Call for miners: join the mining train hit rate testing

kingofaces

Stalker
Joined
Jun 10, 2013
Posts
1,521
Location
US
Avatar Name
Tony KingofAces Hans
A bunch of us have been discussing whether other people mining in the same area before you affects your hit rate, TT, etc., but there really isn't any good data on this. There are a few ways to test this, but the first test to science the answer out if pretty straightforward if we have a few people willing to show up for testing at the same time.

I'd like to have at least three or four people, but we could do this for up to maybe 10 (more than that will make keeping track of data tough for me and likely unneeded). Basically, the first person will start mining, the second person mines where the first did, same for the third, fourth, etc. in the train. If an area truly gets "depleted" by multiple miners, we'll be able to detect it this way since we'd have the right experimental design to account for randomness and a control group (first person). I'd be willing to volunteer to go last if anyone is worried about losing PED at the end of the train. If we don't see differences, that means it doesn't matter if you mine where someone else did recently.

Basically what is needed is:

  1. Each person will drop unamped at predetermined set of coordinates in order (ore and enmatter). We'll shoot for 50 of those double drops a person if we only have a few, but might be ok with 30 if we get enough people.
  2. Use a Finder F-101 that you can max out for depth and radius. If everyone's finders aren't the same, that can confound results.
  3. Wait until the person ahead of you in the train has dropped and pulled their claims before your drop at the next location (where the person ahead of you was).
  4. Out of the 50 or 30 drops, tell me what your hit rate was. You can use LBML to do this automatically or just keep a tally yourself.
  5. Optional. Record the exact size of each claim (not just categorical size). It's possible claim size might vary in these conditions, so I'd like at least the first and last person to track the size of each ore and enmatter claim separately. It can help to have those in between do it, but isn't absolutely needed for the analysis if I get those two people.

Once I have all the data, I can run statistical analyses like I do for actual scientific research to rule out the effect of randomness (i.e., it's "dynamic") and see if there actually are differences in hit rates for anyone in the train, or if it's all just random noise. I'll post the results here shortly after we're finished.

Tentatively, I'd like to try this on the weekend of Feb 23, maybe around 16:00 UTC on Feb 23 (though that can be tweaked depending on who's available). The only requirements are that you can bring an F-101, have enough stats to max it out, and can drop at least 30 drops of double ore/enmatter. That and being able to follow experimental protocol. Exact time and location are to be determined, but I'll provide a set of coordinates to follow the day of. Let me know if you're interested in joining in or have questions.

It's possible people may lose some PED at the tail end of the train if we do find a difference (see #2 in my linked post for a followup experiment if we do). That would only be a small fee for finding out the answer to the question of multiple miners and hit rates, so do it for science if nothing else. :yay: :wise:

Update: Just for ease of searching if people stumble on this thread, the results are posted here.
 
Last edited:
Count me in and thanks for doing this King :)
 
Count me in but won't be available on weekends before march.
I can join for step 2 if you do the first one next week.
 
That's three people total for this weekend so far including myself (I'll get you in Liu if we do round 2.). One or two more people would be great to have.

I'm currently selecting coordinates that I'll PM to people at the end of the week.
 
tested and done. nodes have about a 20 minute refresh rate... server supply for rare resources is an hourly cycle. idk if any of this has changed. if you mine the same depth as someone else in the same spot a few seconds after them you'll NRF your way to hell.

Unless of course that node refreshes with a resource at that perfect time.


the system has a fail safe as well.. 28% HR is achieved planetside running with auto-tool and a massive overlap. corrective waves to get you to 28% hr always come if you cycle enough ped.

TT returns are fixed in a range 87-95% with a mean of 91-92%

markup is the only way to win.

Edit: i have a LBML screenshot somewhere on my pc of 4000+ drops showing a hit rate just a tick under 28% i believe. let me find it. (all outdoors of course)
 
In the interest of good science I would like to help.

Leroy Casper Hunter
 
tested and done. nodes have about a 20 minute refresh rate... server supply for rare resources is an hourly cycle. idk if any of this has changed. if you mine the same depth as someone else in the same spot a few seconds after them you'll NRF your way to hell.

