So I tracked 2,000 drops
February 7, 2017 7:58AM
edited February 2017
The purpose was to get an idea of their model and figure out what was the best way to obtain the infinite amount of gear required. It wasn't really to find out what their exact "drop rate" is.
If you're going to post, I simmed 20 twice and got 1 so your analysis is wrong...just...don't...
I was looking to answer questions like
-What's the cost model of obtaining gear different ways
-Is it better to sim 1 or 2 at a time vs 20?
-Are drops rates variable, I simmed 40 and got 4, is EA messing with me?
First of all, simming 1-2 at a time did not produce any statistical difference than 20 at a time. I suspected this anyway just from how programming works and the fact that the results of 20 produce numbers like 11 and 7 which are prime. What it DOES do is prevent you from seeing both really good and really bad results with an emphasis on the later. People have 1 out of 20 Stun Cuff sims seared into their head forever.
Is the sample size "large enough". I'm going with good enough for an event with a discrete, binary outcome.
"I've had really long stretches of really bad results". I saw really long stretches of both really good and really bad results. As the sample size gets larger, everything regressed to the mean. Common sense dictates they are not going to individually set your drop rate at 10% one day and 40% the next day.
Cantina toons
Farmed Chewie, TFP, Chirpa, B2, Lando, Kylo Ren
162 shards out of 544 sims = 29.78%
In fact, nearly every individual toon had their drop rate around 30% so it wasn't like TFP was 10% and Lando was 50%.
I'm convinced every cantina node/toon is set to 30% and doesn't change.
Gear
1389 sims of Stun Guns, Cuffs, Holo Projectors, Droid Callers, Detonators and Carbanti and few others but those were 90% of the samples. I started to see a pattern evolve. To no surprise, the drop rate depended on the min energy needed to farm it but to my surprise, each of the gear at each energy cost level were identical.
10 energy = 20%
8 energy = 28%
6 energy = 36% (Ok, it was actually 35 but I'm going with 36 is more likely their model)
From this point, let's call them just 10, 8 and 6. Now to the real reason for doing this, the cost basis with 120 energy @ 50 crystals.
Node cost | EV of 120 energy | drop rate | Crystal cost per salvage | Cost for completed piece
10 | 2.4 | 0.2 | 2.4 | 20.83 | 1,041,6
8 | 4.2 | 0.28 | 4.2 | 11.9 | 238
6 | 7.2 | 0.36 | 7.2 | 6.94 | NA
The PvE store, Cuffs are 1400 and Holo Projectors are 300 which place them right in between the cost of 50 crystal refreshes and 100. It also makes sense since you would expect the store to be slightly more than the cost of a cheap refresh.
Using Cuffs and MK3 Holo Projectors as the baseline, (0.8 and 0.74), there are some clear discrepancies in both the PvE and Ship Shipments stores. Furnaces are insanely cheap as everyone knows but Stun Guns are a better buy than Cuffs and Carbanti are insanely overpriced. In the Ship store, Cuffs are the best buy while Holo projectors are insanely overpriced. The 650 gear in the ship store is insanely overpriced as well while in the PvE store there are several G10/11 gear bargains.
I'm working on a gear chart and the best way to obtain them: Farm, PvE or Ship store.
Edit: Column headings, thanks Maegor
If you're going to post, I simmed 20 twice and got 1 so your analysis is wrong...just...don't...
I was looking to answer questions like
-What's the cost model of obtaining gear different ways
-Is it better to sim 1 or 2 at a time vs 20?
-Are drops rates variable, I simmed 40 and got 4, is EA messing with me?
First of all, simming 1-2 at a time did not produce any statistical difference than 20 at a time. I suspected this anyway just from how programming works and the fact that the results of 20 produce numbers like 11 and 7 which are prime. What it DOES do is prevent you from seeing both really good and really bad results with an emphasis on the later. People have 1 out of 20 Stun Cuff sims seared into their head forever.
Is the sample size "large enough". I'm going with good enough for an event with a discrete, binary outcome.
