Some utterly insane tests

[Ruprech]

ty for the crazy tests, cheered me up no end after escaping from my latest 'bad loot period' to the forums +rep

im off to tickle some ambus with my tt enblade

 

[-=rAndom=-]

As you guys are at testing and seem to love that...
If I may suggest kill steal test. go out with 500 peds of ammo each, fist guy starts mob, second finishes and loot it, both with same "efficient" weapons.

 

[Noodles]

Perhaps MA is using a function of the form a/(b+x)?

Edit: So I think that would mean that the average loot you get is a*x/(b+x) where x is the damage done. Which will asymptote at "a" for large x.

Wow, it's so fun to quote yourself!

To expound a bit further on this very unproven function. Take the looting function y = a*x/(b+x).
Let:

y = L/Lo
x = D/Do

where
L is the average loot received
Lo is the base average loot (programmed my MA)
D is the actual damage done
Do is the base damage (i.e. the mob's hit points)

We can then say that b = (a-1) so that L/Lo equals 1 when D/Do equals 1. A good assumption? No idea. Anywho, here's a graph with a=2,5,10,and 20, and b=(a-1). Note that both axes start at a value of 1. Hard to imagine D/Do less than 1, unless kill stealing affects loot . . .

[br]Click to enlarge[/br]

Maybe something with 5<a<10 fits JimmyB's data the best?

Commentary welcome.

Noodles

 

[falkao]

Perhaps MA is using a function of the form a/(b+x)?

..Edit: So I think that would mean that the average loot you get is a*x/(b+x) where x is the damage done. Which will asymptote at "a" for large x.

From the data Jimmy collected, I have the impression that there is some kind of split, so that loot gets split in two parts. The first one is immediately paid out. The second one goes maybe to the loot pool, from where it is drawn later on.

It seems that with a normal dmg multiplier (1 till 2) about 40% of cost is returned. This drops to about 10% with a mult. of 14.


An example:
Assume cost to kill a mob is 1 PED. Let's further assume overall payout rate is 80%. So we can expect a payout of 80 PEC. Instead of getting it immediately, only 40% of the 80 PEC are returned, i.e. 32 PEC. The remaining 48 PEC are saved so that they can accumulate and be paid out as a global or some nice pedder.

Mobs that cost a lot more to kill might have an other relation. So maybe only 10% are paid out immediately. So this would finally lead to more and/or higher globals on them.

I think we see in Jimmy's data only the normal loot and some indication of high loot. Since normal loot is dominating we've got the impression that return is lower when investing more.

Now one can say, ok it is sufficient to wait and you'll get the rest back with a nice global. Unfortunately we have no idea how golbals are drawn. There is some randomization used and we have no idea if it has memory or not (i.e. if the contributors are stored or not).

If you have only short time (only some hours), then going on those mobs might be risky. If there is no payout then you'll have a huge loss. If you have to log out then maybe somebody else will drain the pool for you. So not the best strategy.

I guess Jimmy with his test, has stressed this loot function to its maximum and results are not very convincing with regard to return.

 

[falkao]

Wow, it's so fun to quote yourself!

To expound a bit further on this very unproven function. Take the looting function y = a*x/(b+x).
Let:

y = L/Lo
x = D/Do

where
L is the average loot received
Lo is the base average loot (programmed my MA)
D is the actual damage done
Do is the base damage (i.e. the mob's hit points)

..

It's possible that for normal loot reduction a formula like yours is used. However, loot itself has a skewed distribution you won't get with your formula.

 

[Noodles]

It's possible that for normal loot reduction a formula like yours is used. However, loot itself has a skewed distribution you won't get with your formula.

Right. I was imagining that the output from my proposed formula would work in tandem with the skewed distribution. Likely by changing the size (area/height) of the skewed distribution, but not its shape. Nor would it change the randomly chosen position within the skewed distribution (which for me, always seems to randomly land in the "no loot" position ).

Stated another way, the output from my proposed formula simply multiplies the output from a "standard" loot distribution formula. And there surely must be more than one "standard" loot distribution formula, since some mobs no-loot over half the time, and others never (or rarely) do so.

 

[Shaggy]

So Falco let me see if I understand your hypothesis. When I loot a mob, MA takes some of the ped I spent killing it and adds it to a loot pool for all to global on. Some of the ped I spent killing it is returned to me in loot.

