Lucky Looter

[Axonius]

I was unable to find a post on this. Something seems wrong with Lucky Looter.  It almost seems like it works in reverse. I watch a number of youtubers and with same/similar settings but LESS points in LL, I've seen these folks get far better drops than I and more often.

 

I have 4 points in LL right now AND have the goggles with 9 points. At day 45 I've found one item that is level 6. 

 

Watching others who don't have as many points, they consistently find level 6 items and aren't even as far along in days as I am. Also, I seem to find a lot more totally empty nodes than they do.

 

Anyone else getting a weird feel from LL?

[meganoth]

Moved this to general discussions

 

This is called confirmation bias: https://en.wikipedia.org/wiki/Confirmation_bias

 

Finding lvl6 items has a very low chance. LL and the goggles add some bonus. So lets say your chance is doubled because of LL. That sounds great if your initial chance is 25%. But if it is only 0.01%, it would be  0.02% now, not that much of a difference.

 

You also don't know if your youtubers increase the loot chances (maybe even lie about it) to make their streams more interesting, or snip away the boring parts.

 

[Gazz]

In A18 it's all extremely random.

 

"I think other people find better stuff than I do" is not a lot of data to confirm anything.

[12pack]

1 hour ago, meganoth said:

You also don't know if your youtubers increase the loot chances (maybe even lie about it) to make their streams more interesting, or snip away the boring parts.

i would add, even, some may "stuff" chests, also... further skewing your perception of their results..

[Ti2xGr]

I was wanting to come up with a way to force certain items (like specific level 6 loot) to only be found/unlocked after level <insert random level or gamestage here>.  Anyone know if this is doable? Can we attach to a minimum level?

[Gazz]

That is doable. Look at how quest loot works.

You can write quality and probability templates based on the gamestage and attach it to loot groups.

You may need painkillers for the headache but it's certainly doable.

 

[Ti2xGr]

9 minutes ago, Gazz said:

That is doable. Look at how quest loot works.

You can write quality and probability templates based on the gamestage and attach it to loot groups.

You may need painkillers for the headache but it's certainly doable.

 

Excellent. I'm getting more and more comfortable with these XML edits every day. Thanks for answer! I've got a project for the weekend...

[Boidster]

1 hour ago, Gazz said:

You may need painkillers for the headache but it's certainly doable.

I am currently in the hospital undergoing treatment for the brain aneurysm I got while trying to make sense of loot.xml.

[Gazz]

It's not hard if you're okay with juggling recursion and 2 layers of templates bending the results.

[FileMachete]

Slightly related is something I saw last night.

 

I was wondering if crafting a lvl 4 (grn) SMG was worthwhile, damage wise compared to the lvl 3 I'd looted. Having no idea if the lvl 3 was 'good' or 'bad'.

So I enabled creative item menu and entered SM then added G, highlighted the cell the SMG appeared in and kep hitting Backspace then G, a good few hundred times. Cause the results were weird, not me, I ain't a werido... really.

 

So what caught my attention at first was that lvl 4s damage was only coming in as either 37 or 38. Nothing else.

Which wouldn't have led me down the rabbit hole until I noticed that lvl 3's were mostly _higher_ damage than lvl 4's. Lvl 3's were commonly 42 or even 43.

 

So I didn't have high hopes of crafting a lvl 4 SMG that would be much better than the looted lvl 3 with it's 36 damage. But went ahead and made one.

Turned out with 44 damage, and a few rounds larger magazine, but a bit less range, so on balance a solid upgrade.

 

Guess the point might be that, "Here be Monsters".

From what I've read/learned in a lot of years in IT, no current computer can ever do _true_ randomness without outside influences.

So there may just be a speck of truth to it when folks say, "this game hates me, I never get good loot!"... :-)

[Jugginator]

It is called lucky looter, not guaranteed improved looter

 

Rng be rng. 

2 hours ago, FileMachete said:

 

From what I've read/learned in a lot of years in IT, no current computer can ever do _true_ randomness without outside influences.

 

Yup lol, computers would be sitting there eating tiny tidbits of electric if we didn't tell it to do stuff with it

[beHypE]

15 hours ago, Gazz said:

In A18 it's all extremely random.

 

"I think other people find better stuff than I do" is not a lot of data to confirm anything.

Do you mean A19 fixes that randomness somewhat ?

[Gazz]

Yes, because finding that auger on day 3 really ruins your progression for that game.

[Shelgeyr]

39 minutes ago, Gazz said:

Yes, because finding that auger on day 3 really ruins your progression for that game.

Purely for the potential shared entertainment value - I'm not sure if you're being serious or sarcastic and can't find it in myself to disagree with you either way

[meganoth]

10 hours ago, FileMachete said:

Slightly related is something I saw last night.

 

I was wondering if crafting a lvl 4 (grn) SMG was worthwhile, damage wise compared to the lvl 3 I'd looted. Having no idea if the lvl 3 was 'good' or 'bad'.

So I enabled creative item menu and entered SM then added G, highlighted the cell the SMG appeared in and kep hitting Backspace then G, a good few hundred times. Cause the results were weird, not me, I ain't a werido... really.

