Help: What's your interpretation? (take 2)

[Immortal]

​

I'm interested to hear what others' interpretation of the above graph(s) is. I put it in another thread as a curiosity, but there seem to be so many interpretations I am unsure where to start. Basically I want to do some more testing (I feel I am done with CP for now, even though I said I would stay ;p)
The setup was a 4.26 eco (equivalent to using MM @ Level 85 HIT, 85 DMG, or aimk2@lvl 65HIT 65 DMG)

The options are:
1) This supports TE thinking: using a higher eco weapon (such as maxed MM) would justify saving 190ped (1%) over the entire run, so your nett loss on TT would be 9%, not 10%.
2) This supports EM thinking, in that it doesn't matter with what weapon you process the TT, you will get the same loot return - which is why you received the standard 90% return on TT.
3) This shows that the TT spend from your weapon got converted into loot - no TT from fap or armor was returned
4) It shows that (L) weapons have a loot cap of 90% on TT, regardless of eco
5) It shows that (L) weapon you used has a loot cap of 80% of total expenditure (it ignores TT_spend, but includes the markup on the weapon).
6) It shows that your HIT/DMG with this weapon aren't high enough to gain loot. Even though it is (L) and maxed, the HIT/DMG levels affect your loot return independent of this.
7) The markup value on your skills gained makes up the difference between your spend + loot, therefore you have broken even. Even though your pedcard heads to zero, your nett worth (including markup and skills) increased more than your total spend.
8) Something else...


Data

	spend	loot	globals	wmu	Spend without fap+armor	global %	loot/TTcost	TTspend/markup	cumulative spend	cumulative loot	cumul spend w/m/u	cumu. Spend-fap-armor	cumul loot-spend	cumul. Loot-spend*mu	Return on TT	Return on m/u
1	1220.04	1250	1022	1377.1628	1090.56	0.818	1.025	0.886	1220.04	1250	1377.1628	1090.56	29.96	-127.1628	1.024556572	0.907663204
2	1273.72	1137.29	499	1442.6321	1146.56	0.439	0.893	0.883	2493.76	2387.29	2819.7949	2237.12	-106.47	-432.5049	0.957305434	0.846618313
3	1284.53	896.11	379	1431.978	1138.73	0.423	0.698	0.897	3778.29	3283.4	4251.7729	3375.85	-494.89	-968.3729	0.869017466	0.772242563
4	1267.89	913.79	585	1442.1942	1124.33	0.640	0.721	0.879	5046.18	4197.19	5693.9671	4500.18	-848.99	-1496.7771	0.831755902	0.737129303
5	1254.58	857.74	250	1417.4188	1107	0.291	0.684	0.885	6300.76	5054.93	7111.3859	5607.18	-1245.83	-2056.4559	0.802273059	0.710822064
6	1352.91	1361.63	951	1556.0817	1121.65	0.698	1.006	0.869	7653.67	6416.56	8667.4676	6728.83	-1237.11	-2250.9076	0.838363818	0.740303892
7	1296.28	1158.81	715	1475.4904	1115.52	0.617	0.894	0.879	8949.95	7575.37	10142.958	7844.35	-1374.58	-2567.588	0.846414784	0.746860038
8	1214.57	1454.88	959	1371.4994	1106.48	0.659	1.198	0.886	10164.52	9030.25	11514.4574	8950.83	-1134.27	-2484.2074	0.888408897	0.784253195
9	1255.72	930	274.33	1423.8508	1120.66	0.295	0.741	0.882	11420.24	9960.25	12938.3082	10071.49	-1459.99	-2978.0582	0.872157678	0.769826305
10	1288.84	900	440	1462.1671	1146.42	0.489	0.698	0.881	12709.08	10860.25	14400.4753	11217.91	-1848.83	-3540.2253	0.854526842	0.754159135
11	1267.9	1300	870	1561.7624	1003.7	0.669	1.025	0.812	13976.98	12160.25	15962.2377	12221.61	-1816.73	-3801.9877	0.870019847	0.761813615
12	1116.81	673.76	270	1271.401419	946.55	0.401	0.603	0.878	15093.79	12834.01	17233.63912	13168.16	-2259.78	-4399.629119	0.850284123	0.744706902
13	1183.7	1665.96	1132	1327.1936	1098.3	0.679	1.407	0.892	16277.49	14499.97	18560.83272	14266.46	-1777.52	-4060.862719	0.890798889	0.781213333
14	1217.94	1173.62	549	1373.672	1108.02	0.468	0.964	0.887	17495.43	15673.59	19934.50472	15374.48	-1821.84	-4260.914719	0.895867664	0.786254297
15	1267.61	1184.6	643	1418.0232	1126.6	0.543	0.935	0.894	18763.04	16858.19	21352.52792	16501.08	-1904.85	-4494.337919	0.898478605	0.789517291
16	1210.1	984.73	389	1329.2935	1076.88	0.395	0.814	0.910	19973.14	17842.92	22681.82142	17577.96	-2130.22	-4838.901419	0.893345763	0.786661691

