Some utterly insane tests

Jimmy B

Indioni Jones
Joined
Dec 2, 2005
Posts
12,256
Location
UK
Society
Skillin' Villains
Avatar Name
Oni
Right, decided to do some loot tests after discussion arising in this thread, and elsewhere.

Test 1 - Kill some Ambulimax Young moronically

Equipment used: Gremlin+5B, Fap-5, Fap-2600, H350 (L) with two H-lasers, an I-scope and an A102. Couple of hits with Valor to finish.

Purpose: To test the theory that cost to kill mob, or amount of damage dealt, has an impact on the loot. The new ambulimax were perfect since they can't kill me, and their fast regeneration makes things much faster.

Results (numbers are PED, first number is first ambu killed etc, markup isn't included):

Ammo: 50, 50, 50, 50, 50, 50
Armor: 3.45, 2.63, 2.96, 2.89, 3.21, 3.02
Weapons: 16.58, 16.9, 16.9, 16.74, 17.37, 16.43
Fap: 1.2, 0.8, 0.86, 0.67, 0.76, 0.94
Total Cost: 71.23, 70.33, 70.72, 70.30, 71.34, 70.39
Loot: 5.97, 130.70, 5.88, 8.41, 4.37, 2.48

Overall spend: 424.31 (70.72 avg)
Overall loot: 157.81 (26.3 avg)

After it went ~6, ~130, ~6 I was so sure I was getting 130 on the next one lol :D

Test 2 - Kill some Ambulimax Young sensibly

Now I killed 12 Ambulimax Young with H350 replaced by a Scrof, and the A102 replaced by A106.

Total Ammo: 39.8
Total Armor: 0.77
Total Weapons: 19.85
Didn't fap.
Total Cost: 60.42 (5.035 avg per ambu)

Loots: 1.21, 1.45, 1.97, 2.90, 3.99, 1.19, 1.90, 2.14, 2.12, 40.35, 4.29, 2.01

Total Loot: 65.52 (avg 5.46)

(lowest single ambu loot I received earlier during ambumania was 0.60 I think)

Conclusion, Part 1

Cost to kill was roughly 14 times more in the regeneration experiment than in the more sensible run. The vast majority of this is extra ammo & weapon decay from letting the mob regenerate.

It seems that in the short term at least, the extra damage dealt isn't turned proportionally into loot. However, it does seem there is some extra loot. It kinda looks a bit like the in test 1 was multiplied by 3 from the 'standard' loot, but its hard to tell from such a short test. Quite a lot of the items did come in stacks that divided 3, but not all (e.g. got 1 Kidney Oil in one loot). It could plausibly be that its capped at 3, so may be worth doing a retest just letting the mob fully regenerate twice.

One other possibility is that its somehow calculated based on the weapon - using the formula in falkao's thread listed above, without pauses to allow regeneration it would be reasonably close to 3 times more expensive to kill ambu with H350 than with Scrof. My feeling was that I could have killed the ambus in roughly 1000 ammo which is fairly consistent with that, so it may be worth retesting like that if I have the stomach for it.

The global was 6 Brain (114 PED), 12 residue, 3 skin color set (33pec), 36 Adrenal (7.2 PED), 905 Ammo.

Test 3 - Mindless Exarosaur slaughter

Had about 42 PED ammo worth of decay left in the H350 which I'll never use in normal circumstances, so I went to slaughter some exarosaurs with it. However, in keeping with the insanity theme of the experiment, I didn't loot them.

Total spend was nearly exactly 60 PED. I killed 11 at the end:

Mature: 0.44
Old: NL, 0.35, 0.17, 0.16, 0.08
Guardian: NL
Dominant: NL, NL, 0.82
Alpha: 0.53

Total Loot: 2.55

Conclusion, Part 2

Not looting the mobs you kill is a bad strategy.
 
Last edited:
Equipment used: Gremlin+5B, Fap-5, Fap-2600, H350 (L) with two H-lasers, an I-scope and an A102. Couple of hits with Valor to finish.


Not looting the mobs you kill is a bad strategy.

Ok, hell of a finding:)... but what the hell did you use a Fap-5 for?

Hurrikane
 
Please, please, PLEASE let me use this. :laugh:

lol, feel free :D

Ok, hell of a finding:)... but what the hell did you use a Fap-5 for?

Hurrikane

The fap-5 was for the Ambulimax test. I mostly fapped up with fap-5 occasionally to give the ambu time to regenerate, just used the 2600 to fap when a crit left me slightly vulnerable to death from another crit.

