The Age of Winter 555 corrupted timline cleared to see all types of nodes there

Feedback and Suggestions
62

The thing is, using a rarity tier weight list as a middle man is just adding an unecessary extra step. Might as well simply have a single loot table with your weights adjusted for what you want, since in practice that’s what you actually do when calculating the odds anyway:
-You calculate the chance of getting the desired tier (5/106 in your example for uncommons) then further divide it by the 5 uncommons in your pool which have an individual chance of 1/5 for that table.
Which is the same thing as making a single table where your weight for uncommons = 5/106/5. It’s just an extra step.

This is the same in any of the systems. If you triple the amount of items, you either triple the time required or you have to triple the drop rate.
You also have a global drop multiplier in LE. One which is even different for MG and CoF. Since it’s global, it really isn’t relevant for the discussion of these methods. It’s applied outside of them.
After all, you could implement #3 without the global multiplier, just like you can implement #1 or #2 with or without them. It isn’t relevant for the distribution discussion.

This is the same thing in any of the systems, though. I don’t see how you think it’s different in any of them. With your example, if you add a common (lets disregard the multiplier, since we’ve already seen it will affect everything) then your rare still needs 250h but your commons now need 3h to farm instead of 2h. And if you add a rare one, then your uber now needs 350h while your commons still require only 2h.
Unless you add them with the same ratio, farming times have to change for some or all of them. All you’re doing is deciding that which subset of your total pool gets changed.

Again, if you don’t adjust the global drop rate, it will happen with all 3 systems. And if you adjust the global drop rate, it doesn’t happen with all 3 systems.

PoE doesn’t separate by item type, meaning that when it rolls unique it will roll all uniques. So they separate their 1000+ uniques by tiers.
LE separates by item type. First you determine the item type and only then roll rarity, so uniques are already pre-tiered.

In fact, if PoE only has 3 tiers of rarity for uniques, each table will have 300+ items.
LE, on the other hand, has about 20-30ish per item type. That seems easier to balance since you’re dealing with smaller samples.

It’s no different from the other systems. Each item has a calculable drop chance. If you feel an item has a low drop chance, you decrease reroll chance, if you feel it’s too high you increase it.
Made easier by the fact that you’re only dealing with small subsets each time, which is the item type.

Also, using your example:
You have 50 commons with a 5000 weight, 7 uncommons with a 35 weight and 2 rares with 2 weight. If you want to simply add a single common item, you now have to find the common denominators for everything for your new values. You can’t simply change the 5000 weight to 5500, because that changes the distribution and makes common items drop more.

I think you would make your point clearer if you just created some real examples like I did on the spreadsheet.

63

Umh… no?
All it does is giving more control, that’s it. You can adjust it separately at the end.

All of the examples above with multi-layered loot tables only leads to a Stochastic Matrix in the end to put it all back into one again anyway. The difference solely is that before the final vector is created you have control over the specifics of the process… in detail.

After initialization you remove the computation need through the Stochastic Matrix after all and have the end-result which the game finally works with.
You have a really sophisticated starting system which then gets reduced into a dumb single-state outcome to work with actively.
It combines depth and simplicity into it.

Yes, so why have this intricate current system then? :slight_smile: The final outcome is after all… a single weight-table.
Could’ve done it beforehand, right?
But EHG wanted - because there’s literally no other reason for it - to adjust values dynamically.
The way they did it is just backwards, but the reason is the same as one would do in a tiered weight-table system.

And since the re-roll system itself is also not as effective in execution (I know… I know… neglectable, no matter though) they still could’ve just used that as a singular outcome table.

It’s both reduced to the same state after all, just the outcome is different… and the outcome is what makes or breaks it.

In one you have control over that step directly.
In the other not.
The end-result has the same status, only initialization takes longer/shorter for it.

No? Because I either don’t change the spread to each other (which enforces a specific item to be dropped a specific amount of times based on time-investment) or you don’t change the proportion to each other (which enforces a specific rarity of a item to be dropped a specific amount of times based on time-investment).
Either way you get a constant related to time-investment, which is important. What the constant exactly entails is up for decision and to be based on the respective situation.

In a market situation you want the overall ‘value’ to drop, hence the second option.
In a SSF situation you want the respective quantity to drop, hence the specific items and with that the first option.

Since we have a market situation basically the second is mandatory actually upon thinking about it (another topic but I’ll get shortly into it) and then using the SSF style to change this system for them accordingly to adjust drop-rates through the functionality for target farming given to them.

The current option has extreme effects on a market and also unsavory effects for the individual.

Exactly… so decide! What do you want? ‘A’ item of the rarity or ‘the’ item of the rarity?
That’s the sole difference of my provided situation.

In MG you want ‘a’ item of the rarity and in CoF you want ‘the’ item of the rarity.

Anything else is nonsensical as it changes the effort needed for each situation, which shouldn’t happen.

No? If I keep the proportional drop equivalent at all times and guarantee the quantity multiplier to adjust to it then this doesn’t happen.

