Plentiful Harvest Passive does not equal 50%... TESTED

BergisMacBride

Since I did my run of over 900 nodes (posted above), I quit tabulating but have been keeping an eye on surveys and general harvesting when doing 4-5 nodes in rapid succession. I haven’t really seen any change in the basic ratio (low-mid 30s). If I feel like it’s changed or someone posts a larger sample size showing numbers within the error range for 50%, I’ll do another run of 500-1000 to get more data.

As it stands now, we have multiple folks reporting large samples pointing to an actual benefit in the range of 35%. Until we have proof otherwise, that’s the consensus rate of return from the data.

I’m personally kinda “ok” with the benefit around 35% - that’s 3-4 times better than what we had before. As it stands now imo, it’s good for the farmers but doesn’t completely destroy their market due to oversupply.

However, I would certainly like ZOS to weigh in on this. Tooltip is 50% but so far we clearly aren’t seeing anywhere near those numbers. Is this a bug? Is this an incorrect integer? Or is this rate intended and the tooltip needs to be corrected?

I hate to keep tagging you folks, but @ZOS_GinaBruno or @ZOS_JessicaFolsom, do you guys know about this?

wills43b14_ESO

Hi guys,

There's a lot of misinformation on this page, so I thought I would present both a numerical and analytical approach to clear some stuff up. I don't have an in-game sample to look at how the possible outcomes are distributed. However, the previous theory of harvesting a node to get materials, then having a 50% chance to essentially harvest it a second time sounds like a reasonable theory to test. Essentially, the argument is saying if you "succeed", it's like you harvested two nodes and got a random amount of material from each one.

Let's say a node can generate 1, 2, 3, or 4 materials.

If we harvest a node and fail the dice roll (don't get the passive), we can get 1, 2, 3, or 4 material.

If we harvest it and succeed in our roll (50% chance), then we can get: 1, 2, 3, or 4 material the "first" time AND 1, 2, 3, or 4 material the second time (note: this is NOT doubling - it basically acts like you get to harvest a second node).



Under this assumption, we can come up with a basic algorithm to figure out how many we should get:

Let's take 10,000 samples and add them to a list called "samples"; here is the algorithm:

1. Draw a random value from 1, 2, 3, or 4 (we'll assume each value is equally likely).
   
2. Now, roll a die.
   
3. If the die reads less than or equal to 3, we take the value we just had and add it to our samples list (anywhere from 1-4 material). If it is greater than 3 (50% chance of getting to harvest a second time), draw a second random value from 1, 2, 3, or 4 and add it to the first value we drew and then add it to the samples list (could get anywhere from 2-8 total material in this case).
   
4. Repeat 10k times.

Note: we can harvest 1 + 0, 2 + 0, 3 + 0, 4 + 0, 1 + 1, 2 + 1, ... , or 4 + 4 material based on what happens (anywhere from 1-8 material; but most likely to get 4 or less as many combos = 2, 3, or 4!)

I'm now going to graph how many times we harvested 1, 2, 3, 4, 5, 6, 7, or 8 material.




As we can see, you're mostly going to get 4 or less material. In fact there is a 70% chance you will get 4 or less. You will only get 5 or more 30% of the time. Also, cool fact, the frequency of 4 or more material is 50% (half) of getting less than 4!


CONCLUSION/TL;DR: If this algorithm is true, ZOS is likely correct, it's just not interpreted the way we would have thought.










For math peeps:
E[x + 0.5x] = E[1.5x] = 1.5 * E[x] = 1.5 * mean of x

Mean of values 1:4 (weighted equally) is 2.5 - if we take 1.5 * 2.5, we get a mean of 3.75. If I take the mean of those samples (or at least my 10k samples), we actually get an average of 3.73 material (median of 4, which matches the histogram), which is likely within Monte Carlo Error.

Code in R:

draw = c()
for (i in 1:10000) {
temp = sample(1:4, 1)
rng = runif(1,0,1)
if(rng > 0.5) {
temp = temp + sample(1:4, 1)
}
draw = c(draw, temp)
}

hist(draw)
summary(draw)

redspecter23

wills43b14_ESO wrote: »

Hi guys,

There's a lot of misinformation on this page, so I thought I would present both a numerical and analytical approach to clear some stuff up. I don't have an in-game sample to look at how the possible outcomes are distributed. However, the previous theory of harvesting a node to get materials, then having a 50% chance to essentially harvest it a second time sounds like a reasonable theory to test. Essentially, the argument is saying if you "succeed", it's like you harvested two nodes and got a random amount of material from each one.