I think letting the ambu regenerate may have been a mistake. The 'natural' cost to kill the ambu using the H350 would be around 3 times more than with the scrof, and the numbers do suggest the tt loot may have been multiplied by about 3. So I guess there's another insanity experiment in the making. :D
 
Er, good work, I think :scratch2:

I'm a bit pissed off that you revealed that not looting mobs gives less loot than looting them. I've been trying to keep that one a secret so I could get an edge over everyone else :rolleyes:
 
Does these fit into seed of .5pecs * dmg done with variance of 75%-100% depending on skillgains you sucked out dueing killing process? And that seed is divided or multiplied by area, mob, you etc modifiers.

Doing it like this explains "holes" in loot (as with deposit sizes and crafting clicks) tt values. :)
 
well..uhm.. great test, dont really understand WHY :)
but hey.. its not the first time i dont understand something :)
 
Another nice test. That deserves some +REP
 
The first test on ambus would have been far more revealing if you knew the time-based loot formulas. Without them, as you've said yourself, it would take a lot (100x) more kills to see the pattern.
 
Did it really cost 70+ PED per mob using the H350? That's insane :)
 
Did it really cost 70+ PED per mob using the H350? That's insane :)

actually i am amazed that he actually killed 1 with the 350 :laugh: I think it took at least 30 min / mob killed
 
This may seem insane at first glance, but who would have guessed those massive differences in costs to kill one mob and all down to wrong choice of gear?

This is extreme, but a slightly more sensible choice would have a similar effect on costs so this shows the importance of taking the time to select wisely.

Another tip following on your useful Exa test, it's not great to leave mining finds in the ground either, not least because a little reminder pops up just when it's too damn late to NeeNahh :nana: you :laugh:

+rep if it lets me :laugh:

t
 
please can you compare apolo+101 and scrof+106 on amulimaxes ? :laugh:
(indeed utterly insane)
 
I totally agree with the topic name. These tests are insane :laugh:.

There does indeed seem to be a relation between cost and loot. But we still dont know what causes it. It might be regeneration, but it can also be the armor decay. It seems to me that armor decay better fits the results you have now. However, we do not know how loot progresses when HP becomes higher, and it might well be that regeneration has a lot to do with it. When I think about the hoflist, I remeber a pretty much unknown avatar (at least to me) had a 17k HOF on ambu young. The same day BlackHawk (from my knowledge using MM) had just 14k. This difference could be explained by regeneration.
On the other hand, when I was testing the regeneration time of scips, I noticed that a scip that hit me many times seemed to give me an unusual high loot of about 9 ped. I dont think scip young usually give that kinda loots, although im not very sure about that.

Over all, I do think we need to isolate these aspects in future tests. That will also make the tests less insane :laugh:
 
Actually i don't think you should let it just regenerate. Instead use a more uneconomical weapon to kill it with and another that is more economical.

Also do longer tests ie hunt for an hour atleast for each setup.
 
It could plausibly be that its capped at 3, so may be worth doing a retest just letting the mob fully regenerate twice.

This sounds very plausible indeed. Loot can become higher with higher dmg inflicted, but not infinitely. Can you please also add a rough estimate of the dmg inflicted and/or total HP of mob to your results, based on ammo spent and regen rate? Using Witte's tool for recording damage would be even better in future.
 
This sounds very plausible indeed. Loot can become higher with higher dmg inflicted, but not infinitely. Can you please also add a rough estimate of the dmg inflicted and/or total HP of mob to your results, based on ammo spent and regen rate? Using Witte's tool for recording damage would be even better in future.

You could calculate damage from the weapon decay or ammo spend. Although it seems weird that the ammo spend is constant while decay is different in the tests:scratch2:
 
I don't see this insane but quite interesting. First it is proven that there is a HP or dmg related payout, and second it seems there are different payout schemes.

My hypothesis:

Per done dmg (HP of mob) you do a contribution to a loot pool. From this you'll get a payout weighted for the contributed amount.

Example:
Mean weapon related cost for Korss350L was about 66.8 PEd. This corresponds to about 29,200 HP. Per HP you'll contribute 0.1 PEC to the loot pool, i.e. 29.2 PED. Instead of getting it immediately there is a weighted payout. You'll get 1/5 in 83% of cases and 5 times as much in about 17% of cases. This is still 29.2 PED but makes the game more interesting.