 

[~CHOPPER~]

There is no "I could care less" option.

 

[jjmatrix]

I usually can break even if I calculate skill gain, and if I would sell everything at markup (which I don't); also you have to take into consideration globals and HOFs which most times it's like a bonus, and it can even up a few bad runs.

 

[Executor]

There is no "Loot is random" option.

 

[Axxel]

Nice experiment + rep

 

[Jimmy B]

Nice experiment Imm.

Difficult to draw conclusions from one set of data though. Certainly you've highlighted some possible conclusions, but data from different set ups is needed to distinguish between them really.

90% return on tt looks possible from this, but wasn't Buck's experiment showing tt profit?

Also interesting to see the closeness of the loot and weapon expenditure lines.

So it'd be interesting to see a re-run where you cut your fap and armor expenses down significantly (or increase them significantly). It'd also be interesting to see a re-run with near-maxed imk2.

 

[Xen]

Go hunt the biggest mob you can with the biggest armor/fap you can use, sometimes even get reckless by letting more than one on you if possible.

Then go do the same thing naked with a hired mod fapper.

Then look at your results.

 

[Jimmy B]

Indeed, that'd be interesting Xen.

 

[Skaffen]

Snakes n Ladders

Indeed, that'd be interesting Xen.

Interesting none the less. Results are heavily suggestive of a regulating mechanism within the loot system, which is to say there is less game here than there ought. - I'm sure most have witnessed this as the 'balancing loot' shortly after log in following a bad hunt (not often, but noticable as 'untypical' after a few years)...and this is why EU is not even gambling.

 

[Fade]

Reality can be so discouraging.

 

[Immortal]

There is no "Loot is random" option.

Well loot doesn't appear to be random (what from the data provided gives you the indication it's random?) and a conclusion "loot is random" doesn't help in furthering an analysis.

Nice experiment Imm.

Difficult to draw conclusions from one set of data though. Certainly you've highlighted some possible conclusions, but data from different set ups is needed to distinguish between them really.

With this I used the highest eco, highest damage available. There are about three weapons that I can use that would give me better eco, all of which require about a million ped in skills to use properly. I am looking at long-term data like this to guage "what can the average player do to improve their returns". If TE is true, that discriminates against those who can't get those weapons (99.99% of playerbase). That's one major reason I don't like it, and proposed the EM model(I didn't have figures to back it up though), coupled with the fact MA releases ul weapons with lower eco than (L).

90% return on tt looks possible from this, but wasn't Buck's experiment showing tt profit?

Yes. Same amount of TT, different returns - I believe my weapon eco is higher also. All in all it's a tiny tiny loss on TT

Also interesting to see the closeness of the loot and weapon expenditure lines. So it'd be interesting to see a re-run where you cut your fap and armor expenses down significantly (or increase them significantly). It'd also be interesting to see a re-run with near-maxed imk2.

Go hunt the biggest mob you can with the biggest armor/fap you can use, sometimes even get reckless by letting more than one on you if possible.
Then go do the same thing naked with a hired mod fapper.
Then look at your results.

