Plentiful Harvest CP star underperforming

Haroeris

After doing some farming since the patch dropped, it felt like the frequency of double drops was less than the 50% advertised by the Plentiful Harvest CP star, so I started keeping track. Out of 222 nodes, only 77 (34%) dropped double. The rate from clothier nodes was was even worse: 21 of 78 (27%)

volkeswagon

I didn't realize it was 50%. I thought it was like 20%

redspecter23

I was also noticing something closer to 33%. But it was just anecdotal. I'll try to keep track of some actual numbers. It's possible it's broken or just some insanely bad rng.

phaneub17_ESO

With herb and rune nodes I been getting almost a consistent x2 of everything about 40 to 60% of the time, depending which zone I'm in so its within the correct boundaries. Surveys are more noticeable because of the high yield it already generates, I'm getting 2 out of 6 nodes will generally double so it looks like 33% chance. At any rate its still much higher than the previous 10% chance it used to be, my bags are getting quite full hitting as I'm emptying out my hoard of Surveys I may need to sub again sooner than usual... you monsters!

Lugaldu

So far, I have definitely found much more than before, 50% could be true.

aussie500

It depends a fair bit on RNG as to how many extra drops you get, it is never going to be a straight 50% for everyone.

Qbiken

Think your sample size is way too low to make any conclusions to whether it's working or not.

Haroeris

Yeah small sample sizes and RNG certainly should be considered, though I figured n=200+ was a decent start for just one person within a couple hours. If others see the same thing and can add more, that’d build greater evidence of a potential bias towards lower rates (or refute, which would be great, but seems unlikely that I’m quite that unlucky that I’m getting 1:2 instead of 1:1 double:single drops)

Tevalaur

I've definitely noticed a consistently lower than 50% chance of doubling with full investment into that CP star after many hours of playing. I don't have a way to keep track of the numerical data you provided, but I've known since they changed that star's maximum benefit from 100% it was going to be a problem.

Even if it took twice as many points, I'd rather be told it's a 100% chance and be able to complain without being told I'm biased when I'm not getting the advertised returns. 100% isn't necessary from a yield perspective but more from the fact that I was going to be glad to get away from claims of "bias" or "anecdotal" when I never seemed to reach the advertised double-chance.

From most of my surveys, 1 in 6 nodes doubled. Occasionally it was 2 or even 3 nodes doubled, only once was it more than 3 nodes doubled. And I have run surveys for several hours/day for 2 of the 3 days since the update launched.

edit: 2 typos

Sapphire7

I'm not getting double nodes 50% of the time either. Out of a total of 622 nodes I got doubles 37.45% of the time.

Elo106

Yeah I had the same feeling so I kept track and it was 27% for me.

VaranisArano

I'm starting to keep track via the old pen-and-paper method, but it may take a considerable time before I've got a data set I'm comfortable drawing conclusions from.

Ringod123

This is all down to ZOS's terrible implementation of RNG, this game has the worst clumping in any RNG I have ever seen, has been like this since the game first launched, you have to have a HUGE sample set for you to actually reach the quoted chance.
Opening 100 nodes with a 50% of double can literally give you a return of maybe 20 nodes which held double, the next 100 can then give you 60 or so, once you have opened thousands, you will then see you average at about 50%.

rotten333

I was unimpressed myself after the number of CPs required to get there that could have been used elsewhere. The nodes were slightly better, yes, but the first chest I opened changed my mind. It was fatter than before and more than likely outperformed the nodes for value sake. I rate it worth the cps I spent now.

Fennwitty

I agree it feels bigger than 10%, but not quite 50%.

Maybe some nutty thing like scaled percentages or something but it doesn't match the expectation of a 50/50 shot at double.

virtus753

Fennwitty wrote: »

I agree it feels bigger than 10%, but not quite 50%.

Maybe some nutty thing like scaled percentages or something but it doesn't match the expectation of a 50/50 shot at double.

Yeah, if it said something like "improves the chances of getting double resources," then that could be much less than 50%. Like Battle Mastery/Flawless Ritual doesn't give you an extra 30/60% chance to apply a status effect -- it increases the (relatively small) chance of applying one by that amount of your current chance, so it adds between 0.3-6%/0.6-12% chance in all, iirc.

Plentiful Harvest clearly says it provides 50% chance of getting double resources at max, so if it's less than that it needs to be fixed. It should be pretty straightforward -- although very, very time consuming -- to test. While we're doing that, hopefully ZOS can double-check on their end that it's programmed correctly.

VaranisArano

I ran my first full loop of Craglorn today, and I'm afraid it left me with more questions than answers.

How are you all handling tracking multi-loot nodes? Because I'm seeing some weird behaviors with housing mats and the ore/dust doubling.

Looking at Plentiful Harvest, I would have expected everything in the node to get doubled. Regular range on nodes is 3 to 4, doubled is 5 to 8. Regular range on housing mats and the extra dust in ores was 1 to 2, doubled is 3 to 4. If so, I could track the base ore/wood/silk/reagent number, since if that one thing is doubled, the rest will be too.

