Menu
What's New
By:
Filters

Request for a new mining test. Is the game "random" ?

R

R4tt3xx

I want to believe
Joined
Apr 10, 2005
Posts
2,172
Location
South Africa
Society
Freelancer
Avatar Name
Alexis Sky Greenstar
I have a proposal to make to the community, I would like to know once and for all if there is a randomness component to the game.

Here is the test.

All miners participating, perform the same actions with the same equipment at the same time and record the results.

For example, pick the same relative coordinate on a server, drop a probe at the same on each miners relative server, then move to another coordinate on their respective servers and perform the actions again.

I am hoping that the value of these kind of tests can be seen just be attempting them.

Good Idea or no ?

L*R
 
R4tt3xx said:
I have a proposal to make to the community, I would like to know once and for all if there is a randomness component to the game.

Here is the test.

All miners participating, perform the same actions with the same equipment at the same time and record the results.

For example, pick the same relative coordinate on a server, drop a probe at the same on each miners relative server, then move to another coordinate on their respective servers and perform the actions again.

I am hoping that the value of these kind of tests can be seen just be attempting them.

Good Idea or no ?

L*R
Click to expand...

If it were random, most players would win, and MA would lose. It certainly has a random element as well, but there is an algorithm that follows certain rules. Also MA change this always a bit, so we never can see a pattern. There are so many parameters, they can use to generate your payback. One is sure, we never will find out, what is coded behind the wall :)
 
R4tt3xx said:
I have a proposal to make to the community, I would like to know once and for all if there is a randomness component to the game.
Click to expand...


No need to test - the answer is yes, but only to the degree of providing a controlled amount of variance. Virtually all loot systems in virtually every game are "controlled random". There's a variety of common loot algorithms, my bet is MA uses some sort of custom derivative.

A truly random loot system would be a disaster.
 
I'm not sure what's specifically being tested here, so you need to start with a clear hypothesis and how to test it first. This setup so far would just be people doing different individual uncontrolled runs. You can test things like difference in hit rates, etc. but specific treatments are really needed. You could test whether hit rates are different in different servers at the same time maybe, but that's not really testing randomness

It's already been shown that there is random variation around whatever the "true" average hit rate is for a given time and place. I mentioned earlier before about testing whether claims are randomly distributed, clumped, etc. I still have to pull up the exact code in R or other stats programs for doing that, but the data collection could just be one person carpet bombing an area with maybe 60 drops three or four different times. Get the lat/lon data for each claim, which run it was, etc. and that's all that's really needed.
 
kingofaces said:
I'm not sure what's specifically being tested here, so you need to start with a clear hypothesis and how to test it first. This setup so far would just be people doing different individual uncontrolled runs. You can test things like difference in hit rates, etc. but specific treatments are really needed. You could test whether hit rates are different in different servers at the same time maybe, but that's not really testing randomness

It's already been shown that there is random variation around whatever the "true" average hit rate is for a given time and place. I mentioned earlier before about testing whether claims are randomly distributed, clumped, etc. I still have to pull up the exact code in R or other stats programs for doing that, but the data collection could just be one person carpet bombing an area with maybe 60 drops three or four different times. Get the lat/lon data for each claim, which run it was, etc. and that's all that's really needed.[/QUOTE


100% correct (again)
Click to expand...
 
trance said:
If it were random, most players would win, and MA would lose. It certainly has a random element as well, but there is an algorithm that follows certain rules. Also MA change this always a bit, so we never can see a pattern. There are so many parameters, they can use to generate your payback. One is sure, we never will find out, what is coded behind the wall :)
Click to expand...

This is my thoughts excatly. Players over think to much in this game and try to beat the system with their theorys, but when they dont work out, they belive the game is broken. I dont keep spreadsheets or much data, but i play most days and do ok in game.
 
bigdaddy trim said:
This is my thoughts excatly. Players over think to much in this game and try to beat the system with their theorys, but when they dont work out, they belive the game is broken. I dont keep spreadsheets or much data, but i play most days and do ok in game.
Click to expand...

The system is programmed to decay over time, the more you hit, the more variables you have to take into account in order to make your next attempt successful.

