How Many Mobs Is Fair To Expect A Stable Return?

How many mobs does it take to see 90%+ returns assuming you've got fair gear (65% eff) and looter sk

  • 100

    Votes: 5 6.8%
  • 250

    Votes: 2 2.7%
  • 500

    Votes: 6 8.1%
  • 1000

    Votes: 7 9.5%
  • 2500

    Votes: 6 8.1%
  • 5000

    Votes: 9 12.2%
  • 10,000

    Votes: 15 20.3%
  • 25,000

    Votes: 3 4.1%
  • 50,000+

    Votes: 21 28.4%

  • Total voters
    74

jambon

Stalker
Joined
Dec 14, 2006
Posts
1,806
Location
Ottawa
Society
Shaolin
Avatar Name
James Jambon Zidane
(Let's assume you've got a gun with 65% efficiency and a looter skill of about 40-45 as a basis for the poll question.)

There's lots of talk lately about return %'s with 91% - 98% average seems to be the going target depending on efficiency and looter level. Even MA has chimed in a while ago about average returns:

Net return for all hunting activity since VU 15.15: 94.71%
Net return for hunters cycling at least 100,000 PED since VU 15.15: 97.29%
Net return for hunters cycling at least 10,000 PED since VU 15.15: 94.79%
Net return for accounts created in 2017, cycling at least 100 PED since VU 15.15: 94.87%
Net return for accounts created in 2017: 93.47%

But what is the general consensus of how many mobs does it takes to start seeing close to that return? How do those cycle numbers relate to "loot events" or kills. And does sticking with the same mob make a difference?

I mean going out with 100 peds and killing 3 mulmun looter elite is unlikely to get you in the 90%+ return level (unless you get lucky). The same could be said with going out with 1 ped and killing 3 punies. I think it's fair to assume that MA's specs regarding the return on 100 peds is based on an avatar rolling 100 peds into killing like 500 punies.

So what is a fair sample size? 100? 500? 1000+?

Does changing mobs part way through skew your results? I think it would if the mobs have different kill costs...

I'd love to get your feedback on this.
 
Last edited:
I'll take a shot at this.

1) How many mobs until you see a stable return?

This is a complicated question, and has to do with the number of samples you need to average before you can get close to the expected value of your distribution. Inherent in this is that you have to have some knowledge of the distribution itself.
The larger the variance, the larger the samples required. (I don't have a number for you, but it does exist mathematically).

What I can tell you is that the sample size ideally include several events of the largest multiplier to be realistic. This would be your 3000x hofs, and you should have at least 2 if not more of such events.

The other assumption is MA doesn't want to go bankrupt, so that 3000x hof must come from player losses. You might imagine about maybe 3% to 4% of your peds cycled in goes to building this hof. Therefore, you might expect one 3000x hof per 75k to 100k mobs.

So my guess is 100k + mobs to achieve the expected value.

The above involves a lot of conjecture so take it with a grain of salt.

2) In terms of changing mobs, it shouldn't matter. Only number of sampled events matter. You should not change gear though as that has a role in changing your returns. You also have to assume that MA does not change tt returns somehow during the time that you are performing the study.

Just a few ideas.
 
Damn I answered the wrong question though :D.

for 90%+ return, should see within a few thousand mobs (edited).

I changed it because I think I'm too optimistic...
 
I think MA did a disservice by quoting cycle amount as a metric, because it's not really the important factor for all the reasons already pointed out.

I also don't think there's a right answer because it's not purely about number of loot events total, there's also a time component.

Killing 500 mobs over 60mins will yield a different statistical result than killing 500 of the same mob over 12 hours.
 
I think MA did a disservice by quoting cycle amount as a metric, because it's not really the important factor for all the reasons already pointed out.

I also don't think there's a right answer because it's not purely about number of loot events total, there's also a time component.

Killing 500 mobs over 60mins will yield a different statistical result than killing 500 of the same mob over 12 hours.

Yup. I find this to be very important but failed to get recognized. If you spread out your kills over a period of time, a couple of days or a week for instance, for the same number of kills (at least 1k or more), the returns you get are surprisingly stable I believe. On the same mob & maturity of coz. And with the same setup.
 