Let's say a node can generate 1, 2, 3, or 4 materials.

If we harvest a node and fail the dice roll (don't get the passive), we can get 1, 2, 3, or 4 material.

If we harvest it and succeed in our roll (50% chance), then we can get: 1, 2, 3, or 4 material the "first" time AND 1, 2, 3, or 4 material the second time (note: this is NOT doubling - it basically acts like you get to harvest a second node).



Under this assumption, we can come up with a basic algorithm to figure out how many we should get:

Let's take 10,000 samples and add them to a list called "samples"; here is the algorithm:

1. Draw a random value from 1, 2, 3, or 4 (we'll assume each value is equally likely).
   
2. Now, roll a die.
   
3. If the die reads less than or equal to 3, we take the value we just had and add it to our samples list (anywhere from 1-4 material). If it is greater than 3 (50% chance of getting to harvest a second time), draw a second random value from 1, 2, 3, or 4 and add it to the first value we drew and then add it to the samples list (could get anywhere from 2-8 total material in this case).
   
4. Repeat 10k times.

Note: we can harvest 1 + 0, 2 + 0, 3 + 0, 4 + 0, 1 + 1, 2 + 1, ... , or 4 + 4 material based on what happens (anywhere from 1-8 material; but most likely to get 4 or less as many combos = 2, 3, or 4!)

I'm now going to graph how many times we harvested 1, 2, 3, 4, 5, 6, 7, or 8 material.




As we can see, you're mostly going to get 4 or less material. In fact there is a 70% chance you will get 4 or less. You will only get 5 or more 30% of the time. Also, cool fact, the frequency of 4 or more material is 50% (half) of getting less than 4!


CONCLUSION/TL;DR: If this algorithm is true, ZOS is likely correct, it's just not interpreted the way we would have thought.










For math peeps:
E[x + 0.5x] = E[1.5x] = 1.5 * E[x] = 1.5 * mean of x

Mean of values 1:4 (weighted equally) is 2.5 - if we take 1.5 * 2.5, we get a mean of 3.75. If I take the mean of those samples (or at least my 10k samples), we actually get an average of 3.73 material (median of 4, which matches the histogram), which is likely within Monte Carlo Error.

Code in R:

draw = c()
for (i in 1:10000) {
temp = sample(1:4, 1)
rng = runif(1,0,1)
if(rng > 0.5) {
temp = temp + sample(1:4, 1)
}
draw = c(draw, temp)
}

hist(draw)
summary(draw)

The problem with your assumption is that materials don't drop at a 1, 2, 3, 4 rate. The math after that may or may not be accurate. I didn't check it, but it appears you put a ton of work into it.

If you want a solid, easy to track set of tests, you can go with alchemy nodes. They will typically drop 1 ingredient on a standard pull and 2 on a "double" pull. It's very easy to test over a large amount of nodes what your drop rate actually is for that particular situation. Even when the math is as easy as 1 or 2 only, we are still seeing a trend toward mid 30% chance to double. Over thousands of points of data, this should be statistically near impossible.

wills43b14_ESO

@redspecter23 I haven't played much this patch, how are the "doubles" shown? Do you just get a stack of 2 instead of 1? My example was for something like a clothing crafting node. I forgot what categories you can get, but I just assumed 1-4 to kind of show what's going on. I can do an analysis for 1-3 or one for alchemy types as well.

wills43b14_ESO

And to say anything, I honestly just need some kind of sample. There's not much I can say without one.

wills43b14_ESO

@tmbrinks corrected me on the nodes that they're 3 or 4, so it should be 6, 7, or 8. In which case, you should get < 5 (no double) 50% of the time, and > 5 50% of the time. I take it back, there's something clearly wrong. I was thinking each node gave 1-4 for some reason lol XD.

lillybit

I think I understand what you're saying but nodes don't work like that. If I harvest a silk node without the passive I'll get 3 or 4; with it it'll be 6, 7 or 8. There is no way for it to be doubled and return 4 because a single drop will never be 1 or 2. Likewise alchemy will always return a single mat and doubled it's 2. There isn't a way to confuse two low drops for a single drop in either case.

