Please note how it’s positive in several months, right next to months with extreme negative values. Can it be players activating/deactivating their accounts? If so, it should correlate with faucets. I mean if a rich player quits, he will no longer generate ISK (which made him rich at the first place). But no, there is no correlation:
You can imagine that farmer and active accounts are different, that PvP accounts are activated/deactivated, but it has two problems. At first it makes zero sense to keep lots of liquid ISK on a PvP account. Logging out titan accounts does not show up on the ISK graph, just logging out accounts with large wallets. Also, if it’s players logging out, than it must have some seasonality, ergo, the results should correlate with the previous year, and they don’t:
Finally, if it’s GM action that catches large amount of botters in some months and not in others, it must show up on the daily graph. There is a downloadable file in the monthly report and in that a “moneysupply” file, that provides the total money in circulation every day (while the graph is monthly). We should see huge drops when GMs strike:
We can see only one big drop: in October, which was a positive ISK delta month. So no, the monthly oscillations can’t be explained with ban waves. Also, please notice how smooth that diagram is, compared to the monthly ISK delta. If you have any knowledge of statistics, you know that long time averages should be smoother. It’s practically impossible that a smooth daily graph creates a jumpy monthly graph.
What do we know then: that nothing adds up. There are two explanations to this, and neither of it is good. Namely:
- The economic reports are fabricated according to some agenda instead of reflecting the truth. This case we can know nothing beside the fact that we are being fooled.
- GMs know exactly who are botting and ban a few every day in the months when they want to ban them and unban them when they don’t, according to some agenda
Anyway, these totally irrational graphs prove that GM actions in EVE Online are arbitrary and follow some agenda.
I don’t understand – after all of these years, are there still difficulties detecting Bot usage in EVE?
Also, if I understand EVE correctly, why does Bot usage banning not have more of an effect on Corps, moreso than individual players?
LikeLike
By looking at said data I think they are “regulating” the market by bannings bots with enough isk on their wallets… that could explain those deltas. It’s speculation, but would at least make some sense.
LikeLike
If they banned all the mining bots the game would instantly lose all players.
They keep mineral prices artificially low.
For every ship produced somebody has to spend an hour or a few mining. Mining isn’t fun, isn’t very profitable (without bots) and is a typical newbie activity – in a game with no new players. There is nobody in the game to mine all those battleships and supers other than bots. Without them people would be forced to mine themselves, which the PvPers won’t do.
LikeLike
@stawek: or, they could just increase mining yield at the same time.
Alternatively they could consider the possibility that the game has no new players exactly because all new player activities are dominated by bots and new players can’t make ISK. Similarly the game has no industrial players because they were pushed out by bots.
LikeLiked by 1 person
I see a summing of random variables without averaging, so variances increase.
Did you try to check what events in EVE happened during the positive and small negative delta ISK?
What happens if you assume that the active delta ISK have about -25 bias?
Also, what’s the difference between “delta last year” and “delta 1 year before”? Naively, I would assume that they refer to the same time period.
LikeLike
@retsep: “summing without averaging” makes no sense, since average is the sum divided by the number of elements.
the delta-delta graph points have 12 months between them, so the 2017 feb vs 2016 feb to show seasonality.
LikeLike
I believe that the ISK delta is now calculated when last the player logged in, with the player moving from ‘active’ to ‘inactive’ after a 30 (or maybe more?) days. Therefore, if a rich, high-earning player is banned, the effect on the monetary supply will be immediate (as they can’t login) but their ISK won’t be declared ‘inactive’ and show up in the ISK delta until a month or more after.
Perhaps introducing an off shift will produce a correlation but I still doubt it. I believe the ISK delta is dominated by players joining and leaving the game voluntarily and forced bannings are only a minor component.
LikeLike
“the delta-delta graph points have 12 months between them, so the 2017 feb vs 2016 feb to show seasonality.”
This might be problematic — CCP changed what they call “active players” in Nov2016, so you’re not really comparing apples here.
LikeLike
Will “without ” replacement into “instead of” – “I see a summing of random variables instead of averaging, so variances increase.” make more sense?
In my understanding, ISK delta peris calculated by detracting sum of sinks within the month from the sum of all faucets within a month
LikeLike
@Retsep: no, if that was the case, the total money in circulation was fixed. It’s not. Active ISK delta is the money disappearing/appearing by accounts activating/deactivating/banned.
LikeLike
First about me (sorry for my english it is not my first language): i’m a software engineering and i build bots also for eve.