Yes, I get ‘a’ common, not ‘the’ common.
Once more, which of the sides are you talking about?
The ‘the’ common side or the ‘a’ common side.
We have 2 possible states in my examples, as I’ve provided 2 examples. One for the proportion-based weight-distribution table and the other for the quantity-based weight-distribution table.

Depends as well!

If you simply want ‘a’ item of a category you don’t need to increase the quantity of drops. You get your specific item later but nonetheless get one of a rarity into your possession.

The example there is:

The ‘a item’ way:
If I have a ‘red ring’ and add a ‘ravenous void’ to it then I’ll only see half as many ‘red rings’… but in total I’ll see the same amount of items in this rarity category.

The ‘the item’ way:
If I have a ‘red ring’ and add a ‘ravenous void’ to it then the ‘red ring’ and ‘ravenous void’ drop at the same rate still. I get double as many of the category, but the time-investment for each individual item stays the same.

This is the difference between those 2 models.

One is good for trade, one is good for SSF.

This example is given with adjusting the global drop rate.
In the current system.
By adding some common uniques.

That’s the outcome then.

To avoid that you need to add a equivalent amount of the other rarities as well or they’ll all be less available outside of the category of rarity added to the pool.

That’s shit.

So why not have it in a fixed table then rather then this… whatever it is? :slight_smile:
After all as a player I want to get ‘a’ rare unique (to trade for ‘the’ rare unique) or ‘the’ rare unique if I can’t trade with ‘a’ rare unique.

If a system has different anchor values and hence different outcomes then it is different. What’s there even to discuss?

But I don’t know in the system if the other items will now have a relevant drop-rate compared to the others if I change this one. And I don’t know how many I need to change how much to achieve the proper proportions to each other.

Once more: How do you decide on the proper proportions in the existing system?

No, you have them declared, it’s a value, a multiplier.

In this case (for C++):
/do this for each rarity to dynamically create the array, add new members for each position, yes, can be separated, did it in one to make it easier./
class for type (){
i = 0;
class member
classmemberweight
++i
class member
++i
…}

for (i){
std::vector::push_back() }

Done for the size. Fill with the values.

The denominator:
int tablearray[3] = {100, 20, 1}; //you change it directly.

Done for the denominator for ‘the’ item drop.

int rare = 1, uncommon = 5, common = 20; //this is what you change.
int tablearray[3] ;
tablearray[0] = a x b x c;
tablearray[1] = a x b;
tablearray[2] = a;

Done with the denominator for ‘a’ item drop.

That’s all. Automated.

Now if you add to your 50 items (weight 5000) a 51th item it changes automatically to 5100.
You get ‘the item’.
Or if want ‘a’ item nothing changes as you just make more members.

Hope that helped :stuck_out_tongue:

Edit:

Since I’m a dimwit today seemingly:

You make the vector with vector’<‘i’>’ simply :stuck_out_tongue:
Dunno why my brain went off into making it dumb.

64

You don’t though. Which is what I’m trying to get across. If you add a common unique you might make it so that a subset remains the same (uniques and thus also the uber amulet) but then you’re making every single common item take longer. It’s inevitable.
And if you add a rare item to the pool, the uber amulet also changes and takes longer than 250h now.

You can’t make it so that adding a new item doesn’t change any of the previous items farm times. Either in total or a subset, something has to change. This is true in any of the 3 systems. The only difference is that with #3 you can say that only a subset changes, but items do change. There’s no getting around that.

It doesn’t. The variation between each system is small. Because you never add 50 rare uniques and 1 common. It’s a nonsensical idea. It’s a purely theorical outcome that has no use in reality.
You add new items in an approximate ratio to their rarity. Which makes each of the 3 systems vary only slightly.

A global drop rate multiplier does nothing to change “the” item drop chances. Not even “a” item drop chances. It just changes the time required to get either.
If you have an item that has a 1% chance to drop, you can make x10 drop rates and the chance remains at 1%. You just get more loot. You get the 1% item faster, but you also get all the trash 99% a lot more. This doesn’t change balance.

It really doesn’t matter. If we use your example and we add 5 commons, 3 uncommons and 1 unique, your uber amulet now takes longer to drop.

#3 only favours “A” item. “The” is always screwed and will always take longer in a normal situation where you add a few of each tier.
#1 and #2 are in the middle of “A” and “The”. The chances to get “A” shift only a little and the chances of “The” also only shift a little.

That is what everyone does in every game, including PoE. There isn’t a league where they introduced more rare uniques than common ones. Nor does anyone do that. Like I said before, it’s a theorical argument that has no actual bearing on reality.

Because, as we’ve seen, the way the relative rarity of all items shifts with each system is slightly different. And EHG prefers this one.

If the variations are small, then they’re equivalent and become simply a matter of preference. What’s there even to discuss?