With surveys exactly the same applies, just with higher numbers. The silk will be 18 or 20 and doubles to 36, 38 or 40.

For Enchanting it does work that way, they just seem to drop whatever they feel like and it's really hard to say if it's doubled or not. I'm pretty sure it was mentioned that they weren't included in the calculations for that reason tho.

For everything else it's easy to say with 100% certainty if a node was doubled or not.

redspecter23

wills43b14_ESO wrote: »

@tmbrinks corrected me on the nodes that they're 3 or 4, so it should be 6, 7, or 8. In which case, you should get < 5 (no double) 50% of the time, and > 5 50% of the time.

Those numbers are correct for nodes such as clothing, blacksmithing, etc. Basically those people testing the node drops are just noting if they are less than 5 or greater than 5. As you noted, each of those situations should be approximately 50% over a large enough sample size. The data so far is indicating otherwise.

wills43b14_ESO

Interesting, I wonder if there is a multiplicative factor in there somewhere that's throwing it off. Like they're multiplying two decimals together.

wills43b14_ESO

redspecter23 wrote: »

wills43b14_ESO wrote: »

Hi guys,

There's a lot of misinformation on this page, so I thought I would present both a numerical and analytical approach to clear some stuff up. I don't have an in-game sample to look at how the possible outcomes are distributed. However, the previous theory of harvesting a node to get materials, then having a 50% chance to essentially harvest it a second time sounds like a reasonable theory to test. Essentially, the argument is saying if you "succeed", it's like you harvested two nodes and got a random amount of material from each one.

Let's say a node can generate 1, 2, 3, or 4 materials.

If we harvest a node and fail the dice roll (don't get the passive), we can get 1, 2, 3, or 4 material.

If we harvest it and succeed in our roll (50% chance), then we can get: 1, 2, 3, or 4 material the "first" time AND 1, 2, 3, or 4 material the second time (note: this is NOT doubling - it basically acts like you get to harvest a second node).



Under this assumption, we can come up with a basic algorithm to figure out how many we should get:

Let's take 10,000 samples and add them to a list called "samples"; here is the algorithm:

1. Draw a random value from 1, 2, 3, or 4 (we'll assume each value is equally likely).
   
2. Now, roll a die.
   
3. If the die reads less than or equal to 3, we take the value we just had and add it to our samples list (anywhere from 1-4 material). If it is greater than 3 (50% chance of getting to harvest a second time), draw a second random value from 1, 2, 3, or 4 and add it to the first value we drew and then add it to the samples list (could get anywhere from 2-8 total material in this case).
   
4. Repeat 10k times.

Note: we can harvest 1 + 0, 2 + 0, 3 + 0, 4 + 0, 1 + 1, 2 + 1, ... , or 4 + 4 material based on what happens (anywhere from 1-8 material; but most likely to get 4 or less as many combos = 2, 3, or 4!)

I'm now going to graph how many times we harvested 1, 2, 3, 4, 5, 6, 7, or 8 material.




As we can see, you're mostly going to get 4 or less material. In fact there is a 70% chance you will get 4 or less. You will only get 5 or more 30% of the time. Also, cool fact, the frequency of 4 or more material is 50% (half) of getting less than 4!


CONCLUSION/TL;DR: If this algorithm is true, ZOS is likely correct, it's just not interpreted the way we would have thought.










For math peeps:
E[x + 0.5x] = E[1.5x] = 1.5 * E[x] = 1.5 * mean of x

Mean of values 1:4 (weighted equally) is 2.5 - if we take 1.5 * 2.5, we get a mean of 3.75. If I take the mean of those samples (or at least my 10k samples), we actually get an average of 3.73 material (median of 4, which matches the histogram), which is likely within Monte Carlo Error.

Code in R:

draw = c()
for (i in 1:10000) {
temp = sample(1:4, 1)
rng = runif(1,0,1)
if(rng > 0.5) {
temp = temp + sample(1:4, 1)
}
draw = c(draw, temp)
}

hist(draw)
summary(draw)

The problem with your assumption is that materials don't drop at a 1, 2, 3, 4 rate. The math after that may or may not be accurate. I didn't check it, but it appears you put a ton of work into it.