I don’t sell/share them it just for fun / private usage / i don’t sell coins. My IA do something different like logistics (yes fleet with remote rep/energy) / alpha fleet like people do CTA (it is not
different) / scan and report isk value of freighters for suicide gang and something like this (i wanna build something also with neural network just for fun, i’m pretty sure a good IA can be good for pvp than me).
About problem of bots on eve there are many solutions (ccp if wanna can reduce 90% bot usage):
– first define a target (i think on top there is rtm cause it drop ccp incoming and destroy eve economy).
If our target is rtm, how people make coins? 90% ratting on 0.0.
Then we focus on this situation (ccp don’t need to waste time try to catch
people that do other things cause it is a game cat/mouse (require a lot of time/money to invest), it is also more easy find a solution in this case than others).
Now we have a target and some information to look into:
– most of peoples use commercial bot (few people can build a bot, cause
require skills of reverse engineering), the most used use image recognition.
Why they use this technique? (and we focus on them)
Pros:
– easy to build (not require reverse skills few people can do this thing. It is something like difference people know / dont’ know rocket science)
– less time to build system (we can say 80% more faster than build system with injection/memory reading, It is easy to patch if something change also)
Cons:
– less system efficiency (example with eve i can run a bot with memory/injection that use +3/5% cpu usage over eve client, same thing for ram usage > than i can run also 15/20 client with just a normal pc without issue lag)
– input / output lags (performance / they are higest responce time (is bad example for logistic))
– input issue (think about local chat on 0.0 if you can’t see people on
list because you need to slide it to find a red / ineff about fighters/drones control. Just example
memory/injection can calculate when shoot to frigate (also find best target when there are many available) without miss it in few millisec or read data without see it on your screen)
– require image/overview setting adjust (cause it need to see it on screen to read it)
As you know when ccp change image / colors this type of bots have problems (we
found weakness). Example i had many problems in some situation with
large object colliders like belts / pos structure to avoid it when they
are on my way (i see many bot stuck on it or just try to orbit if they are lucky) it is hard also for people some times to get out of stuck.
A memory/injection can create a 3d map then steering / avoid before go on stuck (if you are smart to create algorithm), image rec can’t do it (they cant understand 3D space, you can say: there are
cars that drive by it self, i can say yes it is true but they move in “2D” space, mulitple cameras with also laser/radar to understand size of objects require also skilled eng to program software and it is not this case)(we found other weakness)
We just found 2 weakness that image recognition can’t solve them. This is why we need to focus on main problem / main group. Catch injection/memory read it is very hard just look other game that do it, there are always bots, it is not possible to avoid them for many reason (think about game protection always breaked, cause the main problem it is pc design this is why exist console to avoid this problem) but we can kill them with logic.
The main problem that have all bots that ratting in 0.0 is: other human.
Yes cause nobody can rat in system if can get gank. This is why botters rents system like pipe / so far to low/hight/not conquerable space.
Cause if not red/neutral enter in local they can continue with business and make coins (consider also drone region are perfect for this business cause not require salvage then they can just run with mothership without waste time to salvage), why there aren’t bot rentters on
wormhole (need to salvage / no local).
Solutions ?
– instead a npc bounty, drop a item when you killed all npc (included frigates) that:
1. you can’t tractor it.
2. will be drop like in a belts in same room / surrounded large collidable
object like a labyrinth also require hacking skills (just to make game more harder) then will
impossible to complete for image rec (very hard also for injection/mem reader)
3. if space of labyrinth to move require battleship size (also for kill frigates that you can’t kill with mothership) will be also better cause require reship more logic)
– add more connections to hi/low sec for so far region (maybe 0.0 are more fun for people cause more easy to roaming also for small ganks, Is not eve a sandbox???)
If you drop main/easy incomming you will kill also rmt market at same time (the price of isk are soo lowest than few years ago only massive incomming of ratting in 0.0 can be good for this business other type of bot are not for rtm just for itself).
This is easy/smart approach, there are also more hard solutions but why waste time if you can kill for free 90% of problems.
Report people it is not the best approach cause it is userless, only script kiddies can be catch if somebody have time to see them (people that farming in 0.0 not have this issue).
Just example i reported many times my stupid miners (to check how much ccp is smart) and never get any ban (i used 3 level of IA to simulate human, i thought basic/easy/so stupid/first level will be ban, never happen (i tested for many months), still here in space and say happy miner!). And miners are more easy to catch instead other types.
If you never build a bot you dont’ know how to find them and what is better compromise to catch them. I don’t understand why CCP love lost money every month with people do RTM, maybe they are so rich and dont’ care about them.
LikeLike
Only for report, Falcon merge the thread about this topic in the forums, with a one more generic about bots. Was interesting for me the attacks from Enorn in reddit, and than DMC (ppl of the