Developer-Notes--3

[Pyhä Antero]

It seems that direct weapon cost (without decay) determines loot. Till now I've assumed that it is hp dmg done.

As it looks now, mobs do record dmg cost (ammo, without decay and misses) and pays out accordingly.

P.S. This might also expalin why I had unknown variability when trying to establish a loot model for hunting. The data I've used was from different weapons.

How about melee? And why you rule out the misses?

 

[falkao]

How about melee? And why you rule out the misses?

Results are preliminary and only a first step. I'm ruling out misses because it makes sense, but who knows.

 

[narfi]

The problem is that only the PED's of the pre-tagger are lost, hence a kill stealer will get lower loot but it is proportional to the dmg he did and hence for him it will remain the same.

Can you pleaded explain my post at the end of page 44 where I explained that with a considerable sample size I maintained a 400ish% return rate?

 

[ermik]

It seems that direct weapon cost (without decay) determines loot. Till now I've assumed that it is hp dmg done.

As it looks now, mobs do record dmg cost (ammo, without decay and misses) and pays out accordingly.

P.S. This might also expalin why I had unknown variability when trying to establish a loot model for hunting. The data I've used was from different weapons.

This conclusion is the same ive had for a long time now.

I cant even remember how many times ive preached to tired soc mates how they should realise how often they come back with loot around the same ammount as the ammo they went out with.

Yes , the yellow line MISS is not included in loot, gives support for the MA statement to make sure to be maxxed on HIT.

 

[ermik]

the obvious follow up would be if ammo determines loot how does melee work.

well melee could be 100% ammo, wich means with melee you should get a tiny bit better loot over time, but more decay on armor + fap.

This could be one of the reasons MA havent released melee amps, because they would become more effecient.

 

[tamlin]

It seems that direct weapon cost (without decay) determines loot.

I like this hypothesis, as it should be quite easy to to find supporting data (either way) by testing e.g. similar-damage laser vs. BLP weapons (excluding the very low decay BLP weapons like S30).

Another (relatively low-cost) test that could potentially be interesting could be UL vs. L on something like S30, CB5 or S50, as the UL versions decay more - even if just a little.

 

[bettlejus]

This could be one of the reasons MA havent released melee amps, because they would become more effecient.

Tell that to Joker, he always had good return with melee

Maybe it's the other way around, with melee you get compensated occasionally and MA haven't released amp for melee in order for us to not loose all of our peds if we don't have a 50K peds ped card

 

[remontoire]

Tell that to Joker, he always had good return with melee

Maybe it's the other way around, with melee you get compensated occasionally and MA haven't released amp for melee in order for us to not loose all of our peds if we don't have a 50K peds ped card

Have you actually done any melee hunting? Melee hunting does not significantly differ from ranged hunting in terms of offensive costs to returned loot.

If and how do enhancer TT is returned in loot is a different question entirely. And if you are using many tiers of high markup enhancers, you will have high losses.

 

[ermik]

Tell that to Joker, he always had good return with melee

Maybe it's the other way around, with melee you get compensated occasionally and MA haven't released amp for melee in order for us to not loose all of our peds if we don't have a 50K peds ped card

joker is not the best example here, huge dps sword smacked full with 3-600% mu enhancers grinding away fast hitting regen monsters isnt really what im talking about.

choose melee when you can assure your defensive cost will be low, since you will have the mob close to you longer time than with a pistol or rifle.

actually the periods ive used melee my returns have always been as most stable.

 

[falkao]

Can you pleaded explain my post at the end of page 44 where I explained that with a considerable sample size I maintained a 400ish% return rate?

Only 10 pages here, but I guess it is this post.

So you're doing only 25% of total dmg. The other hunters do 75% but per hunter less than 25%, otherwise you won't get the kill. With this you still achieve 85% TT on total expenses (yours plus rest) and this is indeed strange. It should be quite lower in normal mobs, close to your 25% assuming same relative ammo cost in your 25% dmg as in the remaining 75%.

All this should be further studied. Do LA's behave differently? Looks like your team hunting were looter gets all.

 

[rhogenbe]

I think that regardless of the weapon used the same loot mechanism is used. That would make sense because that way automatically switching weapons is also included.

I think that all the decay from the hits are the base on which the loot is based.

So this includes weapon decay, ammo decay, attachment decay (amps, scopes and lasers) including the extra weapon decay from enhancers. I do not know how the enhancer decay itself counts in this but that would be hard to measure since it is random and way below 1% of the total decay.