 

So what caught my attention at first was that lvl 4s damage was only coming in as either 37 or 38. Nothing else.

Which wouldn't have led me down the rabbit hole until I noticed that lvl 3's were mostly _higher_ damage than lvl 4's. Lvl 3's were commonly 42 or even 43.

 

So I didn't have high hopes of crafting a lvl 4 SMG that would be much better than the looted lvl 3 with it's 36 damage. But went ahead and made one.

Turned out with 44 damage, and a few rounds larger magazine, but a bit less range, so on balance a solid upgrade.

 

Guess the point might be that, "Here be Monsters".

From what I've read/learned in a lot of years in IT, no current computer can ever do _true_ randomness without outside influences.

So there may just be a speck of truth to it when folks say, "this game hates me, I never get good loot!"... 🙂

 

Tried this myself and got exactly the same values in creative menue as you did.

 

I had thought CM weapons simply had the average value without the random modifier, but somehow this is not the case. It looks as if some dev just created a sample of every weapon at every quality in-game and hardcoded each weapon he got into CM. Resulting in fixed CM weapons with mostly average values and a few outliers. This would not explain the 1 point damage fluctuation you were seeing though.

 

I also built a few SMGs in the game and as expected q4 was better on average than q3

[Darthjake]

Started a new game a while back, put my first point into lucky looter, and then promptly found the goggles.  About an hour in was rummaging around and found a level 5 blunderbuss.  I prefer stealth, but 7D2D needs a wall mount for old weapons that you don't want to get rid of.  That blunderbuss saved me from 2 bears.  I will never get rid of that gun as long as the game is running. 

 

Or until I restart again, but that's another story.

[Axonius]

Good feedback.  Has anyone been able to drill in to see how Lucky Looter points translate to actual delta in percentage?  Maybe it's not even worth the perk points. 

[meganoth]

Not that easy to calculate. There is a different way though. LL provides a bonus to gamestage and gamestage is constantly increasing when you play.

 

Write down your current gamestage every day at the same time (for example at nightfall). From that you can calculate an average increase per day. Now lets say you increase gamestage by 5 every day and LL3 gives you 30 gamestage more than LL0. That would mean that every time you open a box with LL3 you get the loot you would get (30/5=) 6 days in the future with LL0.

 

If that doesn't sound enticing, consider that you also get a bonus to loot speed. And that is the main reason I usually take LL. 

 

 

 

[jdifran]

8 hours ago, Gazz said:

Yes, because finding that auger on day 3 really ruins your progression for that game.

In my most recent playthrough, I found a cement mixer schematic on day one and promptly scrapped it, because having cement that early would just ruin the early survival aspect of the game for me.  I would personally like to see the workbench, cement mixer, and chem station schematics removed from low level loot and have the chance to find one in a destroyed workstation more like 5% instead of whatever it is now (30%?).

[Boidster]

17 hours ago, Gazz said:

It's not hard if you're okay with juggling recursion and 2 layers of templates bending the results.

Boidster: Doctor, it hurts when I do this <solves polynomial equation>

Doctor: Well, don't do that.

 

5 hours ago, meganoth said:

I had thought CM weapons simply had the average value without the random modifier, but somehow this is not the case.

Gazz (I think) has said previously not to use CM to test out the RNG system. If I remember correctly, a few months back I used 'giveself' and it also showed unexpected results especially if quantity>1. The only way I got results consistent with the expected ranges was to load up on mats and run 20 workstations crafting weapons.

 

15 hours ago, FileMachete said:

So I didn't have high hopes of crafting a lvl 4 SMG that would be much better than the looted lvl 3 with it's 36 damage. But went ahead and made one.

Turned out with 44 damage, and a few rounds larger magazine, but a bit less range, so on balance a solid upgrade.

Here's part of the XML (items.xml) which governs SMG attributes:

<passive_effect name="EntityDamage" operation="perc_add" value="-.15,.15"/> <!-- random EntityDmg -->
<passive_effect name="EntityDamage" operation="perc_add" value=".1,.5" tier="2,6"/> <!-- tier bonus -->

If I'm reading it correctly, a T3 SMG has a damage modifier value calculated like this: [-0.15 to +0.15] + [0.1 to 0.5] + [0.1 to 0.5]

And a T4 would be: [-0.15 to +0.15] + [0.1 to 0.5] + [0.1 to 0.5] + [0.1 to 0.5]

 

So the range of modifiers is:

T3: from (-0.15 + .1 + .1) = +0.05 to (0.15 + .5 + .5) = +0.85

T4: from (-0.15 + .1 + .1 + .1 = +0.15 to (0.15 + .5 + .5 + .5) = +1.35

 

Putting that together with 9mm Ball ammo's base entity damage of 32 gives you the actual damage range you could see:

 

T3: from (32 x 1.05) = 34 to (32 x 1.85) = 59

T4: from (32 x 1.15) = 37 to (32 x 2.35) = 75

 

Quite a range and quite an overlap. Too much overlap, IMO.