This was the biggest mob I can hunt - krelt/aurli d1,d2,d4 (to chaser). fap cost in d4 was 45-50ped per compared to 10-20ped in d1/d2. Commonly had multiple big aurli. Not possible naked (they hit way too hard+fast)
Prob with modfapper is they would have to be online all the time I am... for a minimum 50hours (that's how many hours hunting this was)
as for maxed imk2...hmmm... anyone wanna donate 1M ped?
I have two choices of weapon change that have significant more eco: adapted monty or mod korss. These weapons cannot be used to hunt CP, however I can compare on ambu. Closest ul on that have similar damage and avaialble are: Mad IV (forget it, eco=3.53;2.4) ... and everything else is less than 60% of my current setup.). So that's a no on ul weps.

I usually can break even if I calculate skill gain, and if I would sell everything at markup (which I don't); also you have to take into consideration globals and HOFs which most times it's like a bonus, and it can even up a few bad runs.

Yes I was surprised to find that since my time in entropia I am ahead due to skill gain markup. Pedcard doesn't look healthy though The data provided includes loot from all globals, in total
2xhl6
3 small esi's
about 500ped after markup worth of components.
The data includes loot from globals (96 @ 9927ped)
originally posted in this thread:
https://www.planetcalypsoforum.com/forums/hunting/121533-hunting-statistics-again-7.html

I think I will go ambu NE eudoria, then neas, and I can compare returns on taxed land. I might do hoggs after that (those I can do with a fapper, but.... 50 hours of continuous hunting

My ideal is to establish a base signal for loot, and be able to compare it between mobs, regen etc. May even start on skills and recrods eveything there too (I am sure my evade went up significantly...certainly my health jumped a point)

 

[Xen]

Interesting none the less. Results are heavily suggestive of a regulating mechanism within the loot system, which is to say there is less game here than there ought. - I'm sure most have witnessed this as the 'balancing loot' shortly after log in following a bad hunt (not often, but noticable as 'untypical' after a few years)...and this is why EU is not even gambling.

I don't think these results are heavily suggestive of a regulating mechanism. You will get the same results from a slot machine, which is pure chance with the odds set up to return a certain %.

However, that doesn't mean there isn't a regulating mechanism.

 

[Jimmy B]

With this I used the highest eco, highest damage available. There are about three weapons that I can use that would give me better eco, all of which require about a million ped in skills to use properly. I am looking at long-term data like this to guage "what can the average player do to improve their returns". If TE is true, that discriminates against those who can't get those weapons (99.99% of playerbase). That's one major reason I don't like it, and proposed the EM model(I didn't have figures to back it up though), coupled with the fact MA releases ul weapons with lower eco than (L).

Don't know why you quoted me there, as your response isn't relevant to my quote above it at all.

All I said was that to distinguish between your suggested conclusions, more than one data set is needed. That is still true, regardless of what eco you were shooting with.

As to the TE stuff, the people who've spent tens of thousands of dollars on weapons like imk2 in order to benefit from its great economy may disagree with your stance

 

[Skaffen]

Oxy Moron

I. ..., which is pure chance with the odds set up to return a certain %.

This is a contradictory statement. Consider the difference between poker and the slots. What gamers want is the calculation of poker with a set rake, not preset (hidden) instructions with a minimal payout which is prompted by some hidden instruction.

It is true the 2nd can be made look like the 1st, and I believe this switch was made when the amp nerf hit.

 

[Witte]

I cant draw any conclusion from this dataset, maybe just a couple of exclusions. Although I encourage any research in this area, I already made assumptions from personal experience. Which is that loot is totally independent on the person looting or the weapon/equipment used. This opens the door to get better results than others, which in my view is also the reason the game is not treated as gambling. It also makes most sense in a design point of view to me.

Anyhow, if any other conclusions arise from these experiments I will surely accept those, so carry on .

 

[Skaffen]

Pure Dead Brilliant

... Which is that loot is totally independent on the person looting or the weapon/equipment used.