But that's not what's happening.

Example 1: I got one ancester silk node that was 8 silk/1bast, followed by another that was 3 silk/ 3 bast.

Example 2: An ore node that was 3 ore/ 1 regulus/ 3 dust.

Example 3: A wood node that was 3 wood/4 heartwood.

Based on what I'm seeing, this makes me think that rather than doubling the whole node, Plentiful Harvest is doubling each thing in the node individually. (Which unfortunately means my data set is useless because I wasn't tracking everything in the node.)

If true, this could account for the bias in my data set, if Plentiful Harvest is proccing some of the time on the items I wasn't tracking (housing mats and extra dust.).


Does anyone have suggestions for what nodes I could farm that don't have multiple items that could be effected by Plentiful Harvest? I'm thinking alchemy water nodes.


Incidentally, my current (probably fatally flawed) dataset of alchemy reagents, ore, wood, silk, and dust nodes says: 319 undoubled, 181 doubled. Or out of 490 nodes (tiny sample size, I know), I see a 37% doubling chance.

Which is why I suspect that based on the behavior of multi-loot nodes, I'm missing some of the Plentiful Harvest procs from my data set.


Anyone else have any ideas?

tmbrinks

"Double" node doesn't refer to getting 2x what you get.

It means the game is doing a 2nd pull from the potential loot table for that node.

It's why for BS/WW/JC survey nodes. A single pull will get you 18 or 20 raw mats, but a double pull can get you 36, 38 or 40.

It's plausible that some of the overland nodes (like the 3 silk, 3 bast) was a "left-over" node that somebody already looted the silk from (leaving only Bast), but you got the proc for plentiful harvest, and got silk + bast on the 2nd pull.

I collected about 500 surveys over the last few days (was saving them for this passive), and I meant to take a video of my lootlog history so I could track double drops, but forgot to do so. ANCECODALLY, it seemed to be at about 50% (maybe slightly under, but that could be psychology at work). I had a few times I commented that "Wow! all 6 nodes were doubled!!!" and a few times where "Damn, none of those got doubled", but most seemed to be 2, 3 or 4 of the 6 nodes doubled.

VaranisArano

tmbrinks wrote: »

"Double" node doesn't refer to getting 2x what you get.

It means the game is doing a 2nd pull from the potential loot table for that node.

It's why for BS/WW/JC survey nodes. A single pull will get you 18 or 20 raw mats, but a double pull can get you 36, 38 or 40.

It's plausible that some of the overland nodes (like the 3 silk, 3 bast) was a "left-over" node that somebody already looted the silk from (leaving only Bast), but you got the proc for plentiful harvest, and got silk + bast on the 2nd pull.

I collected about 500 surveys over the last few days (was saving them for this passive), and I meant to take a video of my lootlog history so I could track double drops, but forgot to do so. ANCECODALLY, it seemed to be at about 50% (maybe slightly under, but that could be psychology at work). I had a few times I commented that "Wow! all 6 nodes were doubled!!!" and a few times where "Damn, none of those got doubled", but most seemed to be 2, 3 or 4 of the 6 nodes doubled.

That makes a fair bit of sense.

I guess my next question would be if Plentiful Harvest is deciding to do a double pull for everything in the node, or if it decides individually for each thing in the node.

Because my wood node example looks like single wood, double pull heartwood. Or my ore node was single ore, single regulus, double pull dust.

I suppose its not really fair to speculate on so few examples, so I'll have to keep track of more of my "outliers."

FrancisCrawford

My favorite so far has been 13 ruby ash from a single node during the event.

But that's consistent with any reasonable theory. A 4x haul would have had a 1% chance in the old system, a 25% in the new system if it's working as advertised, and somewhere inbetween if people's fears of underperformance are correct.

tmbrinks

VaranisArano wrote: »

tmbrinks wrote: »

"Double" node doesn't refer to getting 2x what you get.

It means the game is doing a 2nd pull from the potential loot table for that node.

It's why for BS/WW/JC survey nodes. A single pull will get you 18 or 20 raw mats, but a double pull can get you 36, 38 or 40.

It's plausible that some of the overland nodes (like the 3 silk, 3 bast) was a "left-over" node that somebody already looted the silk from (leaving only Bast), but you got the proc for plentiful harvest, and got silk + bast on the 2nd pull.

I collected about 500 surveys over the last few days (was saving them for this passive), and I meant to take a video of my lootlog history so I could track double drops, but forgot to do so. ANCECODALLY, it seemed to be at about 50% (maybe slightly under, but that could be psychology at work). I had a few times I commented that "Wow! all 6 nodes were doubled!!!" and a few times where "Damn, none of those got doubled", but most seemed to be 2, 3 or 4 of the 6 nodes doubled.

That makes a fair bit of sense.

I guess my next question would be if Plentiful Harvest is deciding to do a double pull for everything in the node, or if it decides individually for each thing in the node.

Because my wood node example looks like single wood, double pull heartwood. Or my ore node was single ore, single regulus, double pull dust.

I suppose its not really fair to speculate on so few examples, so I'll have to keep track of more of my "outliers."

