LP System has lots of issues | Constructive Feedback and Suggestions

Feedback and Suggestions
21

Exactly, and that’s the line I’ll pick up and construct the argument about.

So, the itemization has a massive variance through the implemented drop-rate that leads to a ‘not good’ outcome.
Hence the variance itself there is the problem.

In this case the variance of a crafting system is solely confined within the rolls, which by themselves have a small variance. You got 4 mods to change and 5 tiers, it goes up, no changes in base types and sub-types.

That alone makes the distribution different.

With the vast normal distribution provided from dropped items it means you can’t see the steps at the edge-cases anymore, which in crafting when you solely affect a single aspect per time is not as much of an issue. It’s after all solely ‘crafting steps possible’.
With item drops though it’s ‘item type, sub-type, 4 times affix roll, existence of affix roll itself, limit breaking of basic roll range per affix’ which is a total of 7 conjunctive rolls on the normal distribution.

This causes outcomes to be extremely centered, the outliers are very very rare, and high quality items are hence only becoming visible at what? 4 deviations? 5? That’s massive. And then the higher you go the deviation range gets exponentially more, 7 deviations from standard is basically the highest we’ve seen for human behavior, but items are closed behind 10+ deviations.

The lower the deviation rate the higher the still potentially existing outcome.
Yes, your item will brick… which feels bad… but it won’t be useful without the affixes anyway, that behavior was solely a problem within the normal deviation ranges we see, not for the others, it enabled the others. Short-term detrimental, long-term positive.
Hence with the new system the short-term is far more positive but the long-term… is a mess.

That signs towards a problem with the setup itself. You can’t shift the ranges as it causes a breakdown in the whole setup itself. It’s the methodology to get to the outcome which is the issue.

I’ll once more provide Path of Exile as a good example. They created those ‘safe steps’ in crafting as a counter-measure to normal distribution deviation. The outliers became so impossible that they needed a hard-reset state to then cause another new normal distribution situation.
This way the rate of deviations needs to be less and can each individually be adjusted.

Last Epoch has no means to handle normal distribution in-built none, zero. That’s the problem. That’s the easiest possible fix. That makes it similar to PoE though for crafting as it’s the only option viable and simple at once.
Everything else enforces a fundamental change in how the systems to reach your goal are created from the ground up.

22

I’m afraid I’m not into pegging (no kink shaming, though). :laughing:

Joke aside, yeah, the main reason I like ARPGs is the combat aspect linked with the drops aspect. Crafting is something that, to me, only dilutes that because then the game is balanced around both drops and crafting.
Both in LE and PoE you’re expected to craft to have successful items (baring buying them ready, but still someone crafted them). And the drops are adjusted accordingly to this expectation.

Personally I’d rather just enjoy my loop of kill stuff/pick up stuff. But I can tolerate LE’s crafting for being transparent, simple to use and letting me improve the items with little hassle.

This is also why I never really had BiS gear in any game other than D2 because it’s a purely drop-centric game.
In PoE I could get around this via trade, but trade is very toxic (in both games) and only leads to burnout for me.

1 Like
23

I kinda like the idea of having some unlock that just lets you craft what you want. Gate it behind some tough challenge with the idea that your first character of the season has to be able to build up and overcome that threshold, but then after that if you have any interest still left in the game, it’s probably trying new builds, which would be a lot smoother experience if you had something like that. It would also let me clear out my nightmare hoard of a stash in Legacy.

Some kind of deterministic system for CoF so we at least have the light at the end of the tunnel would also be nice. This is I think the biggest thing holding the game back for me aside from bugs.

24

That’s fair, but Kulze saying he likes the extreme varibility in the old system but doesn’t like the less extreme variability in the new system which has fewer negative end states (& the main one being that a rune of removal doesn’t remove the affix you want it to) feels inconsistent at best & hypocritical at worst.

But you never could before. So why is it bad now?

No, it’s not (entirely) “factually wrong”, the previous system had a fail state of reducing an affix by one or more tiers. That would make the item objectively worse than before you pressed the button. That state is not available in the current version of crafting.

25

Is it?
Do you realize the sheer difference in scale there?
I have nothing against RNG, but if RNG is used then accordingly to the right probability-rate for whatever you wanna do. Which seemingly nobody seems to get.