On non maxed weapons your skill is also of influence because this will influence the total amount of decay needed to kill a mob.

So a good test should include all these factors.

That leaves what happens with the decay generated from armor decay, FAP decay and the decay from missed shots.

It would probably be good to also have these data since it could be that part of that is also returned in the loot.

In the past i have done some tests with Swine Deluxe + Evil against M2930 ME + A204 on Hogglo youngs. I have lost the data points that i had but i remember clearly that that the Swine gave me on average about 45% more loot on a single kill which is about what you would expect based on the comparable eco of them. I have neither of them maxed and my profession levels on them also differ so it is hard to really compare that more then day that there is a substantial difference in kill cost and loot return between the 2.

 

[rhogenbe]

Only 10 pages here, but I guess it is this post.

So you're doing only 25% of total dmg. The other hunters do 75% but per hunter less than 25%, otherwise you won't get the kill. With this you still achieve 85% TT on total expenses (yours plus rest) and this is indeed strange. It should be quite lower in normal mobs, close to your 25% assuming same relative ammo cost in your 25% dmg as in the remaining 75%.

All this should be further studied. Do LA's behave differently? Looks like your team hunting were looter gets all.

This could be due to items or bigger TT valued stackables. Bigger HP mobs tend to make it difficult for 100% fair loot spread based on damage.

 

[Pyhä Antero]

All this should be further studied. Do LA's behave differently? Looks like your team hunting were looter gets all.

He doesn't use team at all. So it's about same test as "pre-tagger" test, but with different results? The kill cost of "non-tagger"/looter should be compared to (base) loot though.

 

[Fifth]

All this should be further studied. Do LA's behave differently? Looks like your team hunting were looter gets all.

Indeed, it's the most obvious explanation. Only problem, there's no team...

Over the last 4 months I have cycled over 300k ped of which less than 1/4 came from my own weapon. (the rest from support shooters NOT in a team with me.

I have in essence been kill stealing every mob during this period and still receive approximately the expected 85% tt returns on the overall Offensive costs of the entire team.

MA has worked with these subsystems recently, some things may have changed since 2009-2011.
Have a feeling there's something more in there than meets the eye..

Could it be that there's other (hidden, team-like) systems besides the old "manual" team and the new automatic teaming system "shared loot"?

 

[Joat]

Reading some more posts on the forum and I begin to wonder if we all use same client...

I used to help out my socleader with lowering hp on mob, not being in team
and both him and I ended up with terrible loot. Others experience someting
else doing this...

I don't get same return from mkV+eamp13 vs mkV+203, even thou' they have
same ammoburn and same total damage, different decay thou'.
Others experience that loot is based on ammo burned, others from damage done...

If I do something that only cost me a lot, like help others with healing, I most
often get a decent compensation in loot quite soon after I begin to do something
that gives me loot, while others don't experience this at all doing the same...

So again, do we really use same client?

 

[remontoire]

I like this hypothesis, as it should be quite easy to to find supporting data (either way) by testing e.g. similar-damage laser vs. BLP weapons (excluding the very low decay BLP weapons like S30).

Another (relatively low-cost) test that could potentially be interesting could be UL vs. L on something like S30, CB5 or S50, as the UL versions decay more - even if just a little.

Breer M3a(L), 400 ammo, 0.845 decay, 7.6 dps, 2.869 dpp - 17% decay share
Geotrek LP120 Niloticus(L), 200 ammo, 0.923 decay 7.6 dps, 2.853 - 31% decay share

So if really ammo got converted to loot, you would see significant loot differences between these two guns, much more than indicated by teh marginal eco difference. Indeed, LP120 should always only return horrible loot.

 

[Jimmy B]

If it proofs true that ammo is turned into loot, then dmg/pec might be useless and decay/cost might be more relevant.

Nice analysis. Surprising result to me, but does certainly look plausible.

I don't get same return from mkV+eamp13 vs mkV+203, even thou' they have
same ammoburn and same total damage, different decay thou'.
Others experience that loot is based on ammo burned, others from damage done...

Could be that ammo determines small loots whilst decay determines rare higher loots. Xen's tests were only 100 mobs long, I guess you are speaking of more long-term results.

Or it could be that amp decay is treated differently to gun decay. It'd be interesting to see 100 mob tests similar to Xen's on these two set ups.

