Question: How do I prove there is or is not a personal loot/expenses pool/tracker?

mrproper

Forum Beaver
Joined
Dec 5, 2005
Posts
6,274
Location
Romania
Society
Entropia Directory
Avatar Name
Andone mrproper Andrei
Please post your ideas and thoughts about how such an experiment should be conducted to provide cheap, effective and clear data that shows the highest bias between personal loot pool and no personal loot pool, that previous events influence or not future events in a significant manner, or that doing dumb things in an inefficient way is either penalized or rewarded while playing Entropia Universe.
 
I have some evidence against a personal loot pool. Mindark says there is no personal loot pool.
 
look at my tracker page, that is proof that there is no personal loot pool :laugh::laugh:

not even a tenth of what i have put into this game

still having fun though, or have some people forgotten to do that?

just my 2 pecs

regards,

PW
 
{removed} thank you and good bye.
 
Last edited by a moderator:
If you think MA would lie to you about this then it seems strange to trust them with your peds. Why haven't sold up and withdrawn all your money?
 
If you think MA would lie to you about this then it seems strange to trust them with your peds. Why haven't sold up and withdrawn all your money?

Because I play the game trusting a certain mechanism that they deny it exists and I do just fine and will continue to do so for as long as they don't change it. What is strange is how they can't ever clarify something and every developer note will contradict some of their previous statements and add to the clout of doubt.

Should I get convinced that EU is pure luck and that they have no mechanisms to conform to gambling laws and corrective rewards I should quit again.

I quit for a while when I found out they actually HAVE a corrective mechanism, when I found out ECO is a lie and they don't take decay as payment. Either way, I have more than I deposited and I will use it up to have fun and socialize with nice people.

Anyway, do you have any answer for my question?
 
Every object you have provided in the definition is a subject of interpretation:
  • Personal loot pool = ?
    90% tt return? ATH if the losses big enough?
  • Significant manner = ?
  • Dumb things in an inefficent way?
    shotting in air? using unmaxed tools? throwing mining bombs same spot?
If you're looking to prove there is ~90% tt return, take m2100 + e-amp 15 and hunt some smaller mobs (though this is kind of proven already).
highest bias between personal loot pool and no personal loot pool, that previous events influence or not future events in a significant manner, or that doing dumb things in an inefficient way is either penalized or rewarded while playing Entropia Universe.
 
Last edited:
First of all it's possible to get ~90 tt return without a personal loot pool based on this model:

For mobs it is easy for MA, they can easily work out the least cost to kill any mob. Then make a random generator around 90% of that as payback. That way they would not payout over 100% tt. So if you are inefficient you are losing peds quicker than an eco person

Please post your ideas and thoughts about how such an experiment should be conducted to provide cheap, effective and clear data that shows the highest bias between personal loot pool and no personal loot pool, that previous events influence or not future events in a significant manner, or that doing dumb things in an inefficient way is either penalized or rewarded while playing Entropia Universe.


It's possible to test your hypothesis...
The simplest, and most efficient way to do it (that I could come up with now ) would be comparing paired (paired T test) loot return over a month or so..
by paired i meant... you find 2 players at a smilar level of skills and weapon set up.... then start tracking their loot return after 1 of them get uber uber loot. if loot is personal and tracked by MA then the person who got the uber loot should have lower TT return than the person who didn't get any recent uber hof...
Assuming a cyclical pattern... the uber looter should have a steep fall in tt return than the other....
The size of peds spent in each block of hunt would depend on the mobs HP they'd be hunting!

You can also do the pairing of loot return of the same person over a given interval...and replicated it with many individuals.

I guess due to wide variations in loot.. you have to have many pairing... and the exact number of pairing needed would depend on the variablity and other factors...

You can have the data displayed various ways to show the differences..

You may want to start talking to him:
No need peds, and ofc I can't prove what I will write now, since I tracking just my TT whole balance. But just FYI:


02/28/2012 was my peak, since that I did lost 35k TT AFTER MARKUP.