It is possible to be 100% accurate at least for a short period of time, until the decay kicks in and cascade failure occurs, the game is not broken, it's programmed not to loose.
 
kingofaces said:
I'm not sure what's specifically being tested here, so you need to start with a clear hypothesis and how to test it first. This setup so far would just be people doing different individual uncontrolled runs. You can test things like difference in hit rates, etc. but specific treatments are really needed. You could test whether hit rates are different in different servers at the same time maybe, but that's not really testing randomness

It's already been shown that there is random variation around whatever the "true" average hit rate is for a given time and place. I mentioned earlier before about testing whether claims are randomly distributed, clumped, etc. I still have to pull up the exact code in R or other stats programs for doing that, but the data collection could just be one person carpet bombing an area with maybe 60 drops three or four different times. Get the lat/lon data for each claim, which run it was, etc. and that's all that's really needed.
Click to expand...

Claims are and are not randomly distributed. It depends on the player that is at the wheel. There also is no true average, again it depends on how the miner moves and what the system does in order to compensate for those actions.

The "AI" of this system is not static, it adjusts to the players actions and returns. The skill system could also be used to supplement the lack of player skill and is something that is not being accounted for and cannot be accounted for without the correct framework.

I am not looking for a way to "break the game", I am just interested in knowing what the mechanics are and how they work.

And for those that are interested I have a framework, I have had the same one since the days of Project Entropia. It's just evolved and gotten more advanced over the years and soon I am currently working on a macro version that works within the mining tool's radius and diameter.

I will also be posting a little math challenge, if you pass it, you have the basic mathematical understanding that I am looking for in order to evaluate and understand what I am doing.

L*R
 
R4tt3xx said:
The system is programmed to decay over time, the more you hit, the more variables you have to take into account in order to make your next attempt successful.

It is possible to be 100% accurate at least for a short period of time, until the decay kicks in and cascade failure occurs, the game is not broken, it's programmed not to loose.
Click to expand...

And what does this say?
It say, that there is no much of pure random. Decay is also a part of the algorythm.
Yes, ofc. it is coded, so that the bank always win :)

One of the random elements, in loot generation in mining, is the size of the claim. We have the basic size claim II - XXX Into each of that is some space left to generate randomly the size.
So a XIII claim will be between 50 ped - (start of size XIV claim)

One more element of random is the location where the claim pop up. This is in random location of the range of your finder, there where you troped the probe. This is a random function and work like :
get random point in navigable radius of location. I use this function often into unreal engine, if i need to get a random location around some thing, and also need to be in navigable place.

thats exactly the point, why i say, you need to calculate your next drop point from last drop point, and not from where the claim poped up.
 
Last edited:
R4tt3xx said:
I will also be posting a little math challenge, if you pass it, you have the basic mathematical understanding that I am looking for in order to evaluate and understand what I am doing.
Click to expand...


I think you'd be better served having a game developer evaluate your framework rather than a mathematician for whom the engine is a black box. Your theories are always fun to read, but I don't think you understand how impractical many of them are from a development standpoint.

This community concocts a lot of wild and wonderful theories, but I think most people would be surprised at just how mundane the loot algorithm actually is.
 
Detritus said:
I think you'd be better served having a game developer evaluate your framework rather than a mathematician for whom the engine is a black box. Your theories are always fun to read, but I don't think you understand how impractical many of them are from a development standpoint.

This community concocts a lot of wild and wonderful theories, but I think most people would be surprised at just how mundane the loot algorithm actually is.
Click to expand...

I performed a few tests last night with something very similar to carpet bombing and the results were interesting. I did score a global which helped my cash flow, but then lost peds after that.

This test is a small scale version of the sequence I used on Project Entropia all those years ago. What I a suspecting is that for each player there is a "virtual loot pool" around that avatar. Depending on when and where the player attempts to draw from the pool, the results change.

An interesting note last night was that my hitrate with the idea's current configuration dropped off considerably when there were more globals, and would make sense as the value of the loot pool would fluctuate much more with more player activity hence the increase in globals.

I am going to change the values of the experiment to their lowest when the globals are occurring and see what happens.
 
You must log in or register to reply here.
Back
Top