Unless of course that node refreshes with a resource at that perfect time.


the system has a fail safe as well.. 28% HR is achieved planetside running with auto-tool and a massive overlap. corrective waves to get you to 28% hr always come if you cycle enough ped.

TT returns are fixed in a range 87-95% with a mean of 91-92%

markup is the only way to win.

Edit: i have a LBML screenshot somewhere on my pc of 4000+ drops showing a hit rate just a tick under 28% i believe. let me find it. (all outdoors of course)

Are you serious about that HR and TT return??? That is so low. I stop mining when I'm under 35% HR cause I find it terrible then. Overall TT return should be over 100% also, and not only with me but with lots of miners I know.

And for the timing of refresh rate of claims, you are way off also

Testing this is far more complicated than that.
 
Last edited:
Rocket, as far as I'm aware, no one has actually formally statistically tested any of that yet, especially with a proper experimental design (half the mining "testing" done on here isn't really testing because of that). That's opposed to allegories where I've also heard of people following another miner and not getting constant NRFs too that conflicts with what you just said.

The whole point of this little project is to actually show the data and have formal scientific comparisons as opposed to people waxing and waning about randomness, it's dynamic, etc., especially since there's no question the community has varying ideas on this subject. At the end of this, I'd be able to bring this data to any other scientist, and they'd agree with whatever the findings were. Most of our mining knowledge doens't reach that level of rigor, and this will hopefully give people a crash course into how formal research is actually done.

Either way, the data and analysis will be up here a few hours after the testing whether other miners affect your hit rate or not. Worst case scenario is that hit rate is affected and a few of us are out a couple PED, but the question gets largely settled. Best case is that it's just a normal mining run for all of us and we find out that other miners in the area don't affect you.
 
In the interest of good science I would like to help.

Leroy Casper Hunter

Thanks, and that gets us to four people. We should have enough now, but one or two more can easily fit in to increase the sample size.

Just a note that we won't be immediately following the previous miner. There's usually going to be at least a few minutes before you bump into an area someone just dropped at, so we'll be going in 5 minute intervals. I'll be sending out coordinates later tonight, and we'll figure out order once we get everyone at the meeting spot.
 
Everyone has been sent the coordinates so far on PCF here (let me know if you didn't get them). If anyone else wants to join in, let me know within about 24 hours of this post.
 
Are you serious about that HR and TT return??? That is so low. I stop mining when I'm under 35% HR cause I find it terrible then. Overall TT return should be over 100% also, and not only with me but with lots of miners I know.

And for the timing of refresh rate of claims, you are way off also

Testing this is far more complicated than that.

What is your TT return?

low hitrate is compensated with mini-multipliers.. those 10x-30x (200 ped ore hits w/ lvl 5 planetside)

if your hit rate falls too low your compensated over time. there's no way to beat the TT or this game is going out of business/into big debt again. Strash's thread basically establishes that everything you're trying to prove about hit rate being a key to success is a fallacy, and with monumental sums of data.

a better hit rate is only going to allow you to survive without drawdown on your pedcard.. but the standardized TT return of about 90-91% is going to be hit by everyone. i cycled about 1.9 mil ped for 91.66% TT and i was mostly in taxed areas indoors. Strash is at 91.4% on 2.2 mil cycled.

if Eve were still around his data is surely the same, but his thread is still here somewhere
 
Just whatever(2018-today)
unknown.png


I have fews indoor to
but not much as rocket or eve
 
Last edited:
What is your TT return?

low hitrate is compensated with mini-multipliers.. those 10x-30x (200 ped ore hits w/ lvl 5 planetside)

if your hit rate falls too low your compensated over time. there's no way to beat the TT or this game is going out of business/into big debt again. Strash's thread basically establishes that everything you're trying to prove about hit rate being a key to success is a fallacy, and with monumental sums of data.

a better hit rate is only going to allow you to survive without drawdown on your pedcard.. but the standardized TT return of about 90-91% is going to be hit by everyone. i cycled about 1.9 mil ped for 91.66% TT and i was mostly in taxed areas indoors. Strash is at 91.4% on 2.2 mil cycled.

if Eve were still around his data is surely the same, but his thread is still here somewhere

At the moment for 2019 only
TT return only : 108,94%
TT return -decay : 102,48%
TT return - decay +% : 114,73%

But I'm sure we all mine in another way, mine is just safe mining, keeping costs to a minimum, with HR's around 40% and still getting multipliers on the way on 110-120% resources.
 