If you just add a common and it becomes 5100 without changing the weights for uncommons and rares, then commons drop more, which is contrary to your purpose.
As an extreme, let’s add 50 more and get 10000 for commons, 35 for uncommons and 2 for rares. Now your commons drop twice as much as they did before. Which means the uber amulet now takes 500h to farm.

It’s irrelevant for the same reason as above, because you don’t add just 50 commons without adding also uncommons and rares, but your calculations fail in that.
It would have to become 10000, 70 and 4 to maintain the “a” method. Which means that if you add only 1, you’d have to add fractions or find a common denominator to maintain integer values.

65

Ok, let’s go a step back, I’ll provide a base example with all the math visible (but not in numbers cause I suck with numbers)

We have a time constant.

This constant in our example is ‘100 items dropped her hour’. Which is our base quantity value of ‘1’ and out general base value of ‘100’ for the percentage we need to derive from it…
Every point of weight represents ‘1’ item in our table for the start.

Our table is:

94:5:1 (for a total of 100)
In words:
We get 94 common items per hour.
We get 5 uncommon items per hour.
We get 1 rare item per hour.

I want to add a new item, 2 problems.
In ‘Static outcome’ I want to enforce that every specific item drops in the hour.
In ‘Variable outcome’ I want to enforce that the same amount of items per category drop.

How do you do that?

Solution:

Example Static outcome:
Our equation:
Value per table = content quantity in table
Quantity = content quantity in sum of all tables / 100 (=base value)
Global Multiplier = All multipliers put together relevant outside of it.
Table change:
94:5:2
Table weight multiplier:
1 → 1.01 (choosing this table more often, 101 times instead of 100).
Quantity multiplier:
1 - 1.01 (increasing the global item drop from 100 to 101 items).

Explanation:
We have 101 positions instead of 100.
To achieve that every item drops we need to increase the time this table is picked by 1%
This causes the table to be picked 101 times in an hour.
Since we need to drop more items in general we need to increase the global drop-rate by 1% to get 101 instead of 100 items.

Result:
Time to see every item: 1 hour…
Time to acquire ‘the’ rare item: 1 hour.
Time to acquire ‘a’ rare item: 30 minutes.

Example Variable Outcome:
Our equation:
Value of tables are fixed proportionally.
Table change:
94:5:1
Table weight change:
1 → 1
Global Quantity multiplier change:
1 → 1.01

Explanation:
We don’t need to change the value of the table. We want 94 commons to 5 uncommons to 1 rare items. This is upheld.
The content of the table is not equivalent to the weight anymore.
We change the times the table is chosen since the variables inside out tables have increased.
We leave the table weight since it’s only interesting for is to get the relative amount.
We change the quantity to align with total item drops staying in the same timeframe.

Result:
Time to see every item: 2 hours.
Time to see ‘the’ rare item: 2 hours.
Time to see ‘a’ rare item: 1 hour.

Extra:

Example ‘Undefined outcome’, the current LE example in theoretical, no major defined numbers:

3 items. One fixed, one 50% re-roll, one 90% re-roll.

Rare item (your example):

1/3 x 0.1 + (1/3 x 0.5 * 1/2 x 0.1) = 0.078 or 7.8%
Weight = 78 out of 1000.
Uncommon = don’t know but defined.
Common = don’t know but defined.

Adding a new item, result:

Behavior: undefined.
Weight: undefined.
Related time investment: undefined.

‘It depends on what exactly you do but you can’t say without a algorithm what it will do before testing it out personally. You only have a direction, no actual solution’.

Those examples present the whole conundrum.

For MG you want ‘Variable Outcome’.
For CoF you want ‘Static Outcome’.

I don’t know where you would like ‘Undefined Outcome’ because it’s not defined. It is neither anchored in time, in
I don’t know what the goal is to be achieved as it’s not presented to me.
New input causes positive and negative outcomes for MG and CoF at the same time.
I can’t see which measures are to be taken for a decisive outcome, I only know the direction without extra help.

66

This is what I mean. You have MASSIVE time changes on your drops by adding just 1 item. You either get massive time changes to get “a” or “the”. Whereas the other 2 methods only slightly change each of them, meaning that adding a rare you’d end up with something like:
Time to see every item: 1.2h
Time to see “the” rare item: 1.5h
Time to see “a” rare item: 0.9h

You can simulate what the results will become and adjust accordingly.
And if you want to keep the rates, you increase global drop rates a bit so you end up with:
Time to see every item: 1h
Time to see “the” rare item: 1.25h
Time to see “a” rare item: 0.75h

But either way you don’t get such drastic changes as you do with your method. Because it slightly changes both, rather than drastically one or the other like your method.

Yes, but they cause smaller negative outcomes, rather than forcing you into screwing one or the other.

Honestly, I don’t know why you keep insisting on this. This only works for PoE because PoE doesn’t give a s*** about people not wanting to trade and doesn’t mind screwing over SSF players.

67

Yes? So?
It depends how you use em, that’s why I specifically said ‘this is the option for MG’ and ‘this is the option for CoF’.

They have different use-cases.

