Some utterly insane tests

[Jimmy B]

Originally Posted by Shaggy

So mabey the damage/pec of a gun isnt that important? Mabey using a UL gun at level 1 wont cause you to lose more ped on hunts? If ma does compansate with higher loots for being less efficiant is there a point to being efficiant?

It seems to me that finding the cheapest way to kill a mob results in a smaller loss of ped overtime but mabey im wrong. If there is a loot multiplier for people who spend more killing a mob, it isnt enought to make hunting with ep-50 at level 20 as eco as korss. I don't think its enought to justify hunting ambu with korss either but mabey I'm wrong about that too.

Well, like you say, I too also feel my long-term return is best when I've been concentrating on spending as little as possible to kill a mob. I dare say, certain weapons could be better than others for reasons aside from pure economy, and also even the armor you choose to wear, and hunting mobs that are 'in your level' could all also have an effect.

But yes, I was in the past happy that spending less to kill a mob was a good thing. I still am.

However, what has me a little concerned is this whole regeneration thing so I went to test it a bit. My concern is to do with markup. If I factor in dmg/sec and dmg/pec like we have done in falkao's thread then the natural thing to do is to use the weapon combo that kills the mob for the least spend (including markup). However, if loot is related to spend somehow that could make what appears to be the logical thing to do, not actually the best thing to do.

For instance, for some mobs spending 200% on a HL14 results in a cheaper kill than spending 150% on a H400. However, if loot is pushed up for the more expensive kill (eg. due to regeneration), then spending that extra markup may no longer be justified. That is unless somehow loot is related to total spend (including markup) rather than tt spend - given that markup is exchanged between players whereas player-MA interactions are tt-based, it seems more likely to me that if loot is related to spend it would have to be tt spend not markup. Could be wrong though.

[falkao]

Originally Posted by Jimmy B

..

However, what has me a little concerned is this whole regeneration thing so I went to test it a bit. My concern is to do with markup. If I factor in dmg/sec and dmg/pec like we have done in falkao's thread then the natural thing to do is to use the weapon combo that kills the mob for the least spend (including markup). However, if loot is related to spend somehow that could make what appears to be the logical thing to do, not actually the best thing to do.

...

I don't think total cost is attributed to loot, but only weapon cost, or better dmg done. With this you’ll have still an advantage with a faster weapon that does the same dmg. If my assumption is true, then the difference between a fast and a slow weapon comes from armor and fap cost.

[Moribund]

Originally Posted by Jimmy B

lol, feel free

I'll take one of those too.

[Shaggy]

So the next logical step would be to test if the markup of a weapon could effect the loot. Would that be the markup you paid, or the average markup then? Could you buy a Korss400 at 10k% markup and see a big increase over a korss400 at 150%? Seems like this would make it easy to cheat as you could have a friend list a korss at an insane markup, then buy it, and have him return the extra ped. If it is just the average markup then that would that make guns like LR59 look like better hunting weapons. Seems like it would be difficult to test since markup doent vary that much on most L guns. 100%-300% seems to be the range.

[Jimmy B]

Originally Posted by falkao

I don't think total cost is attributed to loot, but only weapon cost, or better dmg done. With this you’ll have still an advantage with a faster weapon that does the same dmg. If my assumption is true, then the difference between a fast and a slow weapon comes from armor and fap cost.

This could be the case yeah. However, in the case of the fast regenerating ambulimax the armor and fap costs are pretty small, around than 2% of the total cost. You're not going to recoup the extra markup costs solely by savings to armor and fap for these mobs. So then you'd be better off using the H400 even though it costs more to kill the mob than with the HL14.

Anyway, we still are far from proving that loot is proportional to damage done imo. I'm not convinced its the case at all.

Originally Posted by Shaggy

So the next logical step would be to test if the markup of a weapon could effect the loot. Would that be the markup you paid, or the average markup then? Could you buy a Korss400 at 10k% markup and see a big increase over a korss400 at 150%? Seems like this would make it easy to cheat as you could have a friend list a korss at an insane markup, then buy it, and have him return the extra ped. If it is just the average markup then that would that make guns like LR59 look like better hunting weapons. Seems like it would be difficult to test since markup doent vary that much on most L guns. 100%-300% seems to be the range.

I'd guess the average markup would be used, although I'm not entirely sure how it could work. I guess the best way forward for testing would be to use an expensive high decay gun and a cheap low decay gun. Collecting a few of the antimateriel guns should do the trick I'm not volunteering for that one

[-=rAndom=-]

I highly doubt MA counts with markup.
It's visible from how they handle "freeloading" (reducing H400 drop rate from feffs, more residue less gems...)