Last edited:
Confounders...

...regardless of maxed out or not:

1. what were the skill levels of the miners across all skills, i.e. total skill count and just mining specific
2. its not double blinded so the reporting agency could be biased, not likely but possible
3. what is the current status of the avatar in their loot cycle
4. were any other buffs involved
5. was it ensured when the last person NOT involved with the test mined the area was remote enough to not influence the results and no one else was mining in the area when the test subjects were mining
6. were any of the test subjects considered 'lucky' avatars :D

Good work though this would argue against your loot being avatar based and more environmental but it would be similar to a slot machine not being able to pay out two winners in a row cuz it knows it already did one...its random and probability based. This would frankly make EU more 'skill-based', i.e. knowing that an area was not mined is helpful but man not being able to know someone has just gone through there 5 minutes before you is a big problem. The respawn time would HAVE to be set for very low or this would create terrible loot patterns particularly in areas that are heavily overmined.


Brick
 
At the moment for 2019 only
TT return only : 108,94%
TT return -decay : 102,48%
TT return - decay +% : 114,73%

But I'm sure we all mine in another way, mine is just safe mining, keeping costs to a minimum, with HR's around 40% and still getting multipliers on the way on 110-120% resources.

would be nice to see some absolute numbers. e.g. 108% on 2000 ped cycled is nice but 108% on 200k peds is epic. and in what timeframe that cycle has happened?
 
...regardless of maxed out or not:

1. what were the skill levels of the miners across all skills, i.e. total skill count and just mining specific
2. its not double blinded so the reporting agency could be biased, not likely but possible
3. what is the current status of the avatar in their loot cycle
4. were any other buffs involved
5. was it ensured when the last person NOT involved with the test mined the area was remote enough to not influence the results and no one else was mining in the area when the test subjects were mining
6. were any of the test subjects considered 'lucky' avatars :D

Brick

  1. Everyone was maxed on their F-101, so skill shouldn't affect anything at that point, but I was level 33 in both professions and Leeloo was 44/41 ore/en. I'd have to check with Casper on theirs, but they were maxed too.

  2. Double blinding typically isn't needed for these types of studies. Qualitative studies or those prone to observer error are more susceptible, but the numbers here were going to be the same no matter who was observing us three.

  3. This is where more replications and future studies would account for unmeasured variability. That being said, I haven't seen good evidence to "loot cycles" based on avatars. If you pull random samples from a set average hit rate on MA's end, you are occasionally going to get "streaks" that appear to be a trend, but are really just part of the random sampling. As much as I've seen, that randomness fully explains "loot cycles" within a per avatar basis at least (I can't say that as much about planet-wide loot cycles). If it's just true randomness, the statistical tests I did would already account for this.

  4. None for me at least that are mining related unless you count an agility boost. None of the mining buffs affect hit rate though as far as I've seen.

  5. Leeloo was watching the area for a believe at least a half an hour beforehand. No other miner was in the general vicinity either at that time or during testing. I won't reveal the exact location yet since we're going to do more testing and want to avoid this problem, but it was about as far as you could get from the nearby TPs and not an area commonly mined for anything. If there was some roundabout way that another miner was affecting our hit rates, that would just support the current findings anyways.

  6. I'm not sure if we could even quantify that.
 
Last edited:
nvm, just noticed the thread
 
Just whatever(2018-today)
unknown.png


I have fews indoor to
but not much as rocket or eve

basically this... it always evens out.

100%+ on less than 1mil cycled (3-5 ped/drop) is just not going to give you the mean average
 
basically this... it always evens out.

100%+ on less than 1mil cycled (3-5 ped/drop) is just not going to give you the mean average

With 1 mil and even 5 ped/drop, that would already be 200,000 drops. That's way more than enough to get you close to the true mean to the point it's overkill. The confidence intervals on a sample size like that would already be tiny.
 
Back
Top