"I've had really long stretches of really bad results". I saw really long stretches of both really good and really bad results. As the sample size gets larger, everything regressed to the mean. Common sense dictates they are not going to individually set your drop rate at 10% one day and 40% the next day.
Cantina toons
Farmed Chewie, TFP, Chirpa, B2, Lando, Kylo Ren
162 shards out of 544 sims = 29.78%
In fact, nearly every individual toon had their drop rate around 30% so it wasn't like TFP was 10% and Lando was 50%.
I'm convinced every cantina node/toon is set to 30% and doesn't change.
Gear
1389 sims of Stun Guns, Cuffs, Holo Projectors, Droid Callers, Detonators and Carbanti and few others but those were 90% of the samples. I started to see a pattern evolve. To no surprise, the drop rate depended on the min energy needed to farm it but to my surprise, each of the gear at each energy cost level were identical.
10 energy = 20%
8 energy = 28%
6 energy = 36% (Ok, it was actually 35 but I'm going with 36 is more likely their model)
From this point, let's call them just 10, 8 and 6. Now to the real reason for doing this, the cost basis with 120 energy @ 50 crystals.
Node cost | EV of 120 energy | drop rate | Crystal cost per salvage | Cost for completed piece
10 | 2.4 | 0.2 | 2.4 | 20.83 | 1,041,6
8 | 4.2 | 0.28 | 4.2 | 11.9 | 238
6 | 7.2 | 0.36 | 7.2 | 6.94 | NA
The PvE store, Cuffs are 1400 and Holo Projectors are 300 which place them right in between the cost of 50 crystal refreshes and 100. It also makes sense since you would expect the store to be slightly more than the cost of a cheap refresh.
Using Cuffs and MK3 Holo Projectors as the baseline, (0.8 and 0.74), there are some clear discrepancies in both the PvE and Ship Shipments stores. Furnaces are insanely cheap as everyone knows but Stun Guns are a better buy than Cuffs and Carbanti are insanely overpriced. In the Ship store, Cuffs are the best buy while Holo projectors are insanely overpriced. The 650 gear in the ship store is insanely overpriced as well while in the PvE store there are several G10/11 gear bargains.
I'm working on a gear chart and the best way to obtain them: Farm, PvE or Ship store.
Edit: Column headings, thanks Maegor
Post edited by sneeekypants on
1
This discussion has been closed.
Replies
Shhhhhh. Fully crafted raid gear is the only gear worth spending crystals on in the store. Don't get that nerfed
Second, I want to make sure I'm understanding this correctly:
The drop rate of a certain piece depends on the MINIMAL energy needed to farm it, but not the actual energy of the node ?
So if a piece can be found both in 8 energy node and 10 energy node, the drop rate will be the same for both node (and will be 28%) ? No use to sim the 10 energy node except to go easy on the sim tickets ? What about the Hard Nodes ? Is there's a drop rate difference ? A 8 energy node compared to a 16 energy Hard Mode node, for instance ?
It's not rocket surgery.
I already have farmed for two weeks a set of 50 stun cuffs, that doesn't invalidate the stats shown here.
It is just that it proves that on long run it is more interesting to farm your gear than buying it with crystals.
However, as a statistician, I would like to underline some of my doubts: in an interview, devs asserted that the drop rates are dependent on each items, so cuffs could get a different drop rate than guns.
Moreover, there are nodes that have combined gear (LS-9F, DS 9B and 9F) making these good for farming and decreasing their theoretical crystal cost.
Additionally, I am pretty sure that the new G11 drop less than cuffs and guns. That is why they are overpriced in fleet shop.
Finally, how do you embed the 750 crystal cost of the new G11 gear? I find them really cheap in the shipments and really overpriced (energy cost and fleet currency cost) anywhere else.
But great work otherwise!
Devs have confirmed a few times there is no difference in the drop rate between nodes that have different energy costs for the same gear.
I'm actually tracking the drops on a couple nodes that have a couple pieces that I need however given the above and logically, the rates shouldn't change. Every node has multiple pieces that can drop regardless of whether they are needed or junk. That junk does have drop rates associated with it that most likely follows the same model and for complexity reasons, I would imagine they don't increase the drop rate for unneeded gear even though it seems like it.