Yeah, there are definitely going to be some more "weird" cases.

My question is... is it possible to get double Aetherial Dust?

I did get a triple Kuta from an enchanting survey when I was collecting them, and I have seen a few double Potent Nirn from a survey (although I got that before this change as well)

Nestor

VaranisArano wrote: »

I guess my next question would be if Plentiful Harvest is deciding to do a double pull for everything in the node, or if it decides individually for each thing in the node.

The way it was explained back when Plentiful Harvest was added, is each item is subject to the Dice Roll. So, the main crafting ingredient is subject to the doubling dice roll. With the secondary items, which used to be limited to mostly Bait items, there is a dice roll to see if you get a secondary item, then a dice roll to see if it is doubled.

Enchanting Nodes are weird in that we always got 2 or 3 items, so not sure if the dice rolls follow the same pattern on those, or if it just applied once to the initial pull. I do little enchanting nodes these days.

VaranisArano

Nestor wrote: »

VaranisArano wrote: »

I guess my next question would be if Plentiful Harvest is deciding to do a double pull for everything in the node, or if it decides individually for each thing in the node.

The way it was explained back when Plentiful Harvest was added, is each item is subject to the Dice Roll. So, the main crafting ingredient is subject to the doubling dice roll. With the secondary items, which used to be limited to mostly Bait items, there is a dice roll to see if you get a secondary item, then a dice roll to see if it is doubled.

Enchanting Nodes are weird in that we always got 2 or 3 items, so not sure if the dice rolls follow the same pattern on those, or if it just applied once to the initial pull. I do little enchanting nodes these days.

Thanks! That seems to fit in with what I'm seeing, and also explains some of the "missing" Plentiful Harvest percentages. I'm just not quite tracking everything that can be doubled.

I may have to focus on nodes like Lorkhan's Tears which don't drop multiple items to see that Plentiful Harvest is doing with straight one-item pulls.

Tevalaur

This seems to be much better performing since last night's "back-end maintenance"

Sapphire7

I'm also noticing a small improvement today out of 1718 nodes 41% where double.

VaranisArano

I started recording just the pure water nodes today so there was only one item in the node to be impacted by Plentiful Harvest. Waters give 2-3 on an undoubled node, 4-6 on a doubled node.

Tiny data set so far, but:

Doubled: 40/110 (36%)
Not Doubled: 70/110 (64%)

VaranisArano

Update on Pure Water nodes:

After 272 Pure Water nodes, I'm about done farming and ready to report my findings to ZOS. I have my Plentiful Harvest star fully maxed at 50% and slotted.

Out of 272 nodes:
97 doubled with 4-6 Lorkhans Tears (36%)
175 undoubled with 2-3 Lorkhans Tears (64%)

Frankly, if this were a coin toss, no one would hesitate to say there's something wrong with those results for a supposed 50% chance. I'll leave it to the statisticians to tell me how likely/unlikely I am to get that result, if they so desire.

I've reported this to ZOS as well. Ticket #210313-008728

Tevalaur

I'm back to FREQUENTLY having surveys where 0 nodes double... with 6 nodes per survey, it is only supposed to be a 1.5% chance of none doubling for those without statistical knowledge. This is definitely bugged when it's happening on roughly 25%-33% of surveys this morning.

Tevalaur

VaranisArano wrote: »

Update on Pure Water nodes:

After 272 Pure Water nodes, I'm about done farming and ready to report my findings to ZOS. I have my Plentiful Harvest star fully maxed at 50% and slotted.

Out of 272 nodes:
97 doubled with 4-6 Lorkhans Tears (36%)
175 undoubled with 2-3 Lorkhans Tears (64%)

Frankly, if this were a coin toss, no one would hesitate to say there's something wrong with those results for a supposed 50% chance. I'll leave it to the statisticians to tell me how likely/unlikely I am to get that result, if they so desire.

I've reported this to ZOS as well. Ticket #210313-008728

Did they respond to the ticket with anything beyond an auto-reply?

VaranisArano

Tevalaur wrote: »

VaranisArano wrote: »

Update on Pure Water nodes:

After 272 Pure Water nodes, I'm about done farming and ready to report my findings to ZOS. I have my Plentiful Harvest star fully maxed at 50% and slotted.

Out of 272 nodes:
97 doubled with 4-6 Lorkhans Tears (36%)
175 undoubled with 2-3 Lorkhans Tears (64%)

Frankly, if this were a coin toss, no one would hesitate to say there's something wrong with those results for a supposed 50% chance. I'll leave it to the statisticians to tell me how likely/unlikely I am to get that result, if they so desire.

I've reported this to ZOS as well. Ticket #210313-008728

Did they respond to the ticket with anything beyond an auto-reply?

I reported it via the category that doesn't usually get a reply, so I didn't expect one.

Rittings

I hadn't spotted this thread... but yeah, I tallied 1,000 nodes today as a good sample size - most of which were my survey nodes...

313 doubles.... for 31.3%

I didn't include any rune stones as they can be a little janky on returns.