If you want a solid, easy to track set of tests, you can go with alchemy nodes. They will typically drop 1 ingredient on a standard pull and 2 on a "double" pull. It's very easy to test over a large amount of nodes what your drop rate actually is for that particular situation. Even when the math is as easy as 1 or 2 only, we are still seeing a trend toward mid 30% chance to double. Over thousands of points of data, this should be statistically near impossible.

Also, I didn't put too much work into it. I think the actual analysis itself took like 10s lol.

Skullstachio

Good thing I don’t use plentiful harvest then.

tmbrinks

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

LoneStar2911

tmbrinks wrote: »

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

They're not arguing whether or not it's better than nothing.
They're arguing that the tooltip is incorrect/false. It's ~30%. Not 50%, like it states. Therefore, ZOS needs to fix the tooltip or fix their algorithm.

lillybit

LoneStar2911 wrote: »

tmbrinks wrote: »

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

They're not arguing whether or not it's better than nothing.
They're arguing that the tooltip is incorrect/false. It's ~30%. Not 50%, like it states. Therefore, ZOS needs to fix the tooltip or fix their algorithm.

This! It also costs a lot of CP for it to be giving that much less than stated.

I've only managed to put 40 in for my alt so I'll have to test if that gives the same 1/3 ish drop next time I catch up with surveys - just in case we're better off leaving it at 40% for now!

Rittings

I still think it’s a shame, and would be easier to track, that they nerfed it from 100% to 50%... it would solve so many issues. Not only that, seeing as it takes a fair few cp to max the passive, it would mean real players vs bots would mean bots would be on the losing side for a change. It could have almost solved their bot problem in one genius move.

lillybit

Rittings wrote: »

I still think it’s a shame, and would be easier to track, that they nerfed it from 100% to 50%... it would solve so many issues. Not only that, seeing as it takes a fair few cp to max the passive, it would mean real players vs bots would mean bots would be on the losing side for a change. It could have almost solved their bot problem in one genius move.

That's what a lot of people were saying on PTS, but rather than sort out the horrible, horrible pathing and other issues they cut it in half, which (almost!) nobody wanted

tmbrinks

LoneStar2911 wrote: »

tmbrinks wrote: »

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

They're not arguing whether or not it's better than nothing.
They're arguing that the tooltip is incorrect/false. It's ~30%. Not 50%, like it states. Therefore, ZOS needs to fix the tooltip or fix their algorithm.

I'm not saying that in any, way, shape, or form.

The post I quoted was saying "Good thing I don't use plentiful harvest then..."

I was saying that 30% is still better than getting 0%, regardless of what the tooltip says (and still better than the 10% we were getting before)

I'm one of the strongest advocators of data-based testing and drop rate research here in this subforum. I hate using anecdotes and "feels" to describe what's happening, preferring hard data.

But 30% is still better than 0%.

I also tested with about 600 survey nodes of my own the other day, got a 40.2% doubling rate. Still statistically below the 50% we should be getting. But still WAY better than the 10% we were getting before.

LoneStar2911

@tmbrinks - Gotcha. I understand what you said before better now.

ThorianB

I am running about 38% on this and 5.7% on homemaker this week. I just started keeping track again last friday. Figured i would keep track for a week and see what happens. I am just shy of 500 normal nodes and 87 print dropped. Not doing a lot of farming this week because of the event.

tmbrinks wrote: »

LoneStar2911 wrote: »

tmbrinks wrote: »

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

They're not arguing whether or not it's better than nothing.
They're arguing that the tooltip is incorrect/false. It's ~30%. Not 50%, like it states. Therefore, ZOS needs to fix the tooltip or fix their algorithm.

I'm not saying that in any, way, shape, or form.

The post I quoted was saying "Good thing I don't use plentiful harvest then..."

I was saying that 30% is still better than getting 0%, regardless of what the tooltip says (and still better than the 10% we were getting before)

I'm one of the strongest advocators of data-based testing and drop rate research here in this subforum. I hate using anecdotes and "feels" to describe what's happening, preferring hard data.

But 30% is still better than 0%.

I also tested with about 600 survey nodes of my own the other day, got a 40.2% doubling rate. Still statistically below the 50% we should be getting. But still WAY better than the 10% we were getting before.