With Scorf you'll do a dmg of abou 2,100 HP. So you'll contribute 2.1 PED to the pool. From this you'll get ½ in 66% of cases and twice as much in 34% of cases.

Well, this is only an example but it shows a principle. I guess there are more slots, and the weighting might be differently.
 
Does these fit into seed of .5pecs * dmg done with variance of 75%-100% depending on skillgains you sucked out dueing killing process? And that seed is divided or multiplied by area, mob, you etc modifiers.

Doing it like this explains "holes" in loot (as with deposit sizes and crafting clicks) tt values. :)

Sorry, you've lost me.

The first test on ambus would have been far more revealing if you knew the time-based loot formulas. Without them, as you've said yourself, it would take a lot (100x) more kills to see the pattern.

Well, my test was mainly designed to determine whether or not loot is proportionally modified by regenerated health, as suggested by a number of people and old tests.

Whilst the shortness of the experiment makes it a little hard to draw concrete conclusions, I think we can say for sure that letting a mob regenerate 14 times does not result in the loot being 14 times greater. However, it also appears the loot has however been modified up somewhat.

Did it really cost 70+ PED per mob using the H350? That's insane :)

Yes, because I allowed the mob to regenerate. You could probably kill it for 15-20ped with a H350+A102 if you just shot it without pause.

actually i am amazed that he actually killed 1 with the 350 :laugh: I think it took at least 30 min / mob killed

About 30min per mob yes.

There does indeed seem to be a relation between cost and loot. But we still dont know what causes it. It might be regeneration, but it can also be the armor decay. It seems to me that armor decay better fits the results you have now. However, we do not know how loot progresses when HP becomes higher, and it might well be that regeneration has a lot to do with it. When I think about the hoflist, I remeber a pretty much unknown avatar (at least to me) had a 17k HOF on ambu young. The same day BlackHawk (from my knowledge using MM) had just 14k. This difference could be explained by regeneration.
On the other hand, when I was testing the regeneration time of scips, I noticed that a scip that hit me many times seemed to give me an unusual high loot of about 9 ped. I dont think scip young usually give that kinda loots, although im not very sure about that.

Over all, I do think we need to isolate these aspects in future tests. That will also make the tests less insane :laugh:

I'm not sure armor decay is that likely but its something I was thinking of testing at some point (something like: stand in front of Atrox Dominant in Ghost+5B for half an hour and then shoot it).

I agree the test results are inconclusive, but at the same time the results are quite interesting.

Actually i don't think you should let it just regenerate. Instead use a more uneconomical weapon to kill it with and another that is more economical.

Also do longer tests ie hunt for an hour atleast for each setup.

The test was mainly done to test the theory the loot is based on damage done, rather than the mob's original HP. So I pumped 14 times as much into them than normal to see if the loots were modified. Looks to me they were modified by a factor of three roughly, although that's not at all conclusive, but it is interesting that killing them naturally (ie. no pauses for regeneration) with the H350 set up would have been about 3 times more expensive than killing them with the more sensible set up.


You could calculate damage from the weapon decay or ammo spend. Although it seems weird that the ammo spend is constant while decay is different in the tests:scratch2:

On average the damage done should have been about 18500-19000. Each test had 1000 shots fired, the decay is different by multiples of the valor decay used to finish the mob (all but the last ambu have a hit or two from the valor).


falkao said:
Example:
Mean weapon related cost for Korss350L was about 66.8 PEd. This corresponds to about 29,200 HP. Per HP you'll contribute 0.1 PEC to the loot pool, i.e. 29.2 PED. Instead of getting it immediately there is a weighted payout. You'll get 1/5 in 83% of cases and 5 times as much in about 17% of cases. This is still 29.2 PED but makes the game more interesting.

Yeah it could work something like this, the test is a bit too short to draw such conclusions with any certainty though. I may do some more tests at some point if I have the stomach for it, its a bit pricy though ;)

please can you compare apolo+101 and scrof+106 on amulimaxes ? :laugh:
(indeed utterly insane)

Sure. With the Scrof+106 it'll cost you about 5 ped to kill one. With the opalo+101 you can empty your entire ped card into one, as you'll never kill it.
 
Last edited:
Purpose: To test the theory that cost to kill mob, or amount of damage dealt, has an impact on the loot. The new ambulimax were perfect since they can't kill me, and their fast regeneration makes things much faster.