You don't see a correlation in this data?! It is beyond dispute. There is some doubt as to how the correlation comes about (mechanism) but, it is there.

 

[Witte]

You don't see a correlation in this data?! It is beyond dispute. There is some doubt as to how the correlation comes about (mechanism) but, it is there.

It is not possible make conclusions from just one dataset. A simply illustration with highly correlated data to prove my point:

 

[Skaffen]

TruthSeeker

It is not possible make conclusions from just one dataset. A simply illustration with highly correlated data to prove my point:

True, however this data invokes cost to kill and % return - so it is perfectly valid (and supported by similar 'experiments').

I know we don't want it to be true but it is getting hard to disguise it.

 

[Jimmy B]

LMAO, great post Witte.

True, however this data invokes cost to kill and % return - so it is perfectly valid (and supported by similar 'experiments').

I know we don't want it to be true but it is getting hard to disguise it.

Witte's post is perfectly valid too. You can just draw the wrong conclusion from it.

Imm has listed a whole host of possible conclusions of his experiment. They can't all be true, and without more data its impossible to deduce which one(s) is true.

That doesn't make it useless, it just means more datasets are needed. Maybe if falkao sees this thread he has some more data he can draw on.

 

[Witte]

True, however this data invokes cost to kill and % return - so it is perfectly valid (and supported by similar 'experiments').

I know we don't want it to be true but it is getting hard to disguise it.

From what I have read all the data is from just 1 setup. So the return % should stay constant when loot is independent of avatar or equipment. It should increase when you manage to kill the creature cheaper, or decrease when it cost more to kill. But maybe I misinterpreted the data?

 

[Immortal]

Don't know why you quoted me there, as your response isn't relevant to my quote above it at all.

Yeah I don't know either I think I did a bad snip

As to the TE stuff, the people who've spent tens of thousands of dollars on weapons like imk2 in order to benefit from its great economy may disagree with your stance

Yes. And the people who want to sell them do too. That goes for amps as well.

From what I have read all the data is from just 1 setup. So the return % should stay constant when loot is independent of avatar or equipment. It should increase when you manage to kill the creature cheaper, or decrease when it cost more to kill. But maybe I misinterpreted the data?

Yes where did people come to the opinion that I was trying to draw the conclusions... I asked people for an interpretation (which is sometimes a conclusion, understandably). Hence the huge numbers of options and the possibility that more than one can be correct.
e.g. if it means XXX, what test can i next run that could verify or disprove XXX
At this stage, I haven't even tested different weapon vs same mob
The more outstanding figure is that,for me, loot appears to follow weapon. If the next test shows it follows weapon expenditure, the next question is "does it follow ammo, or ammo+decay, or decay, or ammo+decay-amp_decay?
And use the simplest dataset imagineable. sigh...more screaming ambu)

 

[Shining]

personally, i think the following applies:

Loot =

0.9 x (ammo cost to kill that mob) plusminus 100%, where the minus part is 5% bigger than the plus part.

So, no matter how eco you kill a mob, there is a treshold which makes sure that on the long run you lose.

Globals and HoFs are randomly triggered events and fed by a pool.

IMHO

 

[davidgr1200]

I don't think it says much more than "The more you hunt , the more loot you get" and "It's very difficult to break even".
ENtropia is like thermodynamics - teh 3 laws:
1. You can't win
2. You can't break even
3. You can't get out of the game (Luckily the 3rd one doesn't hold for ENtropia )

 

[Immortal]

I had some more play with data and did a quartile analysis (like what I did with my fap data) I don't know what the technical term for it is, sorry

I want to see where this goes, so will stay on CP for a little longer

 

[Witte]

I am a bit confused what I am looking at now. What is a quartile analysis? From what I see here it seems your average ROI is increasing linearly? And the periodical behavior I see, is it purely caused by total spend or could it also be caused by time? Did you record any time data?