 

[TimUnleashed]

OMG.

I just noticed this thread for the first time, and boy have I been missing out. Now I see the second level of humour in your comment in another post yesterday Jimmy ("not looting mobs you kill is a bad strategy").

ROFLMAO!

*subscribes to this thread immediately*

I shall have to spend some time reading it properly and get up to speed with what would on the surface now appear to be ground-breaking research.

 

[falkao]

So Falco let me see if I understand your hypothesis. When I loot a mob, MA takes some of the ped I spent killing it and adds it to a loot pool for all to global on. Some of the ped I spent killing it is returned to me in loot.

Indeed, thats how I do interpret Jimmy's data, but its only an opinion atm.

Here the fig.
[br]Click to enlarge[/br]

p seems to depend on HP.

 

[Witte]

Indeed, thats how I do interpret Jimmy's data, but its only an opinion atm.

Here the fig.
[br]Click to enlarge[/br]

p seems to depend on HP.

But that would mean there would never be noloot?

 

[Kolobok]

Indeed, thats how I do interpret Jimmy's data, but its only an opinion atm.

Here the fig.
[br]Click to enlarge[/br]

p seems to depend on HP.

Thanks falkao. At this moment I guess some more data from the field is needed to make any further hypotheses.

 

[falkao]

Thanks falkao. At this moment I guess some more data from the field is needed to make any further hypotheses.

Absolutely, and I'm rather sure we will find differences between mob types. Furthermore it would be interesting to see what a 20k HP mob drops contrasted to Jimmy's 20k Ambu.

Moreover, there is still the possibility of a bias, since the 14 mult. run was the first one.

 

[Witte]

Thanks falkao. At this moment I guess some more data from the field is needed to make any further hypotheses.

I agree. However, it seems to me that amublimax seem rather expensive to test this on, so I suggest to switch to calimusoid. I tested the male ones, and the smaller ones regenerate about 5% each 5s. Or maybe another fast regenerating mob with lower HP, but I dont know any others.

 

[falkao]

I agree. However, it seems to me that amublimax seem rather expensive to test this on, so I suggest to switch to calimusoid. I tested the male ones, and the smaller ones regenerate about 5% each 5s. Or maybe another fast regenerating mob with lower HP, but I dont know any others.

They have max 420-490 HP and can do some 50+ dmg. So they might fit well.
I think if this loot split effect exists, letting it regain to a total of 5000 HP (16 minutes) should show it. It would take, however, some time. So combining forces might help.

 

[Witte]

They have max 420-490 HP and can do some 50+ dmg. So they might fit well.
I think if this loot split effect exists, letting it regain to a total of 5000 HP (16 minutes) should show it. It would take, however, some time. So combining forces might help.

I will do some test when i find time. You can tag multiple mobs to speed up the process. It seems maturities below Berserker do 5% in 5-10 sec (5 most of the times). Berserker and up are slower. I think raider was the most common so it might be an idea to concentrate on that maturity.

 

[Kolobok]

I will do some test when i find time. You can tag multiple mobs to speed up the process. It seems maturities below Berserker do 5% in 5-10 sec (5 most of the times). Berserker and up are slower. I think raider was the most common so it might be an idea to concentrate on that maturity.

One of the problems with fast regen mobs is the variance in final HP or total damage. This is one of the important aspects that makes ambu tests rather cryptic. If you tag one mob at a time, you can track it with your program, while tagging multiple mobs it becomes close to impossible.

Before doing that test I'd suggest to do static tests with different maturities of same slow regen mob. For example, Bull Mature (550 HP) and Bull Alpha (1100). Then compare Bull Alpha with Formidon (1100) and Formicacida (1200).

My own tests that I'm not ready to publish yet show that for having more or less precise idea of how loot slots are distributed at least 50-60 mobs are needed, 100+ being much better. Feel free to PM me if you are interested to have a look at raw data spreadsheets.

 

[Jimmy B]

Interesting thoughts everyone

So Falco let me see if I understand your hypothesis. When I loot a mob, MA takes some of the ped I spent killing it and adds it to a loot pool for all to global on. Some of the ped I spent killing it is returned to me in loot.

Well there's no way of telling for sure where the remainder goes. Could go to a personal loot pool rather than one for all to global on for instance.

