FourSpeed

Hehe, I was thinking the same thing...

...and I like PB's too...

Yikes!!! I'm a closet-masochist! I think I better go see the shrink in Rm 604 now...


Cheers,
4


PS> As for bad powersets, hmmm... Nothing leaps to mind for me - the most
hated toon I have is a Rad/Elec Defender who's performance was all over the
map - when he was good, he was godlike, and when he wasn't, he was insta-dead.

The toon drove me nuts, because you never knew which of the above two options
you were gonna get ... every fight... regardless of mob con. By contrast, my PB or
my E3 Blaster is positively smooth sailing compared to that defender.

I'll be curious to hear your opinion in the Cimerora portion of the DA arcs.

As for that particular mission, it gets a LOT easier once you have your iStuff.

Barrier, Aim, BU, Judgement, and your best AoE or nuke, and Bewm! Dey All Fall Down.

Good fun either way though.


Cheers,
4

I just see it as still more evidence of "testing excellence" on the part of our devs.

/sarcasm


Cheers,
4

This. In fact, my L50 PB doesn't have it at all and does fine (he's a ranged
Bi-Form build - Dwarf is *my* lightform)

I haven't looked at your build in Mid's, but at first glance, it struck me as a Human
build far more than a tri-former.

In any case, I wouldn't worry too much about Lightform or not. One might even make
a case for Quantum Flight over LF, *if* all you're really looking for is a quick
"Oh sh*t" button... <shrug>

GL, HF,
4

The part to "get" is that whether or not you buy or slot a single IO, you can
use your access to the WW/BM Market to SELL the stuff you do pick up.

As others have pointed out, making inf in-game is very, very simple. The guide
in my signature, for instance, is a beginner approach that is failproof.

Also, 300M is more than enough to decently equip most toons with IO's. You'll
want to do a little reading about Frankenslotting in the guides area to understand
how to go about it, but it's quite effective.

I do that with most of my toons (in their L20-30's), and they're fine, and they
consistently outperform simple SO builds throughout their entire careers.

You only need massive amounts of inf if you're planning to seriously twink a toon
with purples and pvp sets - something that can be fun, but is never necessary
to play any of the game's normal content.


Regards,
4

Good thing we're not analyzing chat channels...

As for Pareto in this game, 90/10 can still BE following Pareto. It's a different
ratio than typcial, but it's not completely ludicrous.

It could also be that Pareto may not apply - but we have zero way to prove it.

In order to prove it, you have to plot a LOT of sample data and actually show
that you can fit it to a power curve (or not, if Pareto doesn't apply).

Zero chance of us being able to do that.


Regards,
4

While I've little doubt that there have been exploits and folks that abused them,
I would expect two things:

1> If it's truly widespread, I'd expect MOAR moneh in the system - who cheats
to make less influence??? Keep in mind that I was the one person advocating a
high number (400T+ @100K accts) for in-game influence, but we're simply not
seeing it in the dev number or the market bids. Those latter points are the key
drivers of my revised, current 150T/75K/90-10 estimate.

2> If #1 is false, then we're talking about a very small segment of the overall
population. I would not expect, in that case, that they would skew Pareto much.


Regards,
4

Well, it's certainly a possibility. However, I have some doubts in that regard.

Unfortunately, I can't think of any way we could even test that without a bunch of data.

The other side of that question would also be trying to find plausible reasons
to explain why there'd be different segments. In other words, what *are* these nebulous
"parameters" you refer to and how do they cause any differences?

Also, addressing Fulmen's idea - in RL, we have a *lot* of factors affecting $ acquisition;
folks working multiple jobs, part-time workers, education level, type of vocation etc.
Even so, Pareto still applies. Primarily, I think it's because when all is said and done,
those are still "free choices" that people can make about schooling, career choice,
effort/motivation levels etc.

In CoH, sure there isn't a "need" for inf per se, but by the same token, if you do
*any* content at all (kill a mob, click a blinky, grab an exploration badge etc.)
you still get some inf as well whether you need it or not.

In that way, "what you do" is still a free, player choice.

If you just do normal content though, that will obviously limit your inf ceiling
because, comparatively, those rewards are quite small compared to using various
market, merit and other inf-gaining techniques.

In that way, I still can't really find much reason why Pareto *shouldn't* hold in
the CoH economy (pre Freedom) apart from the one caveat made when I conducted
the survey - ie. Forum readers (my respondents) have more access to information,
and are probably more dedicated to playing all facets of the game than the game's
active population as a whole.

If true, that may prove the conventional thinking that forum players are richer
(on average) than the active population base - that was my rationale for shifting
the categories to the right (towards lower overall wealth) when extrapolating
and it is the shift that changes the Pareto results if you believe the inf total
actually is in the 150T range.