I just can’t find a use-case for our existing system here :stuck_out_tongue:

In MG I wanna have a item which has value and in CoF I wanna have a item I can use.
That’s all the player cares about, or have I missed something?

You start with the MG model and put an overly on top for the CoF model, adjusting values accordingly. This way neither is screwed.

Why starting with MG? Because you can add with that baseline since it adjusts less values. You only need to change the weight of the tables accordingly and it turns into the fitting variant, which should be easer then removing already fixed values. You just ‘add’ onto it.

Have I ever said that?

Where have I even given an inkling of that?

Look at the math there, one method is a reduction of the other.
Hence you can adjust the drops after implementing the Variable Outcome with CoF’s reward system to a Static outcome.

Which isn’t done but that present the respective starting points to achieve that.

Think about solutions before saying someone acts in bad faith.

Also who said this is the exact example of PoE? This actually entails the exact situation of every diablo-clone on the market, LE is the exception.
And exceptions have 2 distinct ways to go:
It improves stuff.
It makes stuff worse.

Which one do you think this one does? You have the answers above already.

68

You said that it either takes twice as long for “a” or twice as long for “the”. This only works when you only care for one of them (in PoE’s case “a”) because you only care about balancing trade.
So when PoE adds new uniques, trade remains the same because you still get “a” regularly, SSF is screwed because it takes longer for “the”.

This isn’t true, however. No other diablo-clone uses a unique tier system. We’ve discussed this. Other diablo-clones use a simple weight table (#2 method) with all uniques thrown in there. Which isn’t that different from LE’s (#1 method).

So the one that is actually the exception is PoE. Because the variance in either time for “a” or “the” will massively increase, whereas the other 2 methods make a much more gradual change, meaning that neither “a” nor “the” are too affected, just slightly.

69

Yes, low level mechanics tend to do such stuff. You create the intricacies above that level to tackle the shortcomings.

You only have a choice between a or b, the solution ‘c’ is shit for both sides as nobody can work with it.

What’s the solution?

Pick one of em and handle the shortcomings.

Which one has the bigger shortcomings? Ah yes, the one which is static, which is for CoF, since it does more then the other.

Hence use the Variable one and then pull it up for the users of CoF through their mechanic to make it the Static one.

All…
and I repeat… all diablo clones with a market use a Variable loot based system as their foundational decision maker for items every step through the whole decision line and for emphasis Last Epoch is the sole existing exception to this premise baseline function while having a market. If you want to prove it otherwise then show me another one.

Nobody… and also once more for emphasis… nobody said that the Static method is not a viable method, but you can’t and shouldn’t ever even show an inkling of that towards anything with a market.

So what to do? As I repeatedly say and you seemingly ignore?

You use the one which is good for the market (since the other way around it doesn’t work) and then add the missing functionality from the one of the market onto the systems related to CoF to make it functional for their specific use-case.

What is the result?

MG doesn’t get swamped with items they shouldn’t get swamped in. As well as starved for items they shouldn’t be starved in, providing a solid progression with their faction in mind.

CoF gets not swamped with low value garbage they don’t need, they have the ability to counter the Variable time investment increase through the methods of their faction which then causes them to acquire items in a similar pace then a trader does, which is the whole damn premise of their faction after all.

Got it finally?

70

And I also repeat, all diablo-clones except PoE use a simple weight table like the #2 method in the sheet. If they have 300 uniques, they have a single table with every single unique inside them with their weights.
And when you add a new one, you either get the same variance you get with LE or you have to tweak all values, like with LE.

The unique weight table started with D2 (which was the first to introduce unique items as general drops), which even has dozens of tables depending on which mob/boss you killed and in which difficulty (and in D2R, another for when a zone is terrorized), with different weights between them.
And we’ve already seen that method #2 behaves much like method #3 with small variances.

The whole “this unique is in the tier1 pool and that unique is in the tier2 pool and you either get one pool or the other” is exclusive to PoE, apparently.

Again, this doesn’t happen in any practical case. This would happen if EHG suddenly went “Oh, what will happen if this cycle we just add 50 rares and 0 commons?” which will never happen.
If you have 15 commons and 3 rares, when you add a new rare you also add 5 commons. You usually maintain the ratio or close to it, because it makes no sense for a game to have more rare items than common ones and no game I know of does that.
So when you do add those items, the ratio of common unique drops to rare unique drops remains the same or close enough.

And CoF still gets their bonuses applied just the same. I honestly don’t see how you think that those methods are any different outside small variations for any practical real world case. There are only vast differences when you theorize scenarios that never happen in real life.

71

Yes, and when’s that unique chosen? Another table, likely the rarity table. And when’s that chosen? When a function for creating a item is triggered. They’re all building onto each other.

So what does that create? A stochastic decision matrix.

What does that create? A single weighted table as a outcome at the end.

Where do you adjust values? In the initial weight tables.

So which ones do you adjust? That depends if you want to enforce a Variable or Static weighted table. Those are the 2 general options.