 

[falkao]

..
My own tests that I'm not ready to publish yet show that for having more or less precise idea of how loot slots are distributed at least 50-60 mobs are needed, 100+ being much better. Feel free to PM me if you are interested to have a look at raw data spreadsheets.

Thats indeed true, for a complete picture even more data might be needed. From personal old data and that provided by Jimmy, there seems to be a gaussian distributed normal loot, wheras exceptional loot has a skewed distribution.
If the loot split theory is correct, then a payout might be triggered only after several killed creatures. So yes, you might have to wait for long time.

Moreover, in my old tests I noticed quite some differences between runs on the same mob and maturity but didn't analyze it in detail. This might bias results.

 

[falkao]

Interesting thoughts everyone



Well there's no way of telling for sure where the remainder goes. Could go to a personal loot pool rather than one for all to global on for instance.

that's correct and this should be tested as well.

 

[Jimmy B]

Thats indeed true, for a complete picture even more data might be needed. From personal old data and that provided by Jimmy, there seems to be a gaussian distributed normal loot, wheras exceptional loot has a skewed distribution.
If the loot split theory is correct, then a payout might be triggered only after several killed creatures. So yes, you might have to wait for long time.

My feeling is that the distribution of loots is indeed banded a bit into certain boxes, but the fast regeneration of the ambus meant damage done/cost to kill varies quite a lot thereby causing the loot distribution to spread out a bit.

 

[Vedder]

Interesting thread. (I'd +rep again if I could)
It seems that loot IS somehow connected to total dmg or cost, unlike what I would have thought.
But I guess it makes sense, since MA then doesn't need a 'loot table' per mob, but simply have 1 calculation for everything. (Maybe 2, but more about that later)

I don't have anything to contribute with except a couple of speculations.

1. My guess is that MA just calculates the total damage done, rather than cost to kill.
   This is much simpler for MA (to implement, but also with regards to performance when calculating loot etc).
   And I really don't see why they would reward inefficiency anyway.
   But ofc you never know.
   
2. Wonder if this means that if you let a smaller mob regenerate enough, it becomes an 'always looting mob'?
   
   Actually this might explain the different loot pattern for your ambus with more than 4 times regen, as I'm pretty sure that always looting mobs have a very different loot pattern than normal mobs. Basically a much larger part of the loot is paid in globals for these mobs, than it is for other mobs.
   
   IIRC mobs with more than 1000 HP used to always loot. Then MA changed it, and now the limit seems to be around 2000 HP, but seems some mobs with slightly less always loot, while some with a bit more sometimes don't. (Disregarding the loot-lag no-looters).
   I think MA might have spiced up their criteria for always loot to include some kinda of toughness factor maybe based on mob dmg and speed or something like that.
   

 

[falkao]

My feeling is that the distribution of loots is indeed banded a bit into certain boxes, but the fast regeneration of the ambus meant damage done/cost to kill varies quite a lot thereby causing the loot distribution to spread out a bit.

Here a fig. what I'm talking about:
[br]Click to enlarge[/br]

This data is 2 years old (Beginning 2006) and shows Cornundacauda loot. It was provided to me by Steffel I think. It contains n = 246 kills for mature till dominant.

In the graph I subdivided loot in classes ranging from 0 to 80 PEC, 80-200 PEC and >200 PEC. Btw. no loot rate was about 55%, therefore the high loot on them. Mean cost to kill one is about 70-80 PEC.

Since I had this data, I did further add HP to the dataset and it seems that payout is reduced with higher HP. (Regression Coeff is -0.6 with a p<0.001). So it seems, that the Ambu finding is valid also for this old data.

The skewed distribution on high loot might be related to missing further loot group.

 

[TimUnleashed]

I have nothing but ideas and speculation to add:

If the loot had anything to do with actual cost including markup, or was affected by the markup data, it would undermine the free market economy and render it irrelevant. I am 100% certain that if there is a relationship between cost and loot, it will be to do with base (TT) cost only. Anything else would be ludicrous and completely at odds with MA's overall strategy.

Thanks Oleg. That put into words a feeling I was having trouble verbalising.

It seems to me logical that you should try to kill the mob with as little cost as possible.

Yes of course, but doesn't it seem counterintuitive that MA should encourage eco by rewarding it with loot? Perhaps the big loots (globals hofs etc) are linked to UNeco-ness, but average loot might not be? Or something?