That shift, btw, results in a mere 2000-3000 players in the 75K active base with
holdings of 10B or more - something I believe that both you (Fulmens) and Arcanaville
were advocating from the outset.

Anyway, I don't see how we can get much closer or more accurate without a LOT
more data, which I highly doubt would be forthcoming <shrug>.

So, without that, we're really where we were at the outset - pick your favorite
number and call it your guess at total inf in-game.

Based on what I can see and dabble with, the numbers I posted represent my best
guess at what we have. I'm at least comfortable that it makes plausible sense
while incorporating the actual data we do have.

I still can't really see it as low as 100T - to me that would be an absolute floor, with
even the simplest extrapolation, and it doesn't begin to account for any gung-ho
activities like power marketing, extreme farming, etc. which would clearly raise that
floor. 12T among an estimated 5.5M scrappers, is a paltry 2M per on average --
WAY too low -- even in I-1, pre-market days (imho). YMMV.


Regards,
4

Basically, I'm settling my estimates in at about 150 Trillion (+-50T) for an active
population base of ~75,000 (+-25K) active accounts with Pareto at ~90/10
for several reasons outlined below:

1> Published Dev Numbers of 12 Trillion for just scrappers along with popularity
of AT's (from BABs) suggesting about 13% of AT's are scrappers amongst a published
account base of at least 500K accounts / 43M toons.

I have assumed that the scrapper number is *only* counting inf actually on characters
(the easiest way to actually datamine it with a very simple SQL query).

That sets a plausible inf floor of ~90 Trillion across all AT's.

2> Observed number of outstanding Market Bids that could reasonably be used
to store influence. That number is ~30,000 currently, which means, at most,
there could be as much as 60 Trillion stored there (if every one of those bids was
2B - pretty unlikely). It's very probably *much* lower.

I used 30 Trillion (ie. the avg of 0T - 60T)

This adjusts the plausible floor to 90T-150T, but it also limits the ceiling as well
due to simple storage limit issues.

3> Statistical behaviour of free, unconstrained, income earning/distribution systems,
which typically have a log-normal distribution across the population (a bell curve
when plotted on a log scale) and generally follow Pareto (The 80/20 Rule).

4> Results from the actual survey I conducted a few months back. While it involved
a very small number of respondents, it did (imho) give plausible evidence of following
Pareto (88/28), and it definitely plotted out as a log-normal distribution, which adds
support that point #3 may well apply to CoH's income system as well.

5> A LOT of dabbling with spreadsheet(s) I have made while conducting the survey
and analyzing its results. In summary, keeping the patterns from point #3, and the
relative distribution ratios from point #4, I can model in-game inf for a given
population of accounts. The only results that don't end up being bizarre, or contradictory,
occur when population is between 50K-100K, total-inf is between 100T-200T,
and Pareto is ~90/10. At 75K, it comes out to 143T with Pareto at 93/11.

So, is the 90/10 Pareto explainable? Maybe, if you consider the specialized knowledge
a player needs to make Billions (ie. Market info, how to circumvent the 2B limit,
use of various merits to make inf, crafting, memorized recipes, etc.).

All of the survey respondents (and avid forum readers) have easy access to that
information. The game's population at large? Maybe not, and that *may* be enough
to skew the distribution into the 90/10 Pareto range across the whole group.

In the survey, the largest category for respondents was 1B-10B per *account*.

In my 150T/75K/90-10 model, the largest category is 10M-100M per account,
something that was actually achievable even in I-1 pre-market days.

Are they correct? No idea tbh. But they come closest to making sense given
the data we have (imho). YMMV


Regards,
4

Fair enough. It is true that it must resolve to a scale invariant power-law curve
to be truly considered adhering to Pareto.

That said,

A> We DO know (generally speaking) that non-constrained income acquisition systems
DO adhere to Pareto.

B> In the absence of any truly random sample data that we can plot, we can't
say with absolute certainly that the income system in CoH IS definitely Pareto.

However, the data we have seen are certainly suggestive that it may be.

If you're sufficiently bored that you feel an overwhelming need to quibble over
the robustness of it, be my guest.

Certainly analyzing it from that perspective gives pretty compelling reason to think
that the bulk of the accounts in existence are truly inactive.

Given the 12T Scrapper number and the missing Market storage bids, an in-game
inf total aof 150T plus or minus is, at least plausible for 50K-100K *active* accounts
if a 90/10 Pareto number applies (based on my various spreadsheet tinkerings).

It's a distinctly lesser number of active accounts if 80/20 is expected at 150T.

So, the only issue I have now, is if there is a realistic rationale for saying the active
accounts as a whole are distinctly poorer (90/10) than the surveyed forum crowd (88/28)
even after shifting the entire distribution a full two wealth categories.