So which one do they have? Likely Variable since the static would cause a mess for overall drop-rates, you tend not to touch the proportion unless you know exactly what you wanna achieve with it, which is solely reducing invested time over the whole category, which is barely ever done. You work that out once.

Sooo… instead of making a algorithm which is piss-easy and automates it for you to enforce everything stays in line with everything else… you let basically a intern do it? Working out those numbers?

‘Hey Bob… think this should be a 10 or a 13? I’m not sure’ ‘Eh… might be a 12? Maybe a 11? Ah, just write ‘something’’.

How inept must those developers be? :stuck_out_tongue:
You have layers upon layers of item drops, diablo clones literally are the most complex itemization systems out there… and you think they do it manually?

The weighted systems of electronical games started with gambling, it’s not exactly known which one was first but it’s expected that a old-style one armed bandit (they came in loads of variants) invented the first electronically commercialized system of weighted distribution through the usage of relays back then.

Using it for a unique table was new because of the number of uniques warranting it but the concept existed decades before and was regularly used, even some MUCKs used that.

Yes, it does. Because back then you didn’t use a stochastic decision matrix to reduce that whole mess into a singular table.

Which is why nowadays we only see a singular table.

It takes this whole process and recompiles it for you into a single usable vector nowadays, back when it was introduced it was still a normal array which you had to manually manage memory-wise and was then often split into segments depending on the number of fields of the array you wanted to go through and the overall range of numbers since that could take a decent time.

No, it isn’t. ‘Tiered’ only means you have several layers of something. If you got 5 values with ‘100’ and 6 values with ‘10’ you have a tiered system by design, just not organized.
Let’s not start on discussing ‘what is a tier?’ here as well.

It already does. It always did, it just makes more sense now why it happens and has become visible with MG.

Chronicon does. Very successful game just to note that here. Same genre even, simple player though.

Practical real world case?
0.9 to 1.0.

What’s chosen first? Item type or rarity?
If rarity is chosen first we had 9 common, 1 uncommon and 3 rare uniques added.
Doesn’t sound very well in line with what you’re saying, so your argument in that case is already screwed, more rares then uncommons by a lot, all is screwed, period.

If item type is handled first.
Well, guess what, also screwed!
A single common bow… guess I don’t need a Shadow String or Flight of the First!

Also screwed.

Wanna go again and re-think your argument?

Or want to math it out? Helmets only 0.9 to 1.0?
Bows only from 0.9 to 1.0?
Overall rarity only from 0.9 to 1.0?

How doesn’t it affect me in a real world case again?

Bow example:
7 types with 0 re-roll, 1 with 20% re-roll, one with 50, one with 60.
Change it to 6 with 0% and you have your outcome.

Tell me again that my 50% or 60% drop won’t be affected by that. Even Wraithlord Arbor will be affected despite a much bigger populated table for helmets.

You’re blindly defending a system by now which I’ve shown in theory that it’s worse.
Ok… in theory.
But the reality is very very close to that theoretical example, there’s not that many extra items in there, actually… the common rate is much much higher then in our theoretical examples.

3 examples here which are against your argument and reality.

And reminder… if your overall chance changes from 1,0% to 1,2% that’s a 20% difference in time investment.

72

But that also happens in LE. You get a table for item types, you get a table for rarity, the only difference comes from the last table of uniques where it uses reroll chance instead of weight.

But I’m honestly tired of discussing this. This whole discussion has all been about moving goalposts.
First that a weight table is more efficient.
Then I showed that the efficiency change is irrelevant, that the reroll table has bigger variance.
Then I created the sheet and showed that the variance is minimal, it became about this other system that has an extra tier.
Then when I showed that this other system does bring even bigger variances, that one was good for trade and the other for SSF and you could adjust with CoF (nevermind that PoE doesn’t have CoF, so every unique they add SSF gets screwed again, which, tbh, isn’t really surprising in PoE).
It keeps getting shifted to different scenarios when in reality the actual variations between weight table and reroll chance are minimal and LE also uses several weight tables as well.
It will never end and you’ll find some other way to shift it and keep it going, so I’m done.

In practical terms, there isn’t much difference between both, just slight ways in which they shift. But you will never be convinced of that, so I won’t bother keeping this up.

73

Exactly my point.

I argued ‘something’s not right with the system of how LE does it’.

And the result is:
Higher in computation (no matter if you say relevant or not, that’s the case)
Harder to manage.
Non-optimal outcome for everyone despite alternative options.

Is that the case or not?

Minimal != non-existent.
Also it’s not minimal. Convert the % into time and you see it’s substantial.

Which is why they have the divination system to counter that :slight_smile: Allows target-farming.
But not the topic. I know you dislike PoE because you don’t understand the mechanics in-depth.

In practical terms… this argument is wrong though.

PoE for example adjusts the drop-rates of uniques around their top-tier uniques, which are Mageblood or Headhunter. They stand in relation to each other.
The balancing is done by enforcing ‘a Mageblood drops every x hours at y content’ and that’s the anchor (yep, back there) for their balance.