1) 95% was planetary mining.
2) Amps: lvl5s, lvl7s and lvl8s (all self-crafted) - 7s and 8s was used PRIOR niksarium and duru rebalance, so I really know what I am doing, it's okay.
3) Something funny? 6500 UNAMPED drops in HELL, with 50% TT return in the end.
4) Planetary drops with amps - more than 100000.
5) Craft? Well, hard to say. All my amps, none was sold to someone.


There is not 10k ped cycled, as you can see. Also, those, who knows me, or those who can see my tracker (including history) - they know that I am "MU-sniper", so there is all ok with "efficiency".


DISCLAIMER (not for you, Artrat, just for some idiots):


I know what happening, and this is not whine, or something like that. I really know what I am doing, and that "loss" didn't even hurt me. So just STFU.
 
Last edited:
Every object you have provided in the definition is a subject of interpretation:

Thank you! And I'm not even a MA employee :)

First of all it's possible to get ~90 tt return without a personal loot pool based on this model

Yes it is. Excluded are the parts where you do stupid shit and continue to get 90%.

The simplest, and most efficient way to do it (that I could come up with now ) would be comparing paired (paired T test) loot return over a month or so..
by paired i meant... you find 2 players at a smilar level of skills and weapon set up.... then start tracking their loot return after 1 of them get uber uber loot. if loot is personal and tracked by MA then the person who got the uber loot should have lower TT return than the person who didn't get any recent uber hof...
Assuming a cyclical pattern... the uber looter should have a steep fall in tt return than the other....
The size of peds spent in each block of hunt would depend on the mobs HP they'd hunting!

You can also do the pairing of loot return of the same person over a given interval...and replicated it with many individuals.

I guess due to wide variations in loot.. you have to have many pairing... and the exact number of pairing needed would depend on the variablity and other factors...

You can have the data displayed various ways to show the differences..


My tests were mostly A/B comparisons, as in me being paired with me with different gear and alternating to detect bias, will see how I can do real pairing with teams. But you have a point. If one player gets a high loot, if we exclude that single loot from the data set, the player that got it should have on average worse returns than the other player.
 
My suggestion would be...

I think this is an interesting question. Please accept I am posting this not to be funny, but because I actually have thought about this carefully, and am still turning ideas over in my mind... But that I am coming from a fairly emotionally even keel with nothing invested in the outcome of this discussion. So, here we go - in the spirit of rational and logical conversation.

From the other thread I see a lot of mental conflict over the difference between an assertion that there are no personal loot pools and the generally well-observed 90% return rate average that we see across the system.

So I'd ask you to first confirm that you are trying to PROVE there is a personal expenses tracker OR correct the exercise to indicate if you are trying to PARTIALLY VALIDATE your previously observed 90% return rate...?

And then once you've thought about that, I offer you the follow-up question:

How can you differentiate your results to show the difference between a 90% return rate based on MA's previously stated "percent of decay they actually return to the system" (ie: an average return system based say solely on a pair of net running-total figures variable stored on a character - total decay and total loot value returned) and a system whereby Mindark is somehow keeping more detailed information on you, perhaps Per Planet, Per Mob or some other breakdown.

IE: my suggestion to you is: given we only see one output (loot returned on a hunting or mining run) we don't actually have the variables available to us to determine what the storage and return mechanism IS: we only know that it works, and that it tends to work "consistently on average".

And actually, that's all _I_ actually want to know.

There are, however, other questions I would humbly submit to you as being of far more interest to me (and maybe even you, and Jimmy, and Falkao, and all the other legends of research we have on here)...

Questions like:
MA have asserted
"avatar skills and efficiency on the tools used do indeed matter a great deal, and have a very significant effect on overall returns in all Entropia Universe professions."


So immediately, I want to know things like:
1) How much difference on overall returns does efficiency on each tool make?
2) Is there currently any substance to the general claim I see here and there which indicates that some weapons seem to return more items in loot than other weapons which return say more oils, or is that just subjective experience?
3) can you trigger a global on a smaller mob by blowing a lot on a monster you're awfully inefficient at by struggling with a few of the overpowered ones and then switching to the easier mob? And is this by design?

I would go so far as to say something like: maybe it is worth figuring out what the efficient equipment to hunt various creatures actually is and what skill levels you should actually have to have a reasonable stab at doing something "efficiently".

I'd actually like to know what causes MISS messages as well... because MISSes mess with efficiency something chronic. If it's network lag, server sync issues, or if it's really just a random chance that I miss 7 times in a row when there's only supposed to be a 9% chance I miss etc...

In short - I'm not sure we can answer your question with the tools and information we are supplied as participants, and to that end I am happy to accept MindArk's statements at face value... But I still find many things to be investigated based on what they are saying.

And I would be happy to assist in my rather feeble and humble capacity if there are in fact any other questions that arise for you and that require someone of my abilities to assist with.
 
How about this:
A - I hunt 50-100 mobs or a specific decay/ammo value and record the results as a sample point.
B - I do the same as A, except before the tests I drop in the same spot some probes worth 10-50% of the hunting costs.

Should there be a correction of my obviously stupid losses, the B samples should be slightly higher. Getting enough samples that any expected bias (10-50% difference) would be an order of magnitude more visible than natural loot variance (for example 30 reruns of A/B for 10%, 10 reruns of A/B for 50%). If there is no correction for my losses, then there is no personal loot pool or tracker. Should at least 2-3 other players repeat the experiment with the same results, then that should settle it.
 
Test Option

How about this:
A - I hunt 50-100 mobs or a specific decay/ammo value and record the results as a sample point.
B - I do the same as A, except before the tests I drop in the same spot some probes worth 10-50% of the hunting costs.

Should there be a correction of my obviously stupid losses, the B samples should be slightly higher. Getting enough samples that any expected bias (10-50% difference) would be an order of magnitude more visible than natural loot variance (for example 30 reruns of A/B for 10%, 10 reruns of A/B for 50%). If there is no correction for my losses, then there is no personal loot pool or tracker. Should at least 2-3 other players repeat the experiment with the same results, then that should settle it.

This method was mentioned in the 30+ pages of the main thread as having worked at one time. People were saying they blew up to a 1000 PED of ammo straight into the ground and then went hunting the next day to get HOFs or ATHs.

The poster also said they believe this no longer works, or would be surprised if it did.

I am wondering if part of the motivation for MA's post is that they have actually noticed this "past feature" of the loot system, and now fixed the hole, and POSTED A MESSAGE clearly stating that from now on, if you do stupid stuff like that, they are going to notice and not pay you out for it.

So... perhaps what the correct message to read from their post is; we have adjusted the loot system to include a little extra intelligence now which can detect more abuses than before.

So if they find a lot of mining drops without position changes, bullets fired and gun decay with any monster damage messages, etc... they can opt to take your decay without incrementing your "total decay done" counter (which is maybe the number they are using to help adjust towards your 90% average over time for instance).

Oh, I've rambled sorry. I was trying to say: this method of testing might work, but I am suspecting it's already been patched and so it will fail for you.

You're welcome to test it, but I suspect it will frustrate you.
 
Last edited:
What i would do if i had the time and spare peds, you never know i might as i am intrigued with MA's announcement

I would do two tests as follows:

To test if a server/personal loot pool exists
Drop 100 bombs in the same spot continuously, then drop 100 bombs normally, record results, rince repeat

If there is a memory, and you get back roughly the same, then to test whether it is server or personal:
Drop 100 bombs in the same spot continuously, move server drop 100 bombs normally, move server, rince repeat

If again you get the same rough tt results, ie between 80%-100% tt returns that would suggest you have a personal loot pool. If on the other hand, first is true but second is not, then you have a server memory. If you have neither then there is no memory.
)
Problems with this are various, but to name a few, other people on same server, no way of knowing if i would be the only one or not. And the volitile loot returns, would have to drop a significant amount of bombs, ie in the region of 10k drops to be statistically significant, (the maths i would leave to falkao ;) )

Just my two pecs

Rgds

Ace

EDIT: and i the person doing this, would not be able to do anything else, otherwise that "could" also screw results

EDIT2: you could do exactly the same with hunting, jsut keep to 50/50 waste/none waste
 
Last edited:
"avatar skills and efficiency on the tools used do indeed matter a great deal, and have a very significant effect on overall returns in all Entropia Universe professions."


So immediately, I want to know things like:
1) How much difference on overall returns does efficiency on each tool make?
2) Is there currently any substance to the general claim I see here and there which indicates that some weapons seem to return more items in loot than other weapons which return say more oils, or is that just subjective experience?
3) can you trigger a global on a smaller mob by blowing a lot on a monster you're awfully inefficient at by struggling with a few of the overpowered ones and then switching to the easier mob? And is this by design?
1) In theory, each mob, item and blueprint has two kinds of loot. Let's call them "success" loot and "residual" loot. Success in hunting is getting some "good" items, residual loots are ammo and crap oils. In mining, success gets you the rare stuff, residual finds get you lyst and oil by default. In crafting, success gets you the desired items, the residual loots are just dust and goo. But TT return seems constant, it's the markup and utility of the loots that changes as you become better at getting them.

2) If it works like I just explained above, then yes. Otherwise, higher tt items drop with higher loots.

3) Which brings us to this. High level hunters can get globals on caraboks and snables. Mob does not matter, but what seems to be their accumulated inefficiency from the bigger mobs. This is an indication of the tracking mechanism, how could 1000 noobs hunt 1000 punies every day, and an uber comes along after a bad FOMA run and gets the only weekly global? That is statistically improbable.

I plan to do number 3 consistently, but it would only work if there is a personal loot pool or a personal tracker that is carried into team setups.
 
This method was mentioned in the 30+ pages of the main thread as having worked at one time. People were saying they blew up to a 1000 PED of ammo straight into the ground and then went hunting the next day to get HOFs or ATHs.

The poster also said they believe this no longer works, or would be surprised if it did.

I am wondering if part of the motivation for MA's post is that they have actually noticed this "past feature" of the loot system, and now fixed the hole, and POSTED A MESSAGE clearly stating that from now on, if you do stupid stuff like that, they are going to notice and not pay you out for it.

So... perhaps what the correct message to read from their post is; we have adjusted the loot system to include a little extra intelligence now which can detect more abuses than before.

So if they find a lot of mining drops without position changes, bullets fired and gun decay with any monster damage messages, etc... they can opt to take your decay without incrementing your "total decay done" counter (which is maybe the number they are using to help adjust towards your 90% average over time for instance).

Oh, I've rambled sorry. I was trying to say: this method of testing might work, but I am suspecting it's already been patched and so it will fail for you.

You're welcome to test it, but I suspect it will frustrate you.

This is what i believe, i think MA saying "no personal loot pool", means...

1) if true and we understand it as stated (MA are not known to be clear about what they write), then the loot algorithm has changed.

2) It has always been like this and there are some freaky stuff in the past that is now unexplainable

Rgds

Ace
 
To test if a server/personal loot pool exists
Drop 100 bombs in the same spot continuously, then drop 100 bombs normally, record results, rince repeat

If there is a memory, and you get back roughly the same, then to test whether it is server or personal:
Drop 100 bombs in the same spot continuously, move server drop 100 bombs normally, move server, rince repeat

If again you get the same rough tt results, ie between 80%-100% tt returns that would suggest you have a personal loot pool. If on the other hand, first is true but second is not, then you have a server memory. If you have neither then there is no memory.)

Good method to determine if it's server or player determined.

As a note, on Cyrene, when special robot components drop, it happens for a limited time on all Cyrene servers at once. And when it stops, it stops on all servers too. Not only that, but if you don't loot a mob killed during the good loot wave, kill another, loot it and have the good loot in it, then the next one is over the period and does not loot anything good, the mob you killed in the period but looted after the period will not have the good loot either. So this is a clear confirmation that loot is calculated at the moment of looting and based on the current state of the loot tables for that mob.

However, during the launch of treasure hunting on Arkadia, I prospected on 3 different servers. On each server, the more varied stuff started dropping ONLY after a huge amount of crap loot (part of tooth) dropped on each server. As I visited the 3 servers, they eventually all looted the varied stuff but only in the order of activity. This also happened on Cyrene, everyone flushed lyst and oil before the good stuff started to drop. And the find rate was so abysmal at times, that loots were huge when they dropped. Somehow I suspect there was no "success" loot and only "residual" drops that "compensated" or "corrected" the losses of the player.

This is why people consider the existence of personal loot pools.
 
You must be very religious and an adept of homeopathy and every other advertised product or activity in TV and social media. Should you have nothing of value to add to this topic, thank you and good bye.

Oh, yes....I am religious. You have nailed my personality perfectly.

Perhaps you are psychic? That will come in handy in this endeavor.
 
Good method to determine if it's server or player determined.

As a note, on Cyrene, when special robot components drop, it happens for a limited time on all Cyrene servers at once. And when it stops, it stops on all servers too. Not only that, but if you don't loot a mob killed during the good loot wave, kill another, loot it and have the good loot in it, then the next one is over the period and does not loot anything good, the mob you killed in the period but looted after the period will not have the good loot either. So this is a clear confirmation that loot is calculated at the moment of looting and based on the current state of the loot tables for that mob.

However, during the launch of treasure hunting on Arkadia, I prospected on 3 different servers. On each server, the more varied stuff started dropping ONLY after a huge amount of crap loot (part of tooth) dropped on each server. As I visited the 3 servers, they eventually all looted the varied stuff but only in the order of activity. This also happened on Cyrene, everyone flushed lyst and oil before the good stuff started to drop. And the find rate was so abysmal at times, that loots were huge when they dropped. Somehow I suspect there was no "success" loot and only "residual" drops that "compensated" or "corrected" the losses of the player.

This is why people consider the existence of personal loot pools.

I understand that there are many problems with my test, but if you rinse repeat enough the findings should be statistically significant

Rgds

Ace
 
Slasher:

To rephrase what others have already pointed out - if you succeed with your test and are able to show that shooting thousand rounds in the air before hunt produces a global, what have you proved?

If you read the statement: "there is no personal loot pool" as: "we never track anything and we never use any such data when calculating the returns"... then this is already your interpretation.
Then again, why not spend your time proving that your own interpretation was wrong...? :silly2:

I mean, you can't prove if "personal loot pool" IN GENERAL exists or doesn't exist because this statement itself is too vague and could mean anything...


Applause:

Here, however, i can see more specific approach:
How about this:
A - I hunt 50-100 mobs or a specific decay/ammo value and record the results as a sample point.
B - I do the same as A, except before the tests I drop in the same spot some probes worth 10-50% of the hunting costs.
If you try to find out if there is 3 independent loot pools for each profession, then yes, that should be possible to prove. You already proposed a good method, should work.
 
I don't think you can prove a personal loot pool for the same reason no one could prove it before the dev note even popped up---unknown time variable. You can run all these tests, etc, but how do you determine the exact run number that balances out to 90%. And if you did hit it, do you just stop there? Or say this dropping 100 bombs in one spot, then doing a normal run...what if (assuming personal loot pools exist) your "payback" wasn't supposed to come till X runs (unknown time variable) later than that normal run.

Pretty much...to prove this, I think your as screwed as you were trying to prove it before the dev notes came into play. It's just a matter of trust MA or not.
 
I don't think you can prove a personal loot pool for the same reason no one could prove it before the dev note even popped up---unknown time variable. You can run all these tests, etc, but how do you determine the exact run number that balances out to 90%. And if you did hit it, do you just stop there? Or say this dropping 100 bombs in one spot, then doing a normal run...what if (assuming personal loot pools exist) your "payback" wasn't supposed to come till X runs (unknown time variable) later than that normal run.

Pretty much...to prove this, I think your as screwed as you were trying to prove it before the dev notes came into play. It's just a matter of trust MA or not.

Two things can help:
- statistical confidence shows continuous improvement of your model, thus confirming it to a high degree
- you can continue running the experiment until everyone or a high degree of players are convinced and ask you to stop :)

Should the payback be delayed in some cases or not the others, is the duty of the experimenter to filter out noise and improve bias as much as possible so statistically the payback drops more often on the correct side. The number of runs will also help increase the quality of the data, by getting more paybacks.

Also repeating the experiment again from clean starting conditions or giving the instructions to other players to repeat it, will show if it's a good model when they can repeat it. It's not perfect and doesn't have to be. If the results are inconclusive or too random, then we are wrong or the importance of the model is very small, either way, we can focus on something else instead. Maybe the gambling part, sounds fun :)
 
Take a weapon that does at least 21 minimum damage.

A) Kill 1000 snable young males, and record individual loots
B) Kill 1000 snable young males, and record individual loots, but loot only every second kill

your cost per kill is fixed, so you should get no kill-cost related variance, also your effective dps is rather constant. This does not cover all possible causes of variance though.

If there is no memory, then the second loot curve and classes will match the first.
 
matrix_red_blue_pill.jpg


I think we were offered a choice yesterday.
 
You must be very religious and an adept of homeopathy and every other advertised product or activity in TV and social media. Should you have nothing of value to add to this topic, thank you and good bye.
Haha. Being an ass to someone because he pointed out the fact that this thread is useless. Good one.
 
Take a weapon that does at least 21 minimum damage.

A) Kill 1000 snable young males, and record individual loots
B) Kill 1000 snable young males, and record individual loots, but loot only every second kill

your cost per kill is fixed, so you should get no kill-cost related variance, also your effective dps is rather constant. This does not cover all possible causes of variance though.

If there is no memory, then the second loot curve and classes will match the first.

I think someone else did this and showed that not looting the mob is somewhat a stupid thing to do. We can at least try to limit ourselves to looting the mob but killing it in stupid ways.

Also I tried to do this style with punies and MF chips but I'm an obsessive looter :)
 
I just don't know what is so hard about understanding that a fixed rate of return is very easy to program.

Much easier, might I add, than having the loot system constantly "watching you."

Not to mention the fact that MA would gain NOTHING from lying to us in this case.
 
How about this:
A - I hunt 50-100 mobs or a specific decay/ammo value and record the results as a sample point.
B - I do the same as A, except before the tests I drop in the same spot some probes worth 10-50% of the hunting costs.

Should there be a correction of my obviously stupid losses, the B samples should be slightly higher. Getting enough samples that any expected bias (10-50% difference) would be an order of magnitude more visible than natural loot variance (for example 30 reruns of A/B for 10%, 10 reruns of A/B for 50%). If there is no correction for my losses, then there is no personal loot pool or tracker. Should at least 2-3 other players repeat the experiment with the same results, then that should settle it.

Developer note one said that hunting vs crafting vs mining are independent. Or at least that these are independent as far as big loots are concerned. In general, mixing different professions into this test is not a good idea, even without that note. To get a valid test, the test needs to, as far as possible:
  • repeatable
  • control causes of variance as much as possible
  • the result must be unambiguous
 
I just don't know what is so hard about understanding that a fixed rate of return is very easy to program.

Much easier, might I add, than having the loot system constantly "watching you."

Not to mention the fact that MA would gain NOTHING from lying to us in this case.

It is human nature to try to understand how something works, nothing different here, we are just talking about loot pools, that the vast majority of PE players do want to know about.

Fixed rate of return has been shown many times by many players, but where it comes from is a different matter

And beyond that, it is really interesting to know and to do tests

Rgds

Ace
 
Back
Top