I think MA did a disservice by quoting cycle amount as a metric, because it's not really the important factor for all the reasons already pointed out.

Most people also don't have a great grasp on statistics or how averages work. If they say the average net TT for a given group was say 95%, there's going to be a ± on that average. That means that while the grand average is 95%, each individual players average is going to vary around that mean. What that variation is isn't known, but if it were say 5%, then the getting 90% TT in that year would just be normal variation for an "in-system" set average of 95%. No one could complain that something was wrong in that case, as that's just how random number generation works.
 
I can't tell you how many loot events it takes, MA keeps all the parameters guarded and it appears that they are constantly 'tweaking' the system anyway, so who knows...

But I can tell you how many I think it should take: somewhere between 5,000 and 10,000 loot events.

My reasoning is, I just spent like 4-5 hours killing 500 robots for a mission, that's about as much as I'm willing to do in a day. Somewhere between 10 and 20 days, my returns should level out. It's not realistic to me that I'd have to persist for 2 months or more before my returns begin to stabilize, that's asking too much commitment from me.

Think of newer players, are they going to persist and keep going despite terrible returns after 3 weeks? I really don't think so...
 
Last edited:
I voted for 2500, because that's where sub 90% is happening WAY less. It's still happening, but a lot less. 2500 is a decent amount too.

Small sample below, % return per day, with the following specs:

- my December month in 2018 with 600k turnover;
- 1st 12 days were on mostly 10 ped mobs, avg of 15k ped/day turnover, around 1500 mobs a day
- rest of Dec was mostly on 35 ped mobs, avg of 22k ped/day turnover, around 700 mobs a day

104.45%
92.60%
91.94%
100.36%
100.06%
92.36%
103.55%
98.54%
85.11%
105.02%
88.29%
92.64%
____________
90.94%
96.44%
99.69%
88.95%
87.56%
97.44%
93.59%
94.01%
95.71%
106.85%
89.64%
156.57%
88.50%
90.01%
108.43%
93.37%
91.28%
116.45%
94.34%

I hope it helps you make an idea.
 
i remember doing the rest of a mission on some 2ped mob which was around 4k kills ending with close to zero multipliers and no multi above 5x with about 55% tt return. worst hunting trip i ever had. all was done in 2 days
 
First of all, the two questions you ask are different in subtle ways. For results to be 'stable' means having no noticeable fluctuations. In our expected dataset, the first occurrence of any good multiplier (even if absolutely 'on time') should provide a good upward kick, the second one less so and so on, because the first one and later ones raised the running average already. Basically, the answer to 'stable' is when one or a few great events makes little change to the outcome. This requires a much longer dataset than what I think you are actually asking...

So, when should you arrive at 90+% returns? Note, this is not the average expectation point (which may be 95%), but a point where finally getting a particularly high multi will shift you above the line for most of the next x mobs until the next nice looter. Thus ask a related question: What effect does a 100x multi have on a run of 1000 mobs?
Answer: it is the equivalant of a 10% rise in output for every single mob in the dataset - the one event could be split across all the mobs and give you the same running total. As we can see, a hypothetical position of 899 ped return for cost 999 peds (1 ped mob) would suddenly jump to 999 return on 1000 mobs - it makes a difference of a whopping 10%.

If you want to be at least at 90% from an expected long-term 95% or so, then the answer we are looking for is: how long until we get a multi that jumps us over 90%, then slow slide in return %, then multi jump again?
We know that 100x multis and higher occur from experience and tracker, so they are all 'in the equation'. We have seen that a 100x multi has a powerful effect on the 1000-mob return figures - it would only have a 1% effect over 10k kills. Even higher multis of 200x, maybe 400x being 'stored up' would have a 2% effect and a 4% effect on 10k kills.

The overall answer is thus: how long until I can be seeing these events that keep me hovering around in the 90%s?
The problem: different mobs seem to have different distributions (or 'personally monitored' loot dynamics...), thus the chunk of returns being withheld from me for a while appears to vary if farming different mobs.