Hence the argument is outright bullshit from your side since no such anchor can exist in the LE system. Since items don’t stand in relation to each other that means you can’t hinge it on a specific item. ‘Wraithlord Arbor’ doesn’t drop 10 times less then ‘Red ring’ for example, which would be a static connection. Instead it changes.

So in PoE a SSF character drops the same amount of every unique during content no matter how many they add as their drop-rates are designed by ‘drops x amount as often as y’ and total drop-rate is designed by adjusting the quantity to that anchor item.

You don’t have that in LE.
You failed consistently to answer ‘what value is the whole equation anchored to’ as well as ‘how to you guarantee that they keep standing in relation to each other?’ and instead trying to dismantle that it doesn’t happen in what I presented.

I presented concepts, you said ‘but if you do this then!’ and I provided answers. Some of those needed to go out of the scope of the system. Then you found fault in it, I went a step further. Now you say I’m moving goalposts.

It would be fun if it weren’t sad.
And with that I’m out of this argument because it leads to nothing. I provided factual proof, you denied it and said ‘but in the grand scheme of things’, I provided factual proof that the current system is extremely close to the example and now you ignore it.

Yes, it’s worthless to argue like this, you’re absolutely right.

74

No, the “optimal system” you provide also has non-optimal outcomes, you just screw over some people a lot more in detriment of others rather than everyone slightly but equally.

I don’t dislike PoE, I just got burned out by it and all its systems being targetted for the top 1%. And when they make a system that targets the 99% like Harvest, they nerf the crap out of it after being forced to keep it in against their will.

That’s just wrong. There is no difference between both systems for this. You make mageblood and headhunter remain at weight 1, or whichever, you make red ring and ravenous void remain at 98 or 99% reroll chance. The effects are the same and you still have your anchor.

You just don’t want to see that both systems can do the same thing with just slight variations in result. And those variations are what justify using one or the other knowing that the results are similar and that the small differences are what you’re trying to accomplish.

This is completely useless for a SSF character, though. They don’t want “a”, they want “the”. And PoE balances everything towards “a” because to GGG only trade matters.
With LE’s system, you get a balance where “a” gets slightly harder but “the” also only gets slightly harder instead of becoming twice as hard.

The main difference between both systems is that one only cares and favours a part of the playerbase while the other tries to even things out.

You didn’t though. You showed an example that has small changes. You just want a system that benefits MG over CoF, rather than one that is equal for both.

And if you want to talk about ignoring, you also ignored the part where I said that your system leads to the point where “the” common is rarer than “the” rare, because of the overabundance of commons vs rares. And, using your own arguments from before, how do you explain to a player that “Oh yeah, unique x is common, but you’ll get y rare before you get it”.

1 Like
75

See?
That’s what I’m talking about that you’re simply ignoring stuff.

I said ‘this system is for this situation’
Then I provided ‘to adjust it to the other situation you have system x on top which causes it to shift into the other system and hence removes the problem’.

So once more… arguing from you for arguing’s sake.

It’s like you’re hyper-focused on a single piece and have no clue what to do with it afterwards.
You can’t look behind your horizon despite being presented with a universe of options beyond.

Yes, and now you talk about things in the game which aren’t holding true, which is the issue with that.

I’m agreeing fully that the reductionist method of implementing content is extremely negative, I hate it as well. But in the ‘grand whole’ the game is solid.

And the argument ‘all’s done for the 1%’ is nonsensical. You’re just - once again - hyper-focused.
Mass melee rework is not focused on those.
The scarab changes so sextants and high level high-friction player agency is removed and instead simply leaving ‘player agency’ is making things available with T1 maps overall.
The current league is very much targeted at every level of player and allows to either focus on a specific sub-set of function to get returns or a grand whole for more returns.

There’s a shit-ton done for the 70% (which reach maps) and not the 1% (which is pinnacle content).
Failing to see that is a mistake on your side. You’re asking me not to speak ill of LE… I don’t understand why, they’re doing many things vastly worse then PoE… and PoE does other things worse then LE. And I’ll gladly call out each aspect of each game individually.

Outcome different = system different.

All needed to be said.
Everything else is a delusion.

What I wrote.

If you drop a Mageblood every 1000 hours of play-time then you’ll drop a ‘Voll’s Protector’ every hour.

Both are ‘the’.
They reduce the ‘a’ number, which is why prices for uniques are generally low outside of high demand and specialized ones.

Exactly what I’m saying, you just don’t understand my words it seems. Repeatedly.

We have a example for 1 common, 1 uncommon, 1 rare.

I provided the situation with the bows:

So here you go.
Once again ignoring what was written.

I didn’t. You made the argument and I expanded on provided a alternative solution (which is why we have 2 systems instead of 1 now) to solve that, which has other issues.

Then since you argued it screws over one side (the current does for both, differently as well) I said that one is a reduction of the other actually and allows to add the adjustments on top to transform it after without affecting the initial state.