I can easily show that its cheaper (including markup) to kill an Ambulimax Young with a HL14 at 200% than it is to kill it with a H400 at 150%. The extra dmg/sec means the ambu doesn't regenerate as much, and this outweighs the extra cost in markup.

Given your tests thus far are only on Ambu, that's fair enough. I wonder if the tables turn though on a less regenerative mob? Is average looted linked to the mob's actual regeneration factor? (to try and coin yet another insane phrase in this insane testing thread).

Will be interesting to see Witte's Calumisoid data, and then maybe look at some data from a slow regen mob and see how it looks...

This rather puts me off using the higher markup guns.

Against ambu, yeah, but against all mobs? Am I missing something that allows you to conclude that? I guess I'm assuming that we're not using common sense, and only using data in this insane thread

An example:
Assume cost to kill a mob is 1 PED. Let's further assume overall payout rate is 80%. So we can expect a payout of 80 PEC. Instead of getting it immediately, only 40% of the 80 PEC are returned, i.e. 32 PEC. The remaining 48 PEC are saved so that they can accumulate and be paid out as a global or some nice pedder.

Mobs that cost a lot more to kill might have an other relation. So maybe only 10% are paid out immediately. So this would finally lead to more and/or higher globals on them.

Again, I wonder if this "immediate return factor", whether it be 10% or 40%, can be linked to the mob's "regeneration factor" (among other frickin variables).

Now one can say, ok it is sufficient to wait and you'll get the rest back with a nice global. Unfortunately we have no idea how golbals are drawn. There is some randomization used and we have no idea if it has memory or not (i.e. if the contributors are stored or not).

If you have only short time (only some hours), then going on those mobs might be risky. If there is no payout then you'll have a huge loss. If you have to log out then maybe somebody else will drain the pool for you. So not the best strategy.

EU is dynamic


Lol, this thread's gold - keep the stats coming in guys - we'll get something helpful out soon!

 

[Kerham]

Very interesting thread, I don't have much to add, I'll spread some rep around

My observation (which i never found the time to twist around in a mathematical background) would be that the factor seem to be the max dmg done by the wep. That would represent the ..uhm "total loot". In average, you'll receive the loot corresponding to your personal dmg on that wep and minins/globals/hofs would reimburse on certain intervals the difference between what dmg you do with the gun and the max dmg of the gun. Since nobody can do permanent max dmg, would always be place for minis/globals even for the highest skilled. Imo that would explain why the very skilled players hof so very rare compared to how often them hunt. By hof i mean a big payout, not the 4-500 pedders.

Together with the dmg, I think is also reimbursed the decay of the weapon, no matter how big it is, and the amp. Let's call it active decay.

While the passive decay armour+fap seem to work like a penalty reimbursed via skills.

Offcourse, there are holes in this observation, like using the E-series amps. But honestly I am not terribly interested in using them.

Reffering to regeneration, I am twisted. On the mobs I hunted, with the old regen system, seemed that there's a margin of "acceptable" regeneration.

And also, and that's my most empiric observation, with absolutely no facts to back it up, I think that exist a sort of ration to kill speed. If you kill too fast or too slow, you're bound to loose.

Somebody, quite famous in the times, told me when I was a little newb, that if I can kill a mob in 10 or less shots, to not use an amp. Would be 2,5 yrs since that advice, didn't managed to fully understand it, other than the possibility of overkill.

 

[Jimmy B]

Thanks for the comments. I'm currently doing some more tests, of a slightly less costly but more time-consuming nature. I'll publish the results when I'm done, I've done 2 and a half out of 5 planned test runs so far.

I haven't really examined the data I've got in any detail yet, but it does tentatively look to me like loot is based on damage done rather than cost to kill.

Yes of course, but doesn't it seem counterintuitive that MA should encourage eco by rewarding it with loot? Perhaps the big loots (globals hofs etc) are linked to UNeco-ness, but average loot might not be? Or something?

To be honest, I think skill/profession level triggers are important with the big loots. Personally most of my best individual loots have been when I've been being eco, but that could just be how it turned out.

I always used to view it like this: If average cost to kill is 1 PED, average loot is something like 95pec and the eco players who get below 95pec make a tt profit, but the majority lose a bit. Its kind of a similar model to online poker which was my stepping stone to EU so it seemed the natural way for it to work to me. Clearly doesn't quite work like that though.