 

[remontoire]

Reading some more posts on the forum and I begin to wonder if we all use same client...

I used to help out my socleader with lowering hp on mob, not being in team
and both him and I ended up with terrible loot. Others experience someting
else doing this...

I don't get same return from mkV+eamp13 vs mkV+203, even thou' they have
same ammoburn and same total damage, different decay thou'.
Others experience that loot is based on ammo burned, others from damage done...

If I do something that only cost me a lot, like help others with healing, I most
often get a decent compensation in loot quite soon after I begin to do something
that gives me loot, while others don't experience this at all doing the same...

So again, do we really use same client?

Maybe avatars are not really born equal. Maybe the avatar loot cycles affect it a lot.

100 kills with two weapons is not really a serious test.

 

[falkao]

Nice analysis. Surprising result to me, but does certainly look plausible.

indeed, but further tests are needed.

..
Or it could be that amp decay is treated differently to gun decay.

The capped amped test did indicate that. Amp decay seems to be treated like ammo as amp decay in mining = bomb.

Nevertheless, to test what's really happening we have to be carefully. When comparing two setups, mob regen/dps have to be considered as well. Hence a dataset needs to have single kills where loot value, number of shots, hits, misses, total ammo, total decay and total amp costs are recorded.

Moreover, empty rate of mob/maturity has to be known.

 

[Joat]

Maybe avatars are not really born equal. Maybe the avatar loot cycles affect it a lot.

100 kills with two weapons is not really a serious test.

Maybe some amount of avatars are set to test things out without knowing it...

To be a bit serious: Yeah, agree, 100kills are nothing. Should be done for weeks or even months.

 

[remontoire]

Or it could be that amp decay is treated differently to gun decay. It'd be interesting to see 100 mob tests similar to Xen's on these two set ups.

Well, we really need to establish how unamped weapon loot works, including melee, before we can tackle amps.

 

[wizzszz]

Over the last 4 months I have cycled over 300k ped of which less than 1/4 came from my own weapon. (the rest from support shooters NOT in a team with me.

I have in essence been kill stealing every mob during this period and still receive approximately the expected 85% tt returns on the overall Offensive costs of the entire team. (its in a 5% taxed land area)

Missed your post, narfi.
So this explains your "hunters needed threads"


However, your observation are pretty exactly what i have seen twice during merry mayhem:
TT return was tied to total ammo spent, regardless of who spent it.


And @everyone:

Shouldn't it be "tagers" or "pre-killers"?

"pre-tagging" simple does not make sense...

 

[LaraD]

indeed, but further tests are needed.



The capped amped test did indicate that. Amp decay seems to be treated like ammo as amp decay in mining = bomb.

Nevertheless, to test what's really happening we have to be carefully. When comparing two setups, mob regen/dps have to be considered as well. Hence a dataset needs to have single kills where loot value, number of shots, hits, misses, total ammo, total decay and total amp costs are recorded.

Moreover, empty rate of mob/maturity has to be known.

I wrote something about in in another thread:

Melee amps have to work different than ranged amps due to internal accounting. So new system code has to be written and tested. As MA is way understaffed that will take some time...and sure is pretty low on MAs priority list. So expect melee amps to arrive later than sooner

 

[remontoire]

Maybe some amount of avatars are set to test things out without knowing it...

To be a bit serious: Yeah, agree, 100kills are nothing. Should be done for weeks or even months.

I think more importantly, we need to have several avatars doing exactly the same test. Unless these all get the same result, we can't really say we have a result. At least as kill costs vs mobs go.

 

[tamlin]

Nevertheless, to test what's really happening we have to be carefully. When comparing two setups, mob regen/dps have to be considered as well. Hence a dataset needs to have single kills where loot value, number of shots, hits, misses, total ammo, total decay and total amp costs are recorded.

Moreover, empty rate of mob/maturity has to be known.

Another potential influence could be the avatar. Is it a "lucky" one (i.e. one in a good streak) or an unlucky one. Has recent loot been shit for one and he's on the way up from it, and are the other on its way down?

A potential idea for this could be to use a higher-level avatar that's been relatively stable in returns for some period willing to do the test, that would only grind the same mob with the same weapon for n kills, then immediately afterwards (with not even a mining drop or crafiting attempt in between) test the other weapon for the same n kills. This would probably require some dedication.

To not put undue economic burden on that avatar, a small fund could be created. Say if all interested pitched in x PED (let's say 10 PED, and let's call it buying loot-shares for the testing?) it could generate quite a substantial amount of hopefully quite valid data.