In actuality, I suspect there may be. Of course showing/proving it is another
thing entirely.

But, for me, 150T +/- for an active population of 50K-100K with a Pareto of ~90/10
for the entire active group, is plausible, and rationally explainable.

150T for 500K with a completely non-standard distribution and an insane 82/0.75 Pareto
is simply ludicrous.

That's what I believe based on what data we have. You can believe anything you like.

At this stage, I don't have anything else to add to the topic.


Regards,
4

In the simplest sense, Pareto *always* applies because the statement:
"X% of a resource is controlled by Y% of the population"
will *always* be true for any arbitrary value X.

More specifically speaking though it's usually expected to be (X%, 100-X%),
and for many natural (or unconstrained) systems it's often 80/20 (see Pareto Principle).

In the survey, I took a lazy approach by adding the top few categories (28% of
respondents) and calculated that they had 88% of the inf reported so that's
what I listed - 88/28. While not rigorously adhering to Pareto, I can eyeball it,
and see that it's not wildly off, or at least, looks largely natural and unconstrained.

In fact, if you arbitrarily set X = 80, then the farther Y diverges from 20, the more
likely you are to find an "unnatural" or constrained situation. Keep in mind though
that Pareto does NOT have to equal exactly 80/20, and in fact, it doesn't even
have to add up to 100. It's just aesthetically nicer when it does

However wild divergences are indictative of some constraint affecting the results.

For example. If Pareto had calculated land ownership vs Italian population, he
wouldn't have found 80/20. Why? Because there's an implied constraint there (not
all of the population owns any land -- the "population" has to be among land owners
who are then, equally influenced by wealth and inclination for how much land they
own).

In the same vein - if we looked at income for the population of the USA, you'll get
a weird Pareto number - why? Again, several segments of the population don't (or
cannot) work (ie. kids, retirees, disabilitied or retired folks, etc.). If you look
at income among "employed population", then you'll see sensible Pareto numbers.

So, in my W.A.G. data above, when I see a Pareto number of 82/0.75 I know there
*must* be a serious constraint having a huge effect on the numbers and, as I said
in that post, that number is completely insane.

In this case, it's like my USA income example above and I believe in the simplest
answer, which is the likelihood that a large % of those 500,000 accounts are probably
not actually participating, and while I'm still not quite happy with the Pareto
number I get even so, it's a LOT better than it was. I'm guessing that we're still
off a bit on the *active* inf earning population.

In the example you quote though, it *shouldn't* affect Pareto at all for the simple
fact that most people don't make money to have money. They make it to equip
their toons. In general, people who like SO's will tend towards lower earnings,
and people who like Twinked builds will tend towards higher earnings. Similarly,
people who like using the market will tend towards higher earnings, and people
who don't will tend towards lower earnings. The key, however, is that there is
NO constraint that forces that (pre-Freedom).

It is a simple, unconstrained player choice that governs it -- ie. natural.

Keep in mind though, the game IS going to diverge from it going forward due to
Freedom. Prior, there were NO constraints on the ability for any person or toon
to make in-game inf.

With Freedom however, there are constraints: Freems and Preems have various
restrictions on market-use, IO use, and even influence caps different than the 2B
VIP cap.

Thus, the rules change for segments of the population going forward, and that, over
time, will have an effect on the Pareto numbers (unless we calculate them for each
particular segment separately).


Regards,
4

Ok, I think you're misunderstanding something here. My 85% inactivity number
is based on the 500K accounts they say have at least 1 L50 on them.

That would bring the number of active accounts into the 75,000 range which,
appears to be right in the range of what you're quoting for paid subscriptions.

When I adjust my spreadsheet for 50K-75K, with shifted categories, *then* I
get inf projections in the 160-220T range which are at least plausible with the
new estimates (~125T or so).


As for VW's behaving as actual RL economies, nobody is saying that things are
exactly comparable.

But, when all is said and done, the game gives equal chance to earn inf and equal
chance to spend inf to everyone equally (prior to Freedom).

There is zero reason to believe that simple distributive statistics don't apply.

There is zero reason to think that given a population of active players, some of
those players will naturally make more (or less) inf than others and there's zero reason
to believe that it would *not* be some sort of bell curve. Indeed the actual survey
data collected *shows* such a curve from the respondents.

There is zero reason to think that X% of players control Y% of total inf is false.

Does X have to be 20 and Y have to be 80 exactly? No - not even in RL cases
where Pareto is clearly established. That said, if X is 0.75% when Y is 80%
there is clearly something odd going on.

The devs have categorically stated there ARE a minimum of 500,000 accounts,
but there is zero reason to believe they are all active - especially in light of your
65K-85K subscription count (link btw?).

In short there is nothing magical in the analysis that defies simple common sense.