That’s the situation again.

And once again, anchor ignored.
Once again, how the current system creates a relation to time ignored.

You’re really good in avoiding the direct questions and talking around :slight_smile: Might be because there’s no answer I guess.
Harsh if you wanna dismantle an argument which is premised on a time-based constant while your counter has no constants.

76

I never said PoE wasn’t a good game. Just that it doesn’t (or didn’t) treat its whole playerbase the same.
All those changes you mentioned happened only lately after I gave up on it. Until then PoE was a game fully focused on the top 1% and on traders.

My whole point is that there is only a slightly different outcome. It’s obvious the system is different. Just not that different to justify all of this discussion.

That is because you keep shifting the discussion. When we started discussing “a” and “the”, your whole premise was that no matter what they add, I always get “a” in the same timeframe. Which was the important thing.
Now apparently it’s “the”.

I didn’t ignore it. I just fail to see what your point actually is. Yes, you get a shift, but it’s a slight shift, which is all I’ve been arguing.
Adding new items will always generate a shift, no matter the method you use.
You just decide which part you want to shift more (using PoE’s method for favoring “a” or “the”) or you allow the shift to be uniform across the board for everyone. Which, in my opinion, is the fairer solution.

Yes, you shifted the discussion again.
Simple weight tables are similar, so now you add a tier system which PoE uses.
It has problems, so now you theorize another system which nobody uses. Which itself already goes against your very first premise which is that there is no need to make a new system because the existing one works perfectly.

When did I ignore that?

Might say the same thing. After all, I even went to the effort of creating a sheet with actual formulas to show deviations for all systems.

Bottom line is that you can’t make a system that doesn’t change when you add new uniques. Something always has to change. It’s inevitable. And the changes presented in all methods are similar.
And yes, before you say it again, they’re not the same. But they’re negligible as well. Especially when you start taking into account real numbers with hundreds of items.

77

@DJSamhein fell right into his trap… hahaha

@Kulze don’t get me wrong, I like your posts, but sometimes they’re too much to digest… I always try to follow, but most of the time you lose me halfway… You should be in court as an attorney or something :sweat_smile:

Anyways, this is indeed a giant moving goalposts situation… This thread is about quantity of valuable nodes on a Timeline, and you both got all the way through here, arguing about drop tables, weight and anchor systems.
I even thought this was the one about the guy not dropping Omnis, based on what’s the current discussion LOL

1 Like
78

Yeah, I know. I suck at explaining. :stuck_out_tongue:

And yeah, the goalposts have massively moved, which I won’t deny. But each of them is a underlying thing.

If I remember right it went from farming methods and their viability with factions to overall how uniques are dropped, which led to the mechanic behind it… which led to me taking affront at how it’s done because I saw an issue in that.

The initial topic is basically done I would say.

Yeah, the situation has changed. Which is likely to a big part because of LE, GGG looooves to take stuff from other games and implement the realizations in their game to enhance it.

The other way around would be quite nice to see as well. Very few things in game design are actual ‘new’ things, it’s 99% re-iterating stuff from other areas and optimizing the recipe after all.

I mean, if we have that system we’ve seen a halving of a specific part in total drops by adding a single item to it, and adjusting all others accordingly at the same time downwards, also in relation to how many are in there.

That’s what I’m taking affront at.
If one item to 2 items causes more then 50% rate-change for the same rarity then I’ll either see more or less of this area. No matter how high the difference is.

And in our example the difference was high because we had few items.

But in the bow example the item count is also very low.

And in total the item count is also ‘not high’.

It’ll fix itself over time - to a degree - but I see an issue with balancing there, since you can’t do it.
I’ve solely provided 2 methods which do provide a balancing baseline, which makes them immediately superior. As a developer you wanna have control over what your stuff does after all. Side-effects which come out are a different thing… but if you adjust something you wanna know at least what it changes and by how much before you do it. Saves time, saves effort, causes less mistakes.

In comparison the situation we have is a in built disaster-machine hence.

True, that’s my bad. My argumentation line wasn’t fitting with the example.

I’m arguing it’s not ‘slight’.
Slight in value… yes.
But as mentioned, if my 0,8% drop-chance changes to a 0,72% chance then that means instead of 100 hours I need 110 suddenly.
Which is not ‘slight’.

Umh… define ‘shift’.
Will I see more of a specific item? Not necessarily.
Will I see more items in total? Not necessarily.
Which direction are we talking about? I can keep every individual item dropping at the same amount in the same timeframe (through quantity) or I can make sure that I get the same amount of items of at least the same category. And I can adjust one system to the other even.
My main argument is that it at least has a starting point where you can say ‘Yes, I’ll get it then!’ but the system we have does not.
So it makes it hard to handle.

Better, nothing’s perfect.

You never answered me despite several times asking!
What is the anchor of the system we have in LE?

