the first stat goes, at 500 sandhawks I was at 187 uniques. This means I found 96 unique sandhawks this 500 runs which is only about 51.3% as many as the first 500. I would expect the drop off to be even more extreme in the next 500 due to the fact that out of the 592 sandhawks left to get, only 61 don’t have both an element and an attachment.

The no attachment category has dropped by 1.5%. Based on the fact that I had a bizarre amount of attachments in the last 50 runs or so, I would guess the true percentage is a little higher. However, based on the last couple hundred runs I would say the true percentage is somewhere between 78% and 79%.

As for the percent of total sandhawks made up of just the 25 basic sandhawks, this has also dropped from 55% to 53.7%. This isn’t a super crazy change and I think it is pretty accurate.

Ok now for some charts! First off, elements:

Element Number Percent of Total
No Element 687 68.7%
Slag 74 7.4%
Fire 71 7.1%
Corrosive 81 8.1%
Shock 84 8.4%

The no element category has been hovering right around 69% which wouldn’t surprise me. The other elements seem to come in waves where the element that is the lowest percentage at one point is the highest percentage a few hundred later. Compared to the results at 500, these percentages are less spread out and I would bet they are equal odds over a large enough number of attempts.

Let’s look at the number of times and odds of each attachment appearing based on the total number of times a sandhawk had an attachment which is 224:

Prefix Number Percent of Total
Flying 39 17.4%
Stopping 25 11.2%
Deft 32 14.3%
Skirmish 41 18.3%
Stoic 41 18.3%
Bladed 46 20.5%

I find it interesting how much higher the odds are for bladed over all other attachments. However, with this small of a sample size, it is hard to say if this is just rng or a legitimate trend.

Next up the number of times each grip was found regardless of whether or not an attachment or element was on a given sandhawk:

Grip Number Percent of Total
Dahl 183 18.3%
Hyperion 198 19.8%
Maliwan 213 21.3%
Tediore 198 19.8%
Bandit 208 20.8%

These percentages seem pretty evenly spread out, but it is a trend I’d like to look into that dahl happens to be the lowest current percent.

Lastly, I would like to just throw out this interesting difference with the most common 25 sandhawks. Of the sandhawks with no element or attachment, the least common one is hyperion grip tediore stock which occurs only 12 times. The next lowest for comparison occurs 16 times. Meanwhile the most common is a two-way tie between bandit grip bandit stock and maliwan grip dahl stock each with 28. These two sandhawks make up over 5% of the total sandhawks on their own! These aren’t super far ahead as there are multiple with 27 each. I do find it weird that one of the 25 common sandhawks would spawn only 12 times while others spawn 28 times. That is a large difference even with a smaller sample.

I leave you with my two nicest sandhawks so far:


Elemental Bitch
(The doc is in.) #19

I wish I had recorded as much details with the Veruc/Lyuda farm.
Those matching grips… Seem close enough thou but…
Awesome job. :+1: