Help: Hunting loot

[falkao]

hunters joining the experiment so far:

m.Loot = mean Loot, m.Cost = mean Cost

Ava		n	m.Loot	m.Cost	Return Rate
Fairy		21	279	517	54.0%
Falk		20	417	464	89.8%
ForrestP		10	195	403	48.5%
Galadea		31	1197	463	258.5%
Janus		10	188	455	41.2%
Schaefer13	37	490	455	107.5%
Jez		49	404	491	82.4%

Total		178	523	472	110.8%

#	Ava	ID	Loot	Shots	HP
1	Janus	1	211	51	1520
2	Janus	2	283	51	1520
3	Janus	3	85	51	1520
4	Janus	4	128	51	1520
5	Janus	5	385	51	1520
6	Janus	6	118	51	1520
7	Janus	7	177	51	1520
8	Janus	8	285	51	1520
9	Janus	9	232	51	1520
10	Janus	10	290	51	1520
11	Falk	1	239	17	1360
12	Falk	2	290	23	1360
13	Falk	3	124	18	1360
14	Falk	4	307	20	1360
15	Falk	5	202	21	1360
16	Falk	6	122	24	1360
17	Falk	7	3770	25	1360
18	Falk	8	378	25	1360
19	Falk	9	170	18	1360
20	Falk	10	523	22	1360
21	Falk	11	75	17	1360
22	Falk	12	266	19	1360
23	Falk	13	325	21	1360
24	Falk	14	129	17	1360
25	Falk	15	261	26	1360
26	Falk	16	396	21	1360
27	Falk	17	333	20	1360
28	Falk	18	69	23	1360
29	Falk	19	374	21	1360
30	Falk	20	370	28	1360
31	ForrestP	1	107	22	1336
32	ForrestP	2	265	22	1336
33	ForrestP	3	185	22	1336
34	ForrestP	4	118	24	1336
35	ForrestP	5	127	23	1336
36	ForrestP	6	323	24	1336
37	ForrestP	7	169	22	1336
38	ForrestP	8	117	21	1336
39	ForrestP	9	249	22	1336
40	ForrestP	10	348	23	1336
41	Galadea	1	168	33	1300
42	Galadea	2	275	34	1300
43	Galadea	3	268	37	1300
44	Galadea	4	11293	44	1300
45	Galadea	5	294	43	1300
46	Galadea	6	289	34	1300
47	Galadea	7	146	38	1300
48	Galadea	8	120	38	1300
49	Galadea	9	1713	40	1300
50	Galadea	10	245	31	1300
51	Galadea	11	269	36	1300
52	Galadea	12	111	29	1300
53	Galadea	13	194	32	1300
54	Galadea	14	181	29	1300
55	Galadea	15	170	30	1300
56	Galadea	16	318	29	1300
57	Galadea	17	209	43	1300
58	Galadea	18	272	27	1300
59	Galadea	19	273	38	1300
60	Galadea	20	185	30	1300
61	Galadea	21	205	39	1300
62	Galadea	22	200	31	1300
63	Galadea	23	116	34	1300
64	Galadea	24	174	31	1300
65	Galadea	25	115	38	1300
66	Galadea	26	323	28	1300
67	Galadea	27	210	41	1300
68	Galadea	28	286	34	1300
69	Galadea	29	417	33	1300
70	Galadea	30	132	38	1300
71	Galadea	31	17935	64	1300
72	Fairy	1	469	48	1430
73	Fairy	2	240	66	1430
74	Fairy	3	181	40	1430
75	Fairy	4	221	45	1430
76	Fairy	5	213	43	1430
77	Fairy	6	154	44	1430
78	Fairy	7	187	43	1430
79	Fairy	8	75	48	1430
80	Fairy	9	259	46	1430
81	Fairy	10	1633	41	1430
82	Fairy	11	182	38	1430
83	Fairy	12	222	37	1430
84	Fairy	13	344	37	1430
85	Fairy	14	275	37	1430
86	Fairy	15	134	44	1430
87	Fairy	16	284	39	1430
88	Fairy	17	274	39	1430
89	Fairy	18	359	32	1430
90	Fairy	19	178	32	1430
91	Fairy	20	274	38	1430
92	Fairy	21	285	35	1430
93	Sch13	1	188	21	1265
94	Sch13	2	187	17	1265
95	Sch13	3	108	20	1265
96	Schaefer13	4	227	20	1265
97	Schaefer13	5	255	27	1265
98	Schaefer13	6	173	18	1265
99	Schaefer13	7	439	26	1265
100	Schaefer13	8	119	20	1265
101	Schaefer13	9	166	20	1265
102	Schaefer13	10	561	23	1265
103	Schaefer13	11	228	18	1265
104	Schaefer13	12	407	25	1265
105	Schaefer13	13	244	21	1265
106	Schaefer13	14	255	24	1265
107	Schaefer13	15	110	22	1265
108	Schaefer13	16	245	18	1265
109	Schaefer13	17	121	23	1265
110	Schaefer13	18	215	19	1265
111	Schaefer13	19	312	21	1265
112	Schaefer13	20	111	20	1265
113	Schaefer13	21	109	19	1265
114	Schaefer13	22	349	25	1265
115	Schaefer13	23	113	24	1265
116	Schaefer13	24	127	22	1265
117	Schaefer13	25	183	21	1265
118	Schaefer13	26	350	19	1265
119	Schaefer13	27	219	18	1265
120	Schaefer13	28	123	22	1265
121	Sch13	29	225	25	1265
122	Sch13	30	165	19	1265
123	Sch13	31	6132	19	1265
124	Sch13	32	114	26	1265
125	Sch13	33	3632	29	1265
126	Sch13	34	359	24	1265
127	Sch13	35	116	26	1265
128	Sch13	36	266	24	1265
129	Sch13	37	386	22	1265
130	Jez	1	1742	26	1412
131	Jez	2	66	28	1412
132	Jez	3	376	30	1412
133	Jez	4	365	29	1412
134	Jez	5	328	28	1412
135	Jez	6	381	27	1412
136	Jez	7	245	27	1412
137	Jez	8	138	31	1412
138	Jez	9	1939	31	1412
139	Jez	10	178	25	1412
140	Jez	11	60	38	1412
141	Jez	12	138	28	1412
142	Jez	13	193	28	1412
143	Jez	14	286	28	1412
144	Jez	15	325	25	1412
145	Jez	16	257	26	1412
146	Jez	17	423	27	1412
147	Jez	18	74	26	1412
148	Jez	19	360	26	1412
149	Jez	20	249	26	1412
150	Jez	21	170	24	1412
151	Jez	22	3879	36	1412
152	Jez	23	144	36	1412
153	Jez	24	136	30	1412
154	Jez	25	75	31	1412
155	Jez	26	181	35	1412
156	Jez	27	114	30	1412
157	Jez	28	178	33	1412
158	Jez	29	276	43	1412
159	Jez	30	181	28	1412
160	Jez	31	324	24	1412
161	Jez	32	301	24	1412
162	Jez	33	235	26	1412
163	Jez	34	191	28	1412
164	Jez	35	351	29	1412
165	Jez	36	139	31	1412
166	Jez	37	196	32	1412
167	Jez	38	187	23	1412
168	Jez	39	116	25	1412
169	Jez	40	387	25	1412
170	Jez	41	338	25	1412
171	Jez	42	248	28	1412
172	Jez	43	191	29	1412
173	Jez	44	191	27	1412
174	Jez	45	174	24	1412
175	Jez	46	204	34	1412
176	Jez	47	355	27	1412
177	Jez	48	3687	30	1412
178	Jez	49	252	21	1412

----
One of the reasons why I’m analyzing loot is because information is missing. I do not understand why MA is not sharing the return rate for example. Nevertheless, there are ways to find that out but help is needed.

I started a while back in collecting Hogg data. I do have only limited time to play, so it took about 5 months to get a sample of 851 Hogg kills. I did share this data and its findings in my Loot thread. The intention of this thread was to show basic analysis techniques plus a glossary of statistical terms. I know the thread is hard to follow and understand if background is missing, but at least the main findings should be understood.

What are the open issues from my point of view.

Class weights

We have a basic understanding of loot classes. We know for example the distribution per class and can calculate expected loot given the mobs hp. However, this is not the full story yet. What is missing are class weights. We already know that class weights do reflect in some way the dmg a mob can do, but we have only limited data about it. So one of the next things is to understand from what those weights do depend. What data is needed?

I do already have a hogg young set. So I know the weights for hogg young. It would be very helpful to have a set on some 500 and 1000 HP mobs. Ambu, Drones or Feffs for example. It would be sufficient to know the number of killed mobs and the number of loot above base loot (HP/20 + 4*HP/10).


Validated Return rate

We have quite some good understanding about kill costs. Jimmy provided some excellent formulas. However, we don’t have datasets from different avas we can compare and validate our assumption. I know my personal return rate on Hoggs for example, but I’m not sure if another ava with same equipment and skills would get the same result. So we have to validate our findings otherwise it’s a simple finding. What data is needed?

Hunting set for a specific mob from different avas including total number of kills, total cost and total loot.


Hunting simulations

If we’re able to estimate the cost to kill with different equipment and do know expected loot, then we’re able to simulate different hunting strategies.


Data collection

Data can be collected by a single ava or a group of avas. To know class weights we only need a set of kills with respective loot. For some questions it would be irrelevant if 100 kills are collected by one ava, or 100 avas take note of only one kill they did. Especially with loot class weights this would be possible. This would facilitate data collection a lot. So I hope we find some volunteers.


Let me know what you do think about it?

 

[Snabble]

ppppppppppppp

 

[falkao]

From what i have seen, I will never be able to advance my loot theory, past that.

that's maybe your destiny?

 

[Ianus]

If we’re able to estimate the cost to kill with different equipment and do know expected loot, then we’re able to simulate different hunting strategies.
Let me know what you do think about it?

Well apart from decay we can easily base on info from entropedia. For example ive calculate based on info found there that an atrox young will regenerate around 160 HP if hit with a p5a. Being way into 3hp rule where loot is increased linearly that should demonstrate for ex that p5a is more then enough for those you dont need an apis or x5 because you pay to much markup for the same results. Would be interesting if someone can come with a tool to compare different wep vs regen mobs, i cant though.

 

[falkao]

Well apart from decay we can easily base on info from entropedia. ..

sure, but that's not exactly what I'm after. I'm looking for a model that explains the observations and there are some unresolved issues. For example: do skills only have an influence on hit and evade or do they directly influence loot?

 

[AJack10600]

It would be very helpful to have a set on some 500 and 1000 HP mobs. Ambu, Drones or Feffs for example. It would be sufficient to know the number of killed mobs and the number of loot above base loot (HP/20 + 4*HP/10).

Can you please elaborate on what data exactly you need... I don't get it ...
sorry

 

[ForrestPark]

Hi Falkao,

If you can define *exactly* what we need to do, I think a lot of people would help.

If we need to, say, kill 50 ambu youngs and record (a), (b) and (c) variables before and after, that would be helpful to us.

Let's try to define exactly what we need to record, then once everyone agrees exactly what needs to be recorded we can get out there

+rep by the way.

Forrest

 

[ruskea]

After killing thousands and thousands of mobs

This is my loot theory

Sometimes the mobs i kill have loot
Sometimes the mobs I kill, do not have loot

From what i have seen, I will never be able to advance my loot theory, past that.

mostly its like this..

one mob from millions of killed might give u a hof
rarely the mob gives loot enuf to u to break even
sometimes u only lose about 30% on the run
almost always u get back less than 50% of the ammo u burned

 

[falkao]

Can you please elaborate on what data exactly you need... I don't get it ...
sorry

sure. What I'm looking for are loot class weights. Loot is made up of loot classes and mean loot can be expressed as

mean loot = C0 + w1*C1 + w2*C2 + w3*C3 + ... + w7*C7

where wi are the weights and Ci are the class means.

Unfortunately weights wi are hard to estimate because from w3 on they are rather low. So quite a large sample is needed to get a more or less precise estimate. Assuming a weight of 5% and having a sample of 1000 kills, error is still about 1.4% and so quite large. With n = 10000 error would be .14%. Therefore, its nearly undoable for me to collect all this data and hence I'm looking for help.

If we get together say 100 avas with 10 kills per ava we would have a sample of 1000 kills in rather short time.


To estimate weights I don't need to know every single loot. It would be sufficient to know how often a class was hit. This could be further simplified in limiting weight estimation to say weights > w1.

A good estimate for the upper limit of loot in C0 + C1 would be 1/2 HP. Hence, when I know how often loot above 1/2 HP was triggered I can estimate the sum of weights above w1. Furthermore, for weights above w4 the tracker data can be used.

What I'm looking for are hunters willing to invest some time. It would be optimal to get a dataset or some datasets for different mobs with single detailed loot. With 10 kills per ava this should be doable, but I doubt that we get together 100 avas.

 

[ruskea]

sure, but that's not exactly what I'm after. I'm looking for a model that explains the observations and there are some unresolved issues. For example: do skills only have an influence on hit and evade or do they directly influence loot?

well by looking my data the hunt loot has only gone worse while getting more lvls and maybe ur skipping the thing that "it matters who loots the mob" = its avatarbased.. there is a player A who hunts 2 days of high hp mobs (osses and malcrus) without getting a single global.. then there is player B who says in a soc chat that "ill come and get a global to piss u off" and comes to ur area and kills one! of the mobs uve been hunting 2 days and scores a global...

it might be ur X = Yx2 > Z but it also could all be avatarbased.. why make it so hard when it can be made so easy.. coz i dont think a week old noob has figured out all the areas that are good for him at that time and coz that scores 5 globals with 5k ammo on every run

 

[ForrestPark]

What I'm looking for are hunters willing to invest some time. It would be optimal to get a dataset or some datasets for different mobs with single detailed loot. With 10 kills per ava this should be doable, but I doubt that we get together 100 avas.

You're the boss on this one. Pick a mob. Tell us what we need to record. I don;t want to kill 1000 chirpies and then be told I was the only collecting Chirpy data

 

[falkao]

..I don;t want to kill 1000 chirpies and then be told I was the only collecting Chirpy data

lol, thx for the proposal ForrestPark

I've got very precious datasets in the past from Jimmy, Kolobok, Woody, Steffel to name some of them. They invested a lot of time to collect it and as mentioned above, also having 4-5 avas providing data is not enough. Furthermore, there are quite some threads about runs on camped mobs, but unfortunately data does not contain that much info. Hence I'm trying to organize it differently.

The optimal would be to get single loot data. This is very time consuming, but limiting it to 10 kills per ava would help.

Let's see how many avas we can get together and decide then what to collect.

 

[ForrestPark]

Okay well I am in.

If we can make it a mob (or small number of mobs to give some choice) that a lot of avatars can kill and can be found near a TP that would be handy.

Do the mobs all need to be the same maturity?

 

[Ianus]

Fine Im in too
How about feffs at fury, is a popular place. So I kill 10 feffs and then detail value of loot for everyone. Does maturity counts, for ex i have to kill 10 raiders or doesnt matter?

 

[ForrestPark]

Do the loots need to be in the same EU session? So if I log out, do I need to start again?

Can we kill them with any weapon, any armor, any fap etc etc? Or should we do Exa young with no armor, no fapping during the kill and unamped opalo (so we are all doing *exactly* the same thing)....?

 

[dbelinfante]

Data can be collected by a single ava or a group of avas. To know class weights we only need a set of kills we respective loot. For some questions it would be irrelevant if 100 kills are collected by one ava, or 100 avas take note of only one kill they did. Especially with loot class weights this would be possible. This would facilitate data collection a lot. So I hope we find some volunteers.


Let me know what you do think about it?

I might start collecting data. I hunt Hoggs regularly. But you haven't made it clear (to me) what it is exactly that you want/need.

 

[falkao]

to answer some questions:

The mob we choose should have the same maturity, or at least data should contain maturity.

Session time and equipment is irrelevant atm.

I would start with a popular mob with HP above 500 HP. If we take Ambu y for example we don't have problems with no loots. If we choose a mob that triggers empty loots as well, we have to take this into consideration. So the 10 killed mobs would be 10 killed mobs with loot plus the number of no loots.

 

[falkao]

I might start collecting data. I hunt Hoggs regularly. But you haven't made it clear (to me) what it is exactly that you want/need.

thx dbelinfante. We could use your data to compare it with my one. What I would need from you in case of Hoggs y is only globals data like the one in the following list:

example
ID #kill global value
1 20 75
2 43 120
…

ID is the global ID. #kill is the number of kills at which you've got the global and "global value" is its value.

 

[dbelinfante]

thx dbelinfante. We could use your data to compare it with my one. What I would need from you in case of Hoggs y is only globals data like the one in the following list:

example
ID #kill global value
1 20 75
2 43 120
…

ID is the global ID. #kill is the number of kills at which you've got the global and "global value" is its value.

Cool. I have a date with some Hoggs tonite. I might do two runs, should global a few times at least. Will keep track, and let you know the results.

BTW: Where do you extrapolate the global ID? EntropiaTracker?

 

[falkao]

BTW: Where do you extrapolate the global ID? EntropiaTracker?

Its your own number you assign to it like first, second, third etc. global.

 

[ForrestPark]

So let's say the mob is going to be Ambu Young.

We then record the TT value of loots from each of the first ten Ambu Youngs that give us loot (which should be the first ten mobs as they supposedly all give loot).

So I assume this should be the first ten we kill so it is a sound sample, not just any ten in a row....we are not trying to out-loot each other

If we die (for whatever reason including CTD) and need to go back, how does this affect the test? Does that mob count?

If we kill a Mature, with five youngs before it and five youngs after it, then we need to subtract the loot from the Mature from the total, yes, and this test is still valid?

So results-wise (and this is an example), we are looking at something like this then:

Mob: Ambu Young
ID PED
1 2.06
2 1.26
3 4.56
4 9.04
5 15.93
6 2.56
7 2.54
8 9.07
9 1.23
10 1.26

I will do mine tonight...

 

[falkao]

So results-wise (and this is an example), we are looking at something like this then:

Mob: Ambu Young
ID PED
1 2.06
2 1.26
...

that's correct. Don't count broken or regained mobs. It should also be stated what weapon and amp you used. If we want to do a regain correction due to different kill speed, number of shots per mob would be optimal otherwise total decay of weapon or amp would be sufficient.

so the list would look like

Mob: Ambu Young
ID PED Shots
1 2.06 25
2 1.26 31

 

[AJack10600]

I hunt ambuls quite often... will try to gather that data...

 

[falkao]

Fine Im in too
How about feffs at fury, is a popular place. So I kill 10 feffs and then detail value of loot for everyone. Does maturity counts, for ex i have to kill 10 raiders or doesnt matter?

would Ambu y be ok for you for the first round?

Data should look like

Mob: Ambu Young
state Weapon plus Amp

ID PED Shots
1 2.06 25
2 1.26 31
...

 

[steffel]

falkao,

as you know i am collecting mining data for you atm. since its easier (constant drop costs = kill cost) to track tt out and tt in maybe alter your request in 1st post? of course we'd imply that hunting and mining returns behave similar.

btw got 20 runs so far with 196 drops each, want me to send to you or wait till its more?

 

[falkao]

falkao,

as you know i am collecting mining data for you atm. since its easier (constant drop costs = kill cost) to track tt out and tt in maybe alter your request in 1st post? of course we'd imply that hunting and mining returns behave similar.

btw got 20 runs so far with 196 drops each, want me to send to you or wait till its more?

we should limit team data collection to hunting data and see how it goes, but I'm ready for your mining data. 196 drop seem reasonable to start with. At least I get a first impression and can tell you more.

 

[Noodles]

I'm in, and have a lot of data for the mobs less than 200 hp already. I can't really kill much above 300 hp. As you know, I keep promising to present the results, but have not yet done so. I have done some analysis, but nothing to my satisfaction. Where can I send you what I have?

I've also been tracking single mining loots for several months (unamped and 101 amped), and have quite a dataset there too. I also did some preliminary analysis of the results, which seems interesting so far, but not enough data points to understand the larger loots. Combined with Steffel's data, hopefully there will be enough points to work with.

Have also been keeping track of basic filter loots on condition. Hard to avoid death by boredom on that one . . .

 

[falkao]

I'm in, and have a lot of data for the mobs less than 200 hp already. I can't really kill much above 300 hp. .. . .

that sounds great. Low lvl mobs are not well analyzed yet. I would start atm with Ambus y (1010 HP) but switch then over to lower ones. I send you a pm with my e-mail.

 

[Ianus]

So for my part ive done the test unfortunately i didnt track pistol decay but it should be simple since i know exactly how much ammo ive used and that is 46.08 ped expense in ammo. Wep used was Korss 400 L with A102
1. 211
2. 283
3. 85
4. 128
5. 385
6. 118
7. 177
8. 285
9. 232
10.290
pec received

 

[Ianus]

So for my part ive done the test unfortunately i didnt track pistol decay but it should be simple since i know exactly how much ammo ive used and that is 46.08 ped expense in ammo. Wep used was Korss 400 L with A102
1. 211
2. 283
3. 85
4. 128
5. 385
6. 118
7. 177
8. 285
9. 232
10.290
pec received

and ofc mobs were ambu young