Also if you look at nodes like some 9x hard nodes which don't have any shard drops, they drop double 10 energy gear.
The great thing...is when they add a shard to those which makes it an exceptional deal like the infamous 9B-Light hard, so there are 3 useful drops off the one node. "not useful" as in I have 1,000 of them so I don't care what the drop rate is or how cheap it is in the store I will never farm or buy them for the next 100 years.
Edit: typo
Thank you for your feedback, I addressed your points in the OP.
Thank you for your feedback, I'll address your points in order. I'm not a stat guy but worked with a lot of numbers as a c coder doing nav systems then market data/quant programming. I'm a pencil pusher now though.
-I'd like a link to where they say they are different, on Reddit dev's acknowledged that the rates are set individually "by hand" so there could be discrepancies due to human error. What that means is there is no global function or variable which sets them, which is a poor way to do it but maybe there was a reason at the time of implementation. It was ambiguous on whether it meant per item or per node. I think with both this statement and the one you provided, neither wording implies that in fact cuffs are different than guns.
-Technically all nodes have multiple gear drops and all those gear have drops which I would assume fit the same model. Since the drop rates are "hard coded", it's highly unlikely a quant tweaks the numbers either programmatically dynamically or through say a config file on their server. Farming is the cheapest route for all gear, if you get multiple gear you need from the same node, it just makes it more of a value than it was before. You could argue furnaces but you can't actually farm the furnace salvage from any node. I farm some of those nodes you mentioned specifically because some have multiple pieces that are often needed. 9B-Light is another rock star node.
-I'm tracking the G11 gear but the sample size is way too small to deduce anything useful from it. The last statement is why it wouldn't make any sense set the drop way way below 20%, they are so cheap in the PvE store compared to cuffs and guns already, they are almost half price. Unless the drop rate is 30%+ there would be zero point in farming any of them unless you are farming a multiple drop node.
Being a little bit of a numbers geek myself, I did a bit of the same, and my model so far (with a smaller sample size than yours, in the hundred(s)) is:
For instance, simming the node 7D (10 energy) for 250 times gave me 66 stun cuffs (~25%), 90 of the blue thing (36%) and 113 of the green one (~50%)
I've also done a monstrous spreadsheet with what gear can be obtained where, what are the bargains and the rip-offs (per store). (LINK)
Feel free to use / copy it (and redistribute it, just add my name somewhere as co-author).
I've also gathered the actual demand for gear (it doesn't matter to me if Carbanti Mk4s are overpriced, since no one needs them), but haven't found an elegant way to combine both worlds. Well in the end, it all boils down to one simple rule: You need 5000 stun cuff pieces, so get 'em any way you can.
Awesome sheet, I'll take a closer look at it when I'm not dead tired from work. I probably got a lot of typos in my above posts but I don't care.
Something to keep in mind, Carbanti are 50 piece purp gear but fit the profile of 20 piece holo projectors. Same drop rate and same energy cost. From the game design point of view, they are concerned about the cost to obtain.
It is most appreciated.
this is a game and i'm sure my opinion is about to offend someone but i am entitled to my opinion. but some of you take this game way, way to seriously. not to mention you need to do more productive things with your time than tracking gear drops.
a bigger opinion in my head is the fact that people can play a game and when then don't like it instead of not playing it anymore they complain to the creators to make it better. it's not like your being forced to play the game. if you don't like it don't play.
sorry about the rant. i didn't get my morning coffee and i'm a bit cranky.
Will you stop being the forum troll? Please?
First of all, great work @sneeekypants.
Second, in your header for the columns you have drop and EV mixed up
So instead of ranting at someone who actually did something to help out the community, why don't you go and **** in someone else's cereal. Like one of those hundreds "am I ready for the palp event" or "nerf wiggs" threads?
@OP: great work! Thanks! Just ignore @spotremix He's being a ****.
I agree with these drop rate, and would add motbobly gold gear but all new gear found in level 9.
Let it all out son. Feel better meatbag?