I really don't see anyone saying " I got 54%" or " I got 63%" . Everyone is reporting under. I am sure there are people getting more than 50%, but i am guessing it is a lot less than people getting under 50%.

My last time keeping track i only did about 120 nodes and pulled just under 40% as doubles. It's the highest so far of the ones i have kept track of.

tmbrinks

ThorianB wrote: »

I am running about 38% on this and 5.7% on homemaker this week. I just started keeping track again last friday. Figured i would keep track for a week and see what happens. I am just shy of 500 normal nodes and 87 print dropped. Not doing a lot of farming this week because of the event.

tmbrinks wrote: »

LoneStar2911 wrote: »

tmbrinks wrote: »

Skullstachio wrote: »

Good thing I don’t use plentiful harvest then.

I mean...

It's supposed to be 50%

We're getting ~30%

Not using it you get 0%

It's still better than nothing...

They're not arguing whether or not it's better than nothing.
They're arguing that the tooltip is incorrect/false. It's ~30%. Not 50%, like it states. Therefore, ZOS needs to fix the tooltip or fix their algorithm.

I'm not saying that in any, way, shape, or form.

The post I quoted was saying "Good thing I don't use plentiful harvest then..."

I was saying that 30% is still better than getting 0%, regardless of what the tooltip says (and still better than the 10% we were getting before)

I'm one of the strongest advocators of data-based testing and drop rate research here in this subforum. I hate using anecdotes and "feels" to describe what's happening, preferring hard data.

But 30% is still better than 0%.

I also tested with about 600 survey nodes of my own the other day, got a 40.2% doubling rate. Still statistically below the 50% we should be getting. But still WAY better than the 10% we were getting before.

I really don't see anyone saying " I got 54%" or " I got 63%" . Everyone is reporting under. I am sure there are people getting more than 50%, but i am guessing it is a lot less than people getting under 50%.

My last time keeping track i only did about 120 nodes and pulled just under 40% as doubles. It's the highest so far of the ones i have kept track of.

Yeah, there is probably some bias in the sampling for that... IF people were getting those numbers... they're unlikely to report them as a "problem" since it's above the sample.

That said, with the 600 samples I did, 99% of the time we'd expect to be between a 45% and a 55% drop rate, so getting 30/35/40%, we know we are in the realm of "statistically improbable" results, which means we can say the drop rate is NOT 50% as advertised

VaranisArano

@ZOS_GinaBruno @ZOS_Adrikoth and any other ZOS developers who care that you've got a problem with Plentiful Harvest...well, you've got a problem with Plentiful Harvest. I beg your pardon for the image heavy post, but its the best way to share the data.

First off, yes, I do have Plentiful Harvest slotted and maxed out.


See that 50% chance? I'm harvesting wood, ore, cloth, water, and flower/mushroom nodes. I should be getting 50% of those nodes doubled. Just so we're clear, a doubled wood/ore/cloth node gives 6-8 mats. A doubled water node give 4-6 mats. A doubled alchemy reagent give 2 mats. I did not track housing mats because those can give 1-4 mats, and so its impossible to tell when they doubled.

If Plentiful Harvest is working correctly, my results should approach 50% in a large sample size.

Spoiler alert: They do not.



Result: 40% Doubled / 60% Single

ZOS, you need to fix this. If I'm going to spend 50 Champion Points on 50%, I'd like to actually see a result that approaches 50% consistently. If I'm only going to get 40%, then perhaps you should consider adjusting the cost accordingly.

VaranisArano

More data:

Half of my data (1,060 node) was done in three batches, more or less three Craglorn loops plus a short run in Bal Foyen. Here’s the breakdown.



So there was some speculation that given our limited data pool, that if the Plentiful Harvest was truly proccing 50% of the time, that players who weren’t reporting must be getting above 50% some of the time.

My observation based on my tally marking was that this would not happen with any real length of farming run, since the disparity between the doubled and undoubled nodes only becomes more apparent with time.

So what about small batches? Most players don’t farm for an hour, like I do. Most players pick up mats over a short period. So what happens with players who only pick up 10 nodes at a time?

I’m glad you asked. I tracked 1000 nodes in batches of 10, noting down how many were doubled/single.
Here I’m displaying results in terms of how many times I got a certain result in that batch of 10. Markings such as 2D/8S should be read as “2 Doubled, 8 Single”.