 

9 minutes ago, jdifran said:

In my most recent playthrough, I found a cement mixer schematic on day one and promptly scrapped it

[meganoth]

48 minutes ago, Boidster said:

Boidster: Doctor, it hurts when I do this <solves polynomial equation>

Doctor: Well, don't do that.

 

Gazz (I think) has said previously not to use CM to test out the RNG system. If I remember correctly, a few months back I used 'giveself' and it also showed unexpected results especially if quantity>1. The only way I got results consistent with the expected ranges was to load up on mats and run 20 workstations crafting weapons.

 

Here's part of the XML (items.xml) which governs SMG attributes:

<passive_effect name="EntityDamage" operation="perc_add" value="-.15,.15"/> <!-- random EntityDmg -->
<passive_effect name="EntityDamage" operation="perc_add" value=".1,.5" tier="2,6"/> <!-- tier bonus -->

If I'm reading it correctly, a T3 SMG has a damage modifier value calculated like this: [-0.15 to +0.15] + [0.1 to 0.5] + [0.1 to 0.5]

And a T4 would be: [-0.15 to +0.15] + [0.1 to 0.5] + [0.1 to 0.5] + [0.1 to 0.5]

 

So the range of modifiers is:

T3: from (-0.15 + .1 + .1) = +0.05 to (0.15 + .5 + .5) = +0.85

T4: from (-0.15 + .1 + .1 + .1 = +0.15 to (0.15 + .5 + .5 + .5) = +1.35

 

Putting that together with 9mm Ball ammo's base entity damage of 32 gives you the actual damage range you could see:

 

T3: from (32 x 1.05) = 34 to (32 x 1.85) = 59

T4: from (32 x 1.15) = 37 to (32 x 2.35) = 75

 

Quite a range and quite an overlap. Too much overlap, IMO.

 

Very sure this is wrong.

The fist line adds a random value between -15% and +0.15% (because a range is given without qualifer).

The second line adds a fixed value dependant on tier qualifier. So tier3 adds 20% fixed and tier4 adds 30% fixed

 

I already tested this interpretation empirically a few weeks ago and it showed exactly the +-15% variation in each tier

 

 

[FileMachete]

I've seen these kinds of entries in the xmls, but never knew for sure what they meant. So thank you both for the useful info!

 

If I'm following this then these would be results of those xml lines?

 

T3: from (-0.15 + .2) = +0.05 to (0.15 + .2) = +0.35

T4: from (-0.15 + .3) = +0.15 to (0.15 + .3) = +0.45

T5: from (-0.15 + .4) = +0.25 to (0.15 + .4) = +0.55

T6: from (-0.15 + .5) = +0.35 to (0.15 + .5) = +0.65

 

T3: from (32 x 1.05) = 34 to (32 x 1.35) = 43

T4: from (32 x 1.15) = 37 to (32 x 1.45) = 46

T5: from (32 x 1.25) = 40 to (32 x 1.55) = 50

T6: from (32 x 1.35) = 43 to (32 x 1.65) = 53

 

Off topic but understanding this a bit more will let me tweak something I've been wanting to for quite a while; up SMG damage compared to Pistol by a bit. Longer barrel adds around ~100 fps, assuming 'basic' ammo, so ~8%+. (if interested: http://www.ballisticsbytheinch.com/9luger.html)

 

[meganoth]

You wont need an expensive GPU if you can gain 100 fps with a barrel mod 😀

 

Ok ok, the f seems to be feet.

 

 

 

 

[Boidster]

22 hours ago, meganoth said:

Very sure this is wrong.

The fist line adds a random value between -15% and +0.15% (because a range is given without qualifer).

The second line adds a fixed value dependant on tier qualifier. So tier3 adds 20% fixed and tier4 adds 30% fixed

Hmm. We may have to agree to disagree, then, as I also tested this a few weeks back. I will find that thread and re-post the table of my results. Then we can discuss where I went wrong.

 

Regarding the first line: I am not sure we disagree here. It adds a random modifier from -0.15 to +0.15, as I said. Are you saying that each tier is applying this modification? So conceivably a T6 could have -0.15*6 = -0.90 as its base modifier? I did not see any evidence of this in my testing, but I am interested in your results.

 

Regarding the second line: I do not think this is correct. Every "value" parameter in the XMLs is a range, when given as "value=x,y". That is, the value is RNG between x and y. The "tier" parameter indicates which tiers (i.e. only tiers 2-6; T1 doesn't get this additional perc_add) that RNG modifier would apply to. I do not think that the game reads the "tier" parameter and divides the "value" range equally from .1 to .5 and assigns increasing values to each tier. At least it did not seem to do that in my previous tests.

 

Back later with my results from the earlier thread.

 

Previous experiment was for tool damage. I am going to redo the experiment using SMGs. Since the only way to reliably test this is to build SMGs at workstations, it will take me a little bit. Step 1: CM 50 workstations into existence...

[Boidster]

Okay, got 100 T1 and 100 T2 SMGs cooking. Coming soon...statistics!