That’s all I wanna know with that! If I have the info I can change my whole stance as I can adjust the concept in my mind… but I can*t find an anchor. So provide me one if there is one! Then I can likely agree to somehow make the whole thing work, without a starting point though I can’t decide on the end-point!

79

But this will always happen no matter which method you use. With your method, either “a” or “the” will require more hours. You can’t keep both static while adding more stuff. One of them has to change always.
LE’s is just more “democratic” in that it changes everything equally.

You need 10h for “a”. You need 50h for “the”. You introduce a new item, one of those values has to change. You can’t keep both equal. You either change to 11h for “a”, to 60h for “the” or to 10.5h for “a” and 55h for “the”.

You can do that in both systems. Quantity is an external multiplier that doesn’t change. And it doesn’t linearly affect all items. It just normalizes the curve.
Even with increased quantity, you can’t make the drop time for every single “the” unique because their individual drop chance has changed relatively to the rest.

If you reduce it to the simplest state, you get 1 unique (weight 10 or reroll 0) that requires 10h to farm and another (weight 1 or reroll 90) that requires 100h, for a total of 10h to get “a” common and 100h to get “the”.
Adding another unique will always change some of those values. You can’t add a new one still require 10h/100h for both “a” and “the”. That’s true for every method. If you add more stuff, stuff requires longer. All you can do (with your method) is making only one or the other change and try to normalize that via external factors (quantity or divinations/CoF), but the system requirements themselves, at its core, will always have to change. There’s no getting around that.

All we’re discussing from the start is the way in which they differ, which, for small samples is “a lot” but for large samples is “not that much”.

It still does. When will I get a red ring? After 1000 unique rings drop. When will I get a Mageblood? After 1000 uniques from tier 1 drop (obviously made up, I don’t know the real chances and didn’t bother look up).

The biggest thing you can say they differ from is that one you can add and easily infer how that will change the overall system and the other you can only approximately calculate that beforehand.
But once both systems have had their item added, they’re still easily calculable.

I answered that both times. Just like you have Mageblood/headhunter with weight 1, you have red ring/ravenous void with reroll 98. How is that different? Your anchor is simply your rarest item in either method.
When you add a new item, you ask yourself “How often do I want this item to drop in regard to a Mageblood/Red ring?” and balance it accordingly.

80

Yes, nonetheless you can decide which it does.
And then you can adjust other values accordingly to make it a nice experience.

But you can’t with the current one in my mind and that’s the whole ‘anchor’ topic.

Any change feels bad for a large slew of people when it shouldn’t for as many as possible. It’s not discerning properly between the wanted and existing state.

Oh yeah, the shift is a necessity. You can only keep 1 value static.
But isn’t that all that’s needed?

If I keep the acquisition rate for rare items the same in MG then the same amount of rares will reach the market. Less of each of them… sure… but the demand is also spread out between them.

And if I make sure that CoF gets each specific item at the same rate then it makes sure they get something to progress in the same timeframe as well.

I really don’t see the issue there. It’s 2 separate factions, they should handle the respective loot different if one is ‘one for all’ and the other is ‘all for one’ in playstyle already. I’m just saying provide each with the method that pleases them rather then a method which causes issues in both.

If you add a unique ring it’s not 1000 anymore though? The percentile changes.

I can’t even say ‘I’ll get any extremely rare ring every 1000 unique drops’.

If you add one of the type it causes the red ring to change for example to every 1800 and any sort of rare to 900.
None of both values is fixed. Neither is 1000 anymore.
I have a different amount of items needed and I also have a different timeframe for the whole category needed.

In my system one of the values stays ‘1000’. Since I want my players to get either that specific item or one of the type every 1000 hours. Whichever fits for my game.

See example above.
I can’t see that ‘1000’ anymore afterwards. Anywhere.
That’s the value I’m talking about, one which is absolutely fixed.

PoE for example has a fixed rate for ‘some’ content where it’s balanced from. I can’t see how that would be done here simply.

I don’t ask ‘how often do I want it to drop in relation to’ but I ask ‘how often does a player actually see it’.

81

Not really. I’d much rather everything shifts slightly than one part doesn’t shift and the other shifts wildly. Which is why I said it was the “fairer” solution.

Well, but that is the issue. You have a fixed rate for “some” content at the expense of bigger fluctuations for the rest.
With LE, you have smaller fluctuations for everything.

At that point, it’s just a matter of prerefence.

You can only keep some of the uniques in that state, though. In this case, you make sure Headhunter/Mageblood remain static in time required, but other uniques will take longer. There’s no getting around that.

Again, it’s just preference. I like LE’s system (or the simple weight table we were using at the start) better because it’s more democratic. Everything gets affected in a similar way.
And if you want to keep red ring at 1000 rings, you just slightly adjust the some of the other rings reroll chances until it goes back to 1000.

But if you don’t adjust it, red ring will require more than 1000 but you will still get 1 red ring for every ~x sunwreaths that drop, because sunwreath is now also a little rarer. It maintains relative rarity between the uniques similar.