Result: while its possible to double over 50% of a small batch, its much more likely to get less than 50% doubled. Specifically, I got less than 5 nodes doubled 70% of the time.

There's a big caveat to that data set: all those small batches don’t really capture what happens when we correctly aggregate the data from those small batches into one whole. Then, we really see the disparity.

Result: 39% doubled/ 61% single

Final Question. What happens if you pull together all the data that everyone on the forums has posted about their drop rates on this and the other thread?

Result: 35% nodes doubled I'm not sure why that result is so much lower than mine, but its certainly trending in the wrong direction for Plentiful Harvest to be working correctly. (Also, sorry that I didn't catch that Microsoft Word spellchecker on your names, folks!)

Anyways, thanks for coming to my TED Talk about why my Plentiful harvest isn't giving the stated results. Fix your game, ZOS!

lillybit

I can add 38%, tracked from my surveys - I have a breakdown but it's on my laptop and I can't get to it right now

Rittings

This is the same with lockpicking results too... but of course, it becomes even more serious to end gamers if we start to question these percentiles, then it would be fair to assume that crit chance percentiles are off too. Anything that isn’t a flat increase, but instead left to what should be rng, is up for scrutiny. Of course you’d have to be on pc to see those figures with the tracker since us on console don’t have that level of feedback available to us.

VaranisArano

Rittings wrote: »

This is the same with lockpicking results too... but of course, it becomes even more serious to end gamers if we start to question these percentiles, then it would be fair to assume that crit chance percentiles are off too. Anything that isn’t a flat increase, but instead left to what should be rng, is up for scrutiny. Of course you’d have to be on pc to see those figures with the tracker since us on console don’t have that level of feedback available to us.

PC has Combat Metrics and ESO Logs. I don't believe anyone on PC ever said their crit chance wasn't working properly. In any case, a lot of ESO Logs data is publicly available.

And I found the opposite result from you with my lockpicking results. https://forums.elderscrollsonline.com/en/discussion/413835/double-checking-zos-force-lock-percentages

Rittings

Yeah I saw your post on that... I went back and logged another 1000 chests in thieves guild heists... still was around the 55-60 range... maybe it’s those particular chests in a closed instance...

FrancisCrawford

Rittings wrote: »

Okay... 170 surveys completed = 1,020 nodes

Double Pulls: 301 .... 29.5%
Single Pulls: 719

It's the worse than my previous pool... but my previous pool did atleast include non-survey nodes.

Here is something extra that seems quite damning too...

If you toss a coin 6 times, the chance of it being heads 6 times is 1/64 (which is pretty much what we should be seeing with the survey nodes... only a 1/64 chance of them all being single or double).

However, 16 of my surveys were ALL single nodes... and ZERO were all double. In fact, I'm yet to see a full double survey and I've done a ton of them. I mean... from 1/64 chance, my findings are closer to 1/10 chance of pulling a full single node survey... and 1/Infinity of a full double.

I think we need some answers on this one.

If the chances were really 1/64 of an all-single-node survey, you should have had 3 or so hits out of 170, not 16.

On the other hand, if the chance per node of it being single-node is more like 2/3, then the chance of an all-single-node 6-node survey is more like 64/729, which is much closer to your actual result.

lillybit

Here are my figures to add to the totals. I chose to keep track specifically of surveys (not enchanting) because I hadn't noticed anyone else do it and was kinda interested in what sort of pattern it would make. I didn't harvest anything else while doing them tho, so still represents 100% of the nodes harvested over a couple of sessions.

1878 total nodes
1164 singles
714 doubles

Nestor

About those small batches.....

I did about 12 Surveys today. First one, 1/6 of the nodes doubled, the second time, 5 of the 6 doubled, the rest, ranged from 0 to 2 Nodes per survey doubling.

Hit about 30 random nodes along the way, 8 of those doubled.

Should I feel like I am getting screwed out of Mats here? I guess if everyone is not getting what is promised, it is kind of OK. But I do feel like I have left a lot of materials, and tempers, on the ground, as it were.

VaranisArano

Rittings wrote: »

Yeah I saw your post on that... I went back and logged another 1000 chests in thieves guild heists... still was around the 55-60 range... maybe it’s those particular chests in a closed instance...

Sure, that might be the case? I was farming overland Wrothgar, for the most part. I dunno.