There are in fact only two ways for a toon to have zero (or minimal) influence,
given the mechanics of the game.

1> Don't play it.
2> Spend whatever you earn.

Arcanaville's idea of what a build may be worth is interesting, but extremely
difficult to quantify.

We know it takes about 20M for an SO build. If a player is comfy with that level
of performance (or constrained to it for Freems), then once that is achieved,
they're going to be net earners if they continue to play that toon.

For IO builds, costs are all over the map. Otoh, if they're into expensive builds,
then pretty much by default, they'd also have to be into inf accumulation to afford
it - it seems likely that those folks would also be liklier to accrue an excess
of inf, if for no other reason than buying the next shiny on that toon, or another.

I'd certainly expect those guys to be above the 3M toon average implied by
125T/43M.

It seems far more plausible that the paupers are liklier to be the DO/SO using
players than the IO using players generally speaking.

Anyway, for myself, I'm comfortable with the idea that, if the numbers we have
are factually accurate, then some of the considerations I've put forth here reconcile
the oddities given by the flipside premise that 500,000 people with 43 million
toons have only managed to amass 125T over 8 years.

For me, that simply doesn't add up correctly based on things we can actually see
and measure in earning rates, market activity, survey analysis and simple statistical
principles.

You all can feel free to believe whatever works for you.


Cheers,
4

Let me turn that around - can you show me any compelling evidence why it
*doesn't* hold other than your vague suspicions about MMO economies?

In an environment where every individual player has exactly equal chance to make
influence and the only actual barrier is their knowledge and effort levels, can you
give me any compelling reason to believe that income distribution would *not* be
some sort of bell curve?

Have you read Edward Castronova's papers? He specifically stated the Everquest economy
would rank it as the 77th country in the world (at that time).

He may not have categorically stated "Pareto holds", but he certainly strongly implied
that Virtual World economies exhibit many of the same characteristics of real-world
economies and many of the same effects and behaviours apply.

He did specifically state this: "Being an elf doesn't make you turn off the rational
economic calculator part of your brain." clearly implying that players approach
these things similarly in VWs as well as RL.

But even ignoring all of that, we have some actual real data that also strongly
implies that some of these factors may apply.

Nobody has refuted the survey data I collected beyond any caveats I mentioned myself.

Nobody has contested the validity of it, and the only point of issue has been
the total projected from that data.

While it is a very small and non-random sample, it still *clearly* shows a very distinct
log-normal distribution, and it does loosely follow Pareto - in actual data.

Please give me some compelling reason to believe that more respondents would
fundamentally alter those real data patterns already clearly visible.


Next, has everyone here forgotten the simple L10 earning experiments we conducted?

100K by L10 was simple.

1M / hr for a L50 is darn near unavoidable.

Those were *measured* facts.

Characters in I1 made 10's of millions simply playing the game content - no market,
a 9999 trade limit, no fancy-shmancy uber shinies to sell...

If you want to make *Billions*, yes, you probably need the Market. But we're
not talking about billions here. 125T/43M <= 3 Million - to make that much,
all you have to do is stop standing around AP yakking, and kill stuff for a single
Saturday afternoon to beat the average.

If you'll recall my analysis of the survey results, you might recall that the biggest
issue in forecasting total in-game influence was having a real number for population.

I used 100K and came up with somewhere north of 400T.

These new numbers give us some compelling evidence of how many *active*
accounts actually are playing on a regular basis.

The market bids are irrefutable - there simply *cannot* be more than about 60T
stored there, and the odds are very high that it's actually substantially less.

Further, with an average of 86 toons / account (43M/500K), if every toon had a
full 2B on them, 172B would be the maximum (on average) that an account could
hold without market storage. I'm fairly comfortable that most of us don't have
2B on every toon we have, and most of us DO use the market to store a significant
(>10%) of our wealth. That seriously limits the upper end of in-game inf.

Given those limits, I'm inclined to agree that what we actually have in-game is
much closer to 150T than 400T.

So, guess what happens if you invoke Occam's Razor, and say the simplest
answer is probably the right answer?

In that case, simple means that 85% or more of the accounts are NOT played.

Guess where the survey numbers (showing typical financial pattern tendancies)
come out when you base it on a population of 50K - 75K?

Yep, instead of 400T you get 200T or so. When you apply a forum bias to earning
levels due to increased dedication and knowledge, by shifting the categories
to right (as mentioned in my analysis), things start making a LOT more sense,
without needing to throw out established rules of thumb or patterns visible in actual
experimental data.

Given the numbers we have, I'm fine with ~150T in the game - provided that only
about 10% or so of the 500K accounts in existence are actually actively playing
the game regularly.

With that one simple adjustment - a lot of inexplicable issues suddenly resolve
themselves.


Regards,
4