Maybe conversely to people's initial logic, it will be easier to get higher percentage figures from mobs that do not withhold peds for massive multis. Do not confuse turnover with mob count here - I mean, say, 10k punies and 10k argo, 10k ambu etc.

In short (not short enough lol), the final piece of the puzzle here makes the crucial difference and provides various true answers. I know you need a way lower bank roll to hunt 10k punies than 10k ambu, but it should actually be harder in mob count terms to get the 90%+ returns off punies, to be honest, where I have had mobs that on paper are 500x multis or even 1k multis (1k multi = 30 peds payout per 3 pec cost - mobs of 10hp, 20 hp, 30 etc). I fully accept they could be a 'return' from some hunt somewhere else, however.

My really short answer is that I'm looking at about 10k mobs, on mobs that cost 1 ped, or so, for example!

This means I don't really expect a 400-pedder that will boost me 4% on those 10k mobs, but still expect to be above 90% on returns from them! If and when biggies come, I may end up at the elusive 95% or above for a while, that MA says is what they expect from watching millions of events.
 
Last edited:
Damn I answered the wrong question though :D.

for 90%+ return, should see within a few thousand mobs (edited).

I changed it because I think I'm too optimistic...

Pretty much this. Close to average should happen within a few thousand loot events. Absolute average you need much larger sample so as to include exceptional loot events.

However, in my experience exceptional loot events are always "coincidentally" immediately followed by awful loot to take it back, so...
 
I voted 100 just because I wish it was that way, I hate when people say "too small a sample size..."
But just knowing basics about statistics, I know the bigger sample size the better.

I've heard something that comes to mind about slot machines. Something like knowing the overall percent of successes, then actually the more you pull the worse your odds are. This means that really your first pull is the best odds you have.

I actually employ this idea in EU sometimes. On higher "risk" mobs like Levi, Proteron, etc I rather kill 10, 20, maybe 50 at most. I figure if I don't get something good in that chunk, then I'm better off stopping then doing another chunk and losing the same amount. So if I get lucky in 20 = great... if I don't then well loss isn't THAT big.
Others say you got to keep going until you get that multiplier.

GL on your endevors
 
Nothing less than 2500 and over 17000 kills are required.

I voted 25k but the more the better. For me kills and turnover are critical factors.

Followed by bankroll management.
 
The question is... what if u kill 25k mobs per hour and the loot is just not there? What if this whole statement of killing 25k+++ mobs is just simply based on hunting in different times? Than it all does not make any sense.

Perhaps, whole cycling thing suggestion to get decent return is based on... u play some days when its bad and u play some days when its good. Kills doesnt matter? :scratch2: When loot isnt thhere, it isnt there? :wise:

Anyone willing to kill 500 mobs 10 times per one hour? :) or 1000? :) and share the results? Discarding top XX and bot XX loots? :)
 
Last edited:
There is a world of difference between 98%-95%-90%.

95% to 90% means accelerated bankruptcy even on low cost mobs, 0.8-2 ped/kill. 95% is barely surviving. 98% you have a certain chance to survive to next deposit (and here is where MU matters).

For me, 17aug-14oct 2019 meant tt in (all, including def) 98.292 ped, tt out 94.944. 96,5% Max looter lvl32, average eff (65ish) with a guilty trip here and there on lr40fen (L) which I absolutely love.

I would say an average cost to kill towards 2,5 ped. Maybe 70k events (a disproportionate amount 1 ped or lower).

On this interval, I had 3 times some 300x and ONE time a 2kx+. Which 2kx accounted for more than 2% of the overall turnover.

You have a high chance of seeing 90%+ even on few hundreds mobs. You have an extremely high chance of seeing 90%+ on couple of thousands.

But how will that actually feel in on your pedcard and in your soul can be based on trading decisions with an impact anywhere from +5 to -5%, depending on how disciplined, patient and informed you are.

I mean to say that, ever since 2.0 came and quite a few years before, as long as you have some decent k of ped on your pedcard, your habits matter more than tt variance and can even matter more than MU itself (although some habits will be related exactly to MU). I would propose a trilateral 3k bankroll-40dps-1,5ped max per kill and multiply from there. Of which you can expect to see one time falling on your face, but multiple times a deposit lasting months.
 
Back
Top