Results (numbers are PED, first number is first ambu killed etc, markup isn't included):

Ammo: 50, 50, 50, 50, 50, 50
Armor: 3.45, 2.63, 2.96, 2.89, 3.21, 3.02
Weapons: 16.58, 16.9, 16.9, 16.74, 17.37, 16.43
Fap: 1.2, 0.8, 0.86, 0.67, 0.76, 0.94
Total Cost: 71.23, 70.33, 70.72, 70.30, 71.34, 70.39
Loot: 5.97, 130.70, 5.88, 8.41, 4.37, 2.48

Overall spend: 424.31 (70.72 avg)
Overall loot: 157.81 (26.3 avg)

After it went ~6, ~130, ~6 I was so sure I was getting 130 on the next one lol :D

I waited sooo long before being able to do that .....
"I'm sorry Jimmy but your data set is too small, you can not take any conlusions based on such a small run"

:D

Now more seriously ...

I'm camping ambus since the new VU and i am not planning to leave them before they gave me that #@#"!!@ 10K loot. I will now keep track of my runs using the different guns i use on them.

My test runs will be like :
  • MKIII + A104
  • MKIII Unamped
  • Karma + A104
  • Karma Unamped
  • Combo MKIII/Karma both with A104
  • Combo MKIII/Karma both Unamped

That should give you guys a good set of data to play with (as i personnaly don't like such games ;)).
 
I waited sooo long before being able to do that .....
"I'm sorry Jimmy but your data set is too small, you can not take any conlusions based on such a small run"

Hehe and yes you're right. We can't draw many hard conclusions. However, we can draw some possible insights and then make a more informed decision on the next way to test. We can for instance pretty conclusively deduce that spending 14 times more than normal killing a mob, does not result in each mob giving you 14 times the loot. Other than that we can only make conjectures and subject them to further testing.

Nothing wrong with making conjectures based on limited data sets. You just have to be aware they're just conjectures and not treat them as hard facts.
 
I waited sooo long before being able to do that .....
"I'm sorry Jimmy but your data set is too small, you can not take any conlusions based on such a small run"

It depends on the conclusion one would take. For sure you can't reveal any hidden parameters, but as Jimmy already mentioned, it is proven now that loot depends on total dmg done, and since dmg costs its total cost to kill it.

The difference in loot is statistically significant, so the sample size is sufficient to conclude this.
 
We had this talk last night with Oleg on soc chat on whether amount of dmg done/ammo spent influences loot.

I usually go on 200ped ammo hunts on ambus - been doing it for a few weeks now. There are long periods when I get just around 1 ped loot from each ambu - sometimes even during the whole hunt. Then next hunt with exactly same equipment, same location (nea's) I end up getting 3+ ped loot per ambu.

So based on that I am a bit sceptical about the whole experiment.
 
A moderately crazy test

OK was going to go a redo the test with the H350 but not pausing to allow the ambu to regenerate. Couldn't find a H350 on auction nor anything that would give me about the same damage per sec. So I used a slower Ruker UL1 instead.

Equipment: Gremlin+5B, fap-2600, fap-5 between mobs, Riker UL1+102 without the scopes & sights, couple of swings of Valor when gun broke very near the end of mob 6. Mobs were killed with as little interruption as possible, a couple of faps were necessary per mob.

Total Ammo: 93.35
Total Armor: 5.52
Total Weps: 42.00
Total Fap: 1.99

Total Spend: 142.86
Average Spend: 23.81

Loots: 6.4, 2.43, 5.61, 2.62, 12.53, 9.4
Total Loot: 38.99

Loots were:

38 Soft hide, 6 residue, 234 ammo
6 Adrenal, 1 residue, 122 ammo
1 kidney, 1 pancreas, 293 ammo, 18 residue
1 kidney, 3 adrenal, 2 residue
14 pancreas, 141 leather ext, 19 residue, 393 ammo
3 kidney, 267 ammo, 3 adrenal, 13 residue

Conclusion

Again this is I think pretty clearly inherently different from 'normal' ambu loot. I'm not convinced there's a huge amount of evidence of any difference from the insane test though.
 
Hmm thats very interesting and thank you for doing such expensive testing for the community.

I have always noticed my loots are different if I hunt with my ewe ep 41 or a manis on ambu's and pretty much everything else.

Also I notice a change in my loot now that I use a fire forge instead of a bravo for tagging.
 