In the crafting system we basically have 2 states… win/lose and sooner or later you loose. Ok. Makes it easy, we get a rough chance for the ‘lose’ condition and then see how many steps we need to get our stuff done.
Usually the range was doable, just felt crap during earlier progression when your gear lost tiers, has no meaning after getting 4T5 since then you either get an upgrade or not… no matter how many tiers it looses, that condition was not relevant by that time… but removal cost us the chance to get top-tier items with more steps to a large degree.

With the new one this is a outright reduction based mechanic which has a limited top-end, you can’t ever get ‘theoretical endless tries’

The gear-drop system is one which starts off with 32 bases. And we have to add that one on top of the crafting one
Helmets gave (potentially) 2 suffix slots filled with a variety of 188 suffixes, that alone creates 17578 states through mathing it out with factorials (Actually less but I don’t wanna write a bazillion lines long equation and explain it too).
Then 2 prefix slots with 20 suffixes which causes 190 states (and for ease we’ll just ‘add them’ instead of using the factorial for the chance of both existing together in varied combinations + empty affix slots)
Then we take the factorial for the different tiers each slot can have (0-7) for 4 slots, which gives us 70 states, together with our 17578+190 = 17768 helmet states for different helmet base/affix combinations we alone through that 1243760 different outcomes for helmets alone.
And that’s not even taking into consideration the weight-distribution for empty affix slots and the weight-distribution decreasing for higher quality items as otherwise 4 T7 items would be as common as a white one! Or any other base with any other combination that could theoretically drop, all together leading to a normal distribution bell-curve when taking the varied states into consideration.

If you can’t see why I dislike a system which enforces you to roll your dice 1243760 times in a row for a fraction of possibilities in a row to get your result versus one where I have to roll for what… at the top most for 100 times in a row while even having more then a 1 in 2 chance for it to happen is a system I agree with more then the basis for any discussion is gone anyway.

RNG is fine, but a change of 1 out of a quattuordecillion, googol, googolplex or any other weird name for such massive numbers happens is beyond sanity.

Crafting chances on their own are at least within sanity range.

It was.
I just didn’t realize the existence before.
And when the cat’s out of the bag you can’t put it back easily.

In a crafting system where it’s 95% of the time a binary-state outcome then in the range of those 95% there is no ‘objectively worse’ item for the user. It’s either a upgrade or a fail. There is no ‘I failed a bit more’.
Those 5% are not the normative cases, you extremely rarely need a similar item upgraded to the top outside of campaign ranges.

1 Like
26

I think you’re not addressing what is being discussed, though.
We had a long discussion with you saying that we should eliminate the outliers because variance is bad and makes the players feel bad.
We even came up with the numbers 1-24 for the current state and you wanting to change it to 5-24 to reduce that variance.

But now you say you liked the previous system better when that system ranges from -24 to +24. That is not consistent with your previous position.

I’m not getting into the pros and cons of variance, we already did on the other thread and I don’t want to rehash it. Just that you’re very against variance but now prefer a system that had higher variance than the current one.

1 Like
27

Yes, and in the item drop system because of the ridiculously huge amount of possible rolls that holds true.

Why? Since it’s a multiplier for every other mechanic building on top.

Because the crafting system is another topic then the item drop system. They’re separate.

You can’t apply the same principles to a base system compared to a auxiliary system trying to reinforce a base system.

Also if we go with your example:

The change from old to new system didn’t do that.

We had a range from -5 to 24 as an example. What EHG did is 100% in line with the issues I presented with the full range for item drops even. Reducing range of a normal distribution function is not reasonable, but exactly that was done.
Instead of having -5 to 24 we now have 0 to 17.
What should’ve been done was 0 to 24.
The negative numbers were the ‘it feels crap’ situation, nothing else.
The issue with the full-scale system when taken into consideration is that the range itself built up on the RNG ranges below were already utterly wrongly set. I would even argue that the core idea of the system doesn’t align with how item drops are handled chance-wise at all, neither in the old or new system. You just at least ‘had a realistic chance’ in the old one, the new one not.

Since crafting has far less varied outcomes though compared to the drop system you saw the upper-end a lot easier though. Hence EHG reduced the top range too to allow a specific range of items (likely 4 T5) to have the same chance to be crafted, automatically negatively affecting the possibilities for more complex crafts to function by design.