Given your tests thus far are only on Ambu, that's fair enough. I wonder if the tables turn though on a less regenerative mob? Is average looted linked to the mob's actual regeneration factor? (to try and coin yet another insane phrase in this insane testing thread).

Will be interesting to see Witte's Calumisoid data, and then maybe look at some data from a slow regen mob and see how it looks...

Against ambu, yeah, but against all mobs? Am I missing something that allows you to conclude that? I guess I'm assuming that we're not using common sense, and only using data in this insane thread

Yeah the comments about the HL14 vs the H400 only apply to the fast regenerating mobs. The thing here is that despite its lower eco (due to markup) the HL14 works out cheaper to kill an Ambu than the H400. However, if as appears to be the case, the extra regeneration when using the H400 actually results in more loot then it may well be better to use the H400. The HL14's advantage is extra dmg/sec resulting in lower regeneration which offsets the extra markup, but if the extra regeneration with the H400 results in extra loot then the better eco of the H400 wins out.

For mobs that do slow regeneration and low damage, the H400 will always be better than the HL14 in theory.

For mobs that do slow regeneration and high damage, the HL14 may well be better since its extra dmg/sec reduces armor and fap decay.

Somebody, quite famous in the times, told me when I was a little newb, that if I can kill a mob in 10 or less shots, to not use an amp. Would be 2,5 yrs since that advice, didn't managed to fully understand it, other than the possibility of overkill.

Yeah, MA have always said use the right weapon for the right mob so its not implausible that something like this is relevant. My tests could potentially see this effect if its there, I'll try to get em done soon but WoF will slow me down a bit

 

[TimUnleashed]

I always used to view it like this: If average cost to kill is 1 PED, average loot is something like 95pec and the eco players who get below 95pec make a tt profit, but the majority lose a bit. Its kind of a similar model to online poker which was my stepping stone to EU so it seemed the natural way for it to work to me. Clearly doesn't quite work like that though.

Right, I see, thank you.

Yeah the comments about the HL14 vs the H400 only apply to the fast regenerating mobs. The thing here is that despite its lower eco (due to markup) the HL14 works out cheaper to kill an Ambu than the H400. However, if as appears to be the case, the extra regeneration when using the H400 actually results in more loot then it may well be better to use the H400. The HL14's advantage is extra dmg/sec resulting in lower regeneration which offsets the extra markup, but if the extra regeneration with the H400 results in extra loot then the better eco of the H400 wins out.

For mobs that do slow regeneration and low damage, the H400 will always be better than the HL14 in theory.

For mobs that do slow regeneration and high damage, the HL14 may well be better since its extra dmg/sec reduces armor and fap decay.

Again, that clarifies it for me, thanks. Something may come of this yet

 

[Kolobok]

Finally there are some data to be published:

Google Docs - Formidon Young Loot Test

This spreadsheet records the data of 4 runs at Formidon valley, with total of 280 mobs killed. Each individual loot is recorded. Some statistics is provided with the first datasheet, along with empirical costs to kill. While I leave the detailed statistical analysis to mathematicians, here are some remarks that are relevant to the subject of this thread.

First, skewed distribution of loot is evident. There are distinct loot slots, five of them can be easily identified. That is, no loot, 82-107 pec, 149-181, 209-278, 313-380 etc. each slot has average value and width, for example, for slot 2 it is 165pec +-10%, similar for other slots.

Second, this distribution is mob dependent. Number of no looters is about 30%, while for low HP mobs it is about 50% usually. The results for ambu show that MA implements different distributions for different mobs. The spreadsheet has some data for Formidon Mature and Formicacida Weak, with very similar HP - numbers look also very similar, so I think the distribution for low regen mobs is very likely closely related to HP.

Third, the average return without globals and good minics is a little bit less than 50% of the avg. cost to kill a mob. That is very close to what I get on bad hunting and mining runs.

Why this distribution exists at all? Because to have fun of globals and hofs MA has to implement a sort of distribution scheme, in which, for example, 50% return is almost guaranteed, 85% is quite likely with some minics and globals, and the rest is payed as hofs to fortunate ones. It is possible (MA side) to predefine the probabilities of individual loots in such a way that the total payout never exceeds the theoretical minimum of expenses and MA never has to worry that it pays more than it gets.