I was hoping someone else would do this testing so I wouldn't have to myself. :D

It's starting to look pretty conclusive to me, don't hunt mobs that you can't kill quickly with your weapon/armor. If they change all mobs like they have ambu, allo and spiders, suddenly a lot of mobs are going to become inefficient to hunt solo for a lot of mid level hunters. This will also marginalize the advantage of adj/imp/mod faps, as just 2 faps and the mob regenerates.

edit: need to spread rep blah
 
I guess the ambu test was a good choice as they loot better than other mobs but would an Exo test be more cost effictive?....guess then it wouldnt be that insane. Kill 50 with opolo then kill 50 with valor. Good test though, keep up the insanity. :yay:
 
It's starting to look pretty conclusive to me, don't hunt mobs that you can't kill quickly with your weapon/armor.

I'm not convinced this is proven yet to be honest. I got quite a few 60pec loots whilst hunting Ambu with HL14 in Ambumania. If we assume that's about as low as it goes, currently the lowest loots in the insane and crazy tests are around 240pec. For the crazy test above, that could conceivably correspond to tt loot being multiplied by about 4 which isn't so far off the extra it costs to kill.

What I think may be indicated is:

(i) It looks like loot is modified up somehow based on cost to kill or damage done. Its not clear exactly how this modification works though.
(ii) Letting the mob regenerate beyond what it would naturally regenerate whilst shooting with the chosen weapon doesn't seem to affect the loot. (which is different to the conclusion Witte drew from a similar test a long time ago)

(ii) needs to be tested a bit further to be sure, by shooting with something sensible and letting the mob regenerate fully twice.

I guess the ambu test was a good choice as they loot better than other mobs but would an Exo test be more cost effictive?....guess then it wouldnt be that insane. Kill 50 with opolo then kill 50 with valor. Good test though, keep up the insanity. :yay:

Ambu were chosen because of their regeneration properties which makes it much easier to spend significantly more than normal on them.

The problem with the valor/opalo test on Exaros, is that the difference in cost to kill isn't particularly big. That makes it very hard to tell whether differences in loot observed are just noise or not.

We had this talk last night with Oleg on soc chat on whether amount of dmg done/ammo spent influences loot.

I usually go on 200ped ammo hunts on ambus - been doing it for a few weeks now. There are long periods when I get just around 1 ped loot from each ambu - sometimes even during the whole hunt. Then next hunt with exactly same equipment, same location (nea's) I end up getting 3+ ped loot per ambu.

So based on that I am a bit sceptical about the whole experiment.

Its a short run yes. I'm aware of the dangers of leaping to conclusions with insufficient data to back it up.

However, we can draw some conclusions, and also each test points at other things to test that will broaden our knowledge.

We have certainly learnt with little room left for uncertainty that pumping 14 times as much ped into an ambu than normal does not give you 14 times as much loot on each individual mob. For that to be true it would have to be possible to loot 17pec from an ambu when killing them normally. I've never seen that since the new ambus arrived.
 
Last edited:
I'm not convinced this is proven yet to be honest. I got quite a few 60pec loots whilst hunting Ambu with HL14 in Ambumania. If we assume that's about as low as it goes, currently the lowest loots in the insane and crazy tests are around 240pec. For the crazy test above, that could conceivably correspond to tt loot being multiplied by about 4 which isn't so far off the extra it costs to kill.

What I think may be indicated is:

(i) It looks like loot is modified up somehow based on cost to kill or damage done. Its not clear exactly how this modification works though.
(ii) Letting the mob regenerate beyond what it would naturally regenerate whilst shooting with the chosen weapon doesn't seem to affect the loot. (which is different to the conclusion Witte drew from a similar test a long time ago)

(ii) needs to be tested a bit further to be sure, by shooting with something sensible and letting the mob regenerate fully twice.

If i and ii are true, wouldn't that mean that there must be a complicated, arbitrary way to determine loot? MA would have to differentiate between regeneration while "actively shooting" and regeneration while just standing there. How would they do that?

Or, perhaps there is a hard limit for how much regeneration will be compensated?

Perhaps a better test would be to kill as fast as possible with an h400, and then another weapon with around 2x the dam/sec?
 
Armour decay having an effect got me thinking, and remembering.

I have no facts or figures to back this up.

I suspect it does play a part, and perhaps a large part. Mobs I used to need armour for, as I stood no chance without it, and can now kill without armour seem to loot far worse than they used to.

Putting armour on (pixie etc) still does not improve the loot, as I can now use the opalo I used to use at 10/10, and evade the attacks, so take no decay on the pixie.
I used to love Snabs, 2 years ago, now they are a waste of time.
Armour decay would explain it all.

Now, I take no armour decay, as they cant hit me.
 
Back
Top