Originally Posted by Jimmy B

I'd guess the average markup would be used, although I'm not entirely sure how it could work. I guess the best way forward for testing would be to use an expensive high decay gun and a cheap low decay gun. Collecting a few of the antimateriel guns should do the trick I'm not volunteering for that one

cmon Jimmy!
you know you want to! :P

[Jimmy B]

Right did a couple more tests. Firstly, I shot ambus with the scrof+106 set up, shooting them as far as possible down to death then letting them regenerate. I did a set of 6 with two regenerations (roughly tripling cost of kill and damage done, to about 15 ped each). Then a set of 6 with one regeneration (roughly 10 ped each). I then went on a 'normal' ambu hunt with the scrof+106 and recorded the returns from the youngs. Data is shown below, with the previous results also included:



Seems to me that the data strongly suggests that loot is proportional to, or at least related to, cost to kill or damage done. The first test suggests that either there's a cap on this (possibly at about 4 times normal cost), or that the distribution is very different for huge HP killed. Seeing a sample of Daspletor Stalker loots would be interesting for comparison to help determine whether its a cap or a different distribution. A crazy armor test is needed to determine whether its based on damage done or on overall cost. Markup Weps and uneco weps would be interesting to test, but difficult to get sufficient multipliers on cost.

The loot from the normal hunt was pretty well spread out. There's a few small gaps (between 0.83-1.18 and 1.55-1.82) but other than that nothing really discernable. Biggest 'normal' loot was 4.29 then the minis were 19.28, 34.06, 40.35 and 43.24. Average of the loots below 5 ped was about 2 ped or roughly 40% of cost. Average of all loots was about 80%. Interesting that the minis made roughly half of loot, but too short a test to really read anything into that.

[Julius]

[quote=Shaggy;1218592]So the next logical step would be to test if the markup of a weapon could effect the loot. Would that be the markup you paid, or the average markup then?
[\QUOTE]

We know that MA tracks the markup through the AH sales, which is what gives us the nifty little "Market Value" tab... They would (obviously) have no way to easily track the declared value in grey/black markets like the PA shout-a-thon, which operate outside the "main" financial system. (There's a sociology project for someone; do people who come from countries/cultures with a strong history of black markets do better in EU? Hmm...)

This could definitely be another reason to manipulate prices in the marketplace; when you see a spike in resale prices for Korss 400, you will know someone is going out on a massive hunting spree!



Love this thread; massive gj to all. Wish I was knowledgeable enough in mathematical modeling to know what you were talking about all the time, but the parts I get are very interesting!

[falkao]

Originally Posted by Jimmy B

Right did a couple more tests. Firstly, I shot ambus with the scrof+106 set up, shooting them as far as possible down to death then letting them regenerate. I did a set of 6 with two regenerations (roughly tripling cost of kill and damage done, to about 15 ped each). Then a set of 6 with one regeneration (roughly 10 ped each). I then went on a 'normal' ambu hunt with the scrof+106 and recorded the returns from the youngs. Data is shown below, with the previous results also included:
...

seems there is some kind of penalization when killing the mob not in the shortest time. Payout for cost mult. 1 is about 80% (n=66), for the rest it is only 35% (n=24). For single multipliers it is 80%, 42%, 34%, 27% and 38%.
It seem that with higher multipliers, normal loot is reduced with relation to cost but on the other site there is more extra loot. So loot gets more variability and hence will become more casino like.

So my dmg done theory is on weak leggs.

[Jimmy B]

Originally Posted by falkao

seems there is some kind of penalization when killing the mob not in the shortest time. Payout for cost mult. 1 is about 80% (n=66), for the rest it is only 35% (n=24). For single multipliers it is 80%, 42%, 34%, 27% and 38%.
It seem that with higher multipliers, normal loot is reduced with relation to cost but on the other site there is more extra loot. So loot gets more variability and hence will become more casino like.

So my dmg done theory is on weak leggs.

I think the test is too short to be able to draw those conclusions to be honest. None of n=6, n=24 nor n=66 are big enough to say anything about total loot return.

If you ignore the first row, which could well be explained by the cap argument. The next 3 rows return about 32%. There's no minis or globals amongst them - if you pick 3 rows without minis or globals from the normal hunt you can find returns below 32%.