As a side note - it seems that this scheme is implemented not based on static probabilities, but as a time-based scheme, as it is a common feeling that globals and minics come in batches. Some people interpret it as a "reimbursement" for their expenses, but personally I consider this idea very misleading (and encouraging for more losses) and don't see any grounds for it. Expenses can be raised very easily - shot 10k ammo in the open air or drop 100 bombs at PA - what next?

Now some guesswork. My feeling is that the variance of +-10% of loot slots is not entirely casual. It accounts for three effects: first is pre-defined statistical variance, second is the difference in the damage done by the hunter to the mob. Jimmy's tests will prove it as usual beyond any doubt, I believe.

The third effect is the damage that the mob inflicts to the hunter. It is also likely that MA tracks this and pays for this, too. Difficult to prove, but there are some indications for this. First, there are mobs that have same HP but different damage. Second, some observation: during the fourth run I got a lagged mob that ran forth and back unable to hit me (and unable to regenerate, I think). I killed it finally and got extremely unusual low loot - 82 pec. Never before I had this low loot. Then I paid attention to the damage that a mob inflicts to the hunter and there came the next mob that hit me only 2 times, for 9 dmg amd 23 dmg (rest absorbed by the armor) and gave loot of 84 pecs, the second minimum. Usually formidons hit me 4-6 times for 1-40 dmg average and at the end I have little HP left if I don't fap in the middle.

To summarize: MA know the theoretical cost to kill a mob of a certain HP and pays for this. It is based on HP taken or damage inflicted (the difference is that the last option makes the concept of overkill void). It is possible that damage to the hunter is compensated, too. It is up to hunter to be as efficient as possible to minimize costs and get closer to the theoretical cost (misses, weapon markup, armor and fap decay are his enemies). The payout is based on some dynamic distribution schemes, in which half of the return is allocated to low loots and another half to high loots.

Hopefully further tests will show that my guesswork is total bs and present some "ultimate" evidence how it works lol. Loot discussion is everlasting, have fun

 

[Noodles]

Second, this distribution is mob dependent. Number of no looters is about 30%, while for low HP mobs it is about 50% usually. The results for ambu show that MA implements different distributions for different mobs. The spreadsheet has some data for Formidon Mature and Formicacida Weak, with very similar HP - numbers look also very similar, so I think the distribution for low regen mobs is very likely closely related to HP.

+rep for great data!

That confirms at least three different loot distributions; 55% no loot for some noob mobs, 30% for these, and nearly 0% for some large ones.

I have been doing nearly identical tests as these on noob mobs (though not including cost to kill), broken down by type and maturity (to verify the claims in this thread and this thread). Don't have enough data yet to publish, but maybe in a couple weeks.

Is the data beginning to suggest that there might be a continuum of no-loot percentages, from perhaps 55% for noob mobs to 0% for the tougher mobs, based on HP? Or it could sure be done discretely, since there are only three confirmed no-loot percentages so far.

Or correct me if I'm wrong .

What fun!

Noodles

 

[Jimmy B]

Cheers Kolobok

I haven't put my data into a spreadsheet yet but mine is also in slots. The tests I have done are:

102 Plumatergus Young (40 HP) with tt handgun no attachments
102 Plumatergus Young with LR59+A106
Currently on a run with tt handgun+two H-lasers and an I-scope (increasing cost per shot by a little over 25%)

Also plan to do Mann MPH Dlx or Mux-1 (very uneco, at least doubling cost per shot) and possibly something else.

I shall also try a crazy armor test at some point on some trox.

The Overkill with the one-shot set up is interesting as it allows no regeneration. The first three slots are 5-6, 9-11, 14-17. There's no evidence of increased loot due to extra cost to kill from this one, leading me to think that damage done is the key rather than cost per shot. The M2100 allows a bit of regeneration and the slots do appear to be shifted up a bit, and also show a trend for longer kills (I've recorded hits/misses/damage done) to end up generally higher in the slots than quick kills (less regen). There were also more loots in lower slots in general, and more no loots, in the overkill test interestingly (could just be a bad run though).

Interestingly only missed 4 times with the LR59, and two of those were when the mob moved just before I shot, and 1 was on a pretty steep slope. I might try that run again to see if I can reproduce that hit rate, it is admittedly a rather small sample. Took 46 kills before the first miss!

 

[Legion]

I made a post here about my loot theory, slightly OT but might be interested for those ppl reading and discussing here.
https://www.planetcalypsoforum.com/...-yet-another-theory-thread-d-loot-distro.html