We also wouldn’t be talking about crafting at all if there wasn’t a underlying issue with the drop system itself. Because at the end of the day it comes down to ‘How often can I pull the lever for it?’ and hence ‘How quickly will I get a relevant outcome?’.
Since finding a good base to even start intending to craft is by itself like a miracle sent by a creator people might or might not believe in obviously the times this lever is pulled are very low, hence positive outcomes beyond 4 T5 become exceedingly more and more rare until they’re a pipe-dream.

1 Like
28

Yes, it’s not as big as you purport it to be. It’s your prerogative to be inconsistent, as much as it is my prerogative for that to leave a bad taste in my mouth.

Using that logic, the chance for a 4lp Ravenous Void is 50% since it either drops or it doesn’t, that’s 50:50…

29

So… your chance to drop an item is not much different to instead craft a dropped item into a finished product?

Is that really what you wanna say?

Yes! Yes it absolutely is! :slight_smile: That’s how chance works.
Now repeat that 50/50 roll several times in a row and the same outcome has to happen a specific amount of times in a row without fail. That’s how probability is set up, from the basis. All probability world-wide.
All probability starts with a singular ‘Existence: Yes/no’ position and from there you expand it to the number of cases before expanding it on the nest thing affecting it… and so on until you get the respective actual percentile.

Our ‘perception’ state as a player is ‘win/loose’ which is derived state from the underlying probability related to the actual percentile chance this outcome actually has.
A player doesn’t enjoy to ‘loose’, games are there to make you ‘win’ or overcome a ‘loose’ condition through some means (be it effort or skill, doesn’t matter).
If the ‘win’ condition is stacked against you then you stop, it’s simple. So if your ‘win’ condition after taking every step into account goes below a specific amount then it’s over, period.

A coin-flip is a 50/50 probability (and not even that, it can land on the edge).
A lottery in my country is 6 out of 45, a binominal coefficient hence. Which relates to 8.145.060 times you need to do it to actually ‘win’.

The difference between a lottery and the game is… in the game 95% of the time your perceived ‘win’ is solely the jackpot, in a lottery you get partial payouts below.
Imagine a lottery where only 1 person out of 8,145.060 people wins per draw. Do you think that would be something anyone but the hardest of hardcore gamblers would even take part in?
That’s why lotteries have several ‘win’ conditions below the actual top outcome.

Obviously in crafting it’s less then that. So let’s instead look at it directly what our chance is:
We can say a common craft takes between 1-16 FP baseline (cheaper ones available but more expensive ones too). This means medium outcome is 7,5 FP. With glyphs of hope this reduces to 5,625. We also have a chance to get a critical success without FP used and doubling out outcome. The chance for that? Not sure. But it’s somewhere around 10%, which means a multiplier of 0,9 which provices 2 rolls at once hence it’s a 0,8 multiplier for ease of usage.
Means 4,5 FP per roll.

That means in a 45 FP (A really good FP count) item we get commonly 10 rolls. That’s median amount.
That means our chance of getting a 10 tier item to 20 tier is 50%.
That is solely if all affixes are the right ones.
Glyphs of chaos are there to get the right one, which is another reduction in chance to finish a item, those are more then well used. A single use already reduces median outcome though since the hope modifier is gone, hence you can’t roll 10 times anymore, you’re at least limited to median 9 rolls, so already 11 tiers needed instead of 10.
Then if you need to use a removal for a T5 on an item which is unfitting causes you to loose the critical modifier as well, it can’t crit. Also it can outright destroy your item with a 25% chance (if you got a T7 on it). A single rune of removal reduces your immediate potential success to 12,5% (50% of 25%) success to achieve the outcome you want if we wanna hit it with the first time (unless you wanna reduce your chance further.)

That takes into consideration that your exalted item is at least 45 FP and tier 16 to begin with, then you have a 12,5% chance to do it. And that doesn’t even take into account the lost FP through the Rune of Removal in itself.
That’s a 1 in 8 positive outcome. You need 8 bases already for that.

Worse if 2 affixes are not ‘right’. With 3? Don’t even need to try, your chance is vastly below 5% win-chance. Who can get 20 ‘+4 to mana strike’ chests on the right base per cycle? Neither CoF nor MG does, they’re that expensive and sought after.