Just some random ideas.

 

[wizzszz]

A similar result was seen when using capped amps. Hence one might conclude that ammo is turned into loot but not decay. Hence, although loot is higher with DLxE, the quite high decay might not get compensated.

Avoiding weapons where the much of the costs are from decay is only easy for ranged weapons.
What about melee? Obviously not all decay is gone, but it is pretty much impossible w/o long term tests to tell how much of the decay determines loot and which is gone.

Which brings us back to my original question:
Where do the losses go?
Or is it all gone forever and MA does "tax" players so very differently based on weapon decay?

 

[tamlin]

[about Narfi's mercenaries]

However, your observation are pretty exactly what i have seen twice during merry mayhem:
TT return was tied to total ammo spent, regardless of who spent it.

As I have been part of that as doing support-damage (all at Narfi's expense, but that's nothing the system can know about), I can with almost 100% certainty dispel any rumors of personal loot pool and "system pays you back".

While I haven't spent multi-thousands of Narfi's money (yet ), I have gotten exactly zero deviation from my normal loot pattern when I hunt.

Had there been a "personal loot pool", or even a "checks and balances" system in place, I would have gotten way above average loot.

Applying Occam's razor; kill-stealing pays (as Narfi during these runs effectively work like an actual kill-stealer does). Looter takes it all (including money spent by all the others doing dmg).

There could ofc be other interactions not considered, but I think the simplicity of this makes it the most plausible.

 

[wizzszz]

[about Narfi's mercenaries]

As I have been part of that as doing support-damage (all at Narfi's expense, but that's nothing the system can know about), I can with almost 100% certainty dispel any rumors of personal loot pool and "system pays you back".

While I haven't spent multi-thousands of Narfi's money (yet ), I have gotten exactly zero deviation from my normal loot pattern when I hunt.

Had there been a "personal loot pool", or even a "checks and balances" system in place, I would have gotten way above average loot.

Applying Occam's razor; kill-stealing pays (as Narfi during these runs effectively work like an actual kill-stealer does). Looter takes it all (including money spent by all the others doing dmg).

There could ofc be other interactions not considered, but I think the simplicity of this makes it the most plausible.

Narfis (an your) observations do not allow this conclusion.


It seems Narfi has looted all the spent ammo out, so why should the system compensate you for something that has already been paid out (to Narfi)?

Moreover, the personal loot pool theory does not even come into play here, as the expected loot minus fee (90%) minus LA tax (=85%) has been fully paid out already anyway - or in other words: Your personal loot pool didn't increase, so why should it pay out compensations?


I agree with the "kill-stealing" pays conclusion though, but imo it would have been better to not find that out.
(or at least not post that in public)

 

[tamlin]

Narfis (an your) observations do not allow this conclusion.

I think I understand where you are coming from, and on examining my line of thought I have come to agree with this. Almost. There are still holes in the hypothesis.

It seems Narfi has looted all the spent ammo out, so why should the system compensate you for something that has already been paid out (to Narfi)?

Indeed, and that talks against "personal" pool. It's rather validation of a per-mob payback (which would be the simples implementation, which the well-known razor would be in favor of).

I agree with the "kill-stealing" pays conclusion though, but imo it would have been better to not find that out.
(or at least not post that in public)

While I can understand this view, as it can indeed increase the risk of more kill-stealing asshats short-term, it also hopefully raises the public's awareness of this problem. Only if enough people put (economic) pressure on MindArk will they ever even care to consider ordering someone to look into potentially designing a potential future change that could eventually maybe mature to the extent it would be implemented.

 

[rastamannen]

Well ive only tried different weapons and looked at globals and ive seen a pretty big difference depending on what weapon i use.

Ive used 2 different kind of models i have 10 in HA on both models.

So both guns has almost exactly the same dmg/pec but one has almost double dmg/sec compared to the other.
Ive cycled around 6k peds on both guns and ive havent tracked my returns but ive compared global sizes just for fun (I know that my returns are higher on the dmg/sec gun since ive made profit with it).

Gun with half dmg/sec: Average global size= 86,3 ped
Gun with doouble dmg/sec: Average global size= 193,4 ped (wich is quite high, had one HOF @ 1134 peds, without HOF its 143 peds average)

Have no idea if this means something, but it works for me so im happy
90% of the globals are on small trox.