 

[Immortal]

I am a bit confused what I am looking at now. What is a quartile analysis? From what I see here it seems your average ROI is increasing linearly? And the periodical behavior I see, is it purely caused by total spend or could it also be caused by time? Did you record any time data?

(Q1+2*Q2+Q3)/4 is a smoothing function (i think I remember the name now) but basically is used to identify peaks and troguhs
and yes it looks periodic over weapon_TT spent ... or total_TT....or_markup so I chose the one with the least variance (weaponTT+amp+ammo)

Basically I have been away from EU for four days...so If the next dataset(s) is out of character it could be time dependent(and not TT_dependent_, or that the previous 20 just had this spooky trend...more tests I have to do. My loot analysis assumes that it is relative to TT spent (a way of measuring time?) so I'm sticking with that. Alternatively it can be looked at as "time between i repaired and restocked' with in non-calypseian time that varies from a few minutes to a day or so

 

[Executor]

Well loot doesn't appear to be random (what from the data provided gives you the indication it's random?) and a conclusion "loot is random" doesn't help in furthering an analysis.

What I meant by "loot is random" is:
- I do not believe loot is 90% of TT spent or any other variable that you search.
- I believe that each mob has a loot range, plus the globs, hofs etc.

As a result:
- The more you kill, the more you get loot.
- The more you want to kill, the more you need to spend.

Put these 2 lines in agraph and we have the one you show.


To use an extreme angle, I would say that if we were able to kill 1,000 mob with spending 0 PED (a not "decaying blade" for example), then I am sure we still would have loot.

According to your therories, we would have only no looters right?

 

[Immortal]

What I meant by "loot is random" is:
- I do not believe loot is 90% of TT spent or any other variable that you search.
- I believe that each mob has a loot range, plus the globs, hofs etc.

As a result:
- The more you kill, the more you get loot.
- The more you want to kill, the more you need to spend.

Put these 2 lines in agraph and we have the one you show.


To use an extreme angle, I would say that if we were able to kill 1,000 mob with spending 0 PED (a not "decaying blade" for example), then I am sure we still would have loot.

I think what you mean is
Loot=A±B
If that is the case 'B' is random, in that it is unceratain, but doesn't make A random (or uncertain)

If loot were not related to TT spent, then either we or MA go very broke very quick.

According to your therories, we would have only no looters right?

You will have to explain where you deduced that from.

 

[Witte]

I think what you mean is
Loot=A±B
If that is the case 'B' is random, in that it is unceratain, but doesn't make A random (or uncertain)

If loot were not related to TT spent, then either we or MA go very broke very quick.

I think it is related to global TT spend, not to personal TT spend. BTW, it would be interesting to get more details on this periodic behavior that is visible. Some questions that rise are is it the same on every spawn/mob? If not, is it possible to hop from peak to peak? And what about teams?

My personal experience is that you can stay in a constant peak, although it is not easy. This has been my loot philosophy for ages, and the results I get are good (meaning I don't need to deposit). Although I always thought these waves were time dependent, not TT spend dependent.

 

[Immortal]

I think it is related to global TT spend, not to personal TT spend. BTW, it would be interesting to get more details on this periodic behavior that is visible. Some questions that rise are is it the same on every spawn/mob? If not, is it possible to hop from peak to peak? And what about teams?

My personal experience is that you can stay in a constant peak, although it is not easy. This has been my loot philosophy for ages, and the results I get are good (meaning I don't need to deposit). Although I always thought these waves were time dependent, not TT spend dependent.

The loot was 'somewhat' in team i hunted maybe 600ped of the 20k inteam (not big i know)

and yes I want to see where the apparent periodicity goes...which is why I'm staying (pedcard sure is low with all this ultragay TT crap.... i really really think this is an ultra awesome waste of time with what I've been looting..the best thing is....heart oil? :O)

I'm almost inclined to say it's independent of mob+location...since I've been switching between d1,d2,d4

Or maybe the periodicity is a fluke. Or that loot isn't mob, but location specific.
(i like the theory of "you have gained experience in your looting ability")

.../me want bbb to compare his data for periodicity :O