And therein lies the whole topic.

1 Like
30

No, that’s not particularly close to what I said. Since we’re talking about crafting we get to ignore the drop chance because if you’re crafting on an item then by definition it’s already dropped. What you’re talking about (& moving goalposts/strawmanning) is the sunk cost of how likely the item is to have dropped in the first place.

Cool, so I should be seeing 4LP Ravenous Voids raining from the sky, because as you’ve just confirmed, the drop chance for that is 50:50 & not something longer than the age of the universe.

31

This is obviously hyperbole and doesn’t help your argument. It’s somewhere between the lifetime of planet earth and the lifetime of our solar system. Using the lifetime of the universe is just ridiculously absurd, so get your astronomical figures properly correct, please.

32

Actually it’s not. When one of the devs was asked about it some years ago, they gave a number that was pretty big. I don’t remember what it was & I want to say it was something like 10^18, but I could just be making that up.

The age of the earth (4.5x10^9) isn’t that far off the age of the sun (4.6x10^9). The age of the universe in seconds is ~4.4x10^17, so I don’t think my memory is that far off with 10^18.

I’m at work so I can’t check discord.

My astronomical units are just fine thanks.

Edit: this was obviously before they reduced the drop rates for non-CoF.

33

When I did the math for a 4LP red ring, I got that you get one red ring approx every 1000 ring drops and the chance for 4LP was ~ 3e-10%. So a total value of ~ 3e-13%. Which is several magnitudes lower than your number.

Maybe Ravenous Void is rarer than a 4LP red ring. Or maybe I fugded the math on the red ring.

34

Well, I have no clue on how to calculate this, but according to MaxRoll’s planner, the chances of dropping a Red Ring in both echo rewards or rune of ascendance is 1 in 733, and the odds of it having 4LP is 1 in 332B :rofl: :rofl: :rofl:

35

That would be 3e-12, which is 2 magnitudes away from LETools chance which lists as 3e-10. Given both sites track record, I’m way more inclined to trust LETools. So that would make it 1 in 3B instead.

36

Ravenous Void has a level for lp of 105, Red Ring has “only” 100. We don’t know how that converts into actual chance though. The gloves aren’t rarer than the ring, both have a 98% reroll rate.

37

LETools says Void rolling 4LP is 1e-11% chance. So not that far off from red ring. Assuming you could get 1 in 1000 drops as well, it would still make it 1e-14%, so several magnitudes below e-18%.
So somewhere between the full lifetime of earth (until death) and the full lifetime of the solar system (until death).

But, as I said, my numbers (or LETools’) could be off.

38

omg, now this is getting stupid… for 4LP Ravenous, MaxRoll says it’s 1 in 9.6T. :rofl:

@DJSamhein I really don’t know how to translate the chances in % to the odds in 1:X when numbers start to go really big like this. I don’t even know what that 1e-% mean lol…

So I always used MaxRoll for that, since I assumed it was just that, direct translation for the % chances on LETools but shown as odds.
It would be a shame if the discrepancies are really that large on both sites :sweat_smile:

39

I’m curious as to what you think that looks like/how you’d define it. Even if the sun went supernova (it won’t) there would still be a remnant left over. And since the sun is ~99.8% of the mass of the solar system, we don’t really care what happens to the roundings).

Either the devs got the initial chance wring, they’ve massively buffed the drop chance or LE tools is wrong.

40

Drop chance conversation (I’m focused on the chance of an exalted dropping with all the affixes that I care about, maybe a T6 and a T7)

A cycle lasts three-ish months. I don’t know what the average cycle player’s cycle-time is, but I assume it’s something less than a month and a half.

As someone that likes cycles and is at a loss as to why I’d want to play legacy I get only a sense of annoyance for items that I have less than 1:1000 chance of seeing in the time I’ll play in cycle.

Does this game need to cater to me? No, I’m just another data point.

Heat death of the universe versus “only” the age of this solar system conversation.

I find it embarrassing in a kind of, “sigh. this is my team?”, way. So. For real? You three are arguing about whether a drop won’t realistically happen during the life of this planet, the sol star, or the current age of the universe and think the difference between those things matters at all to the span of time you’ll exist on this planet, let alone playing this game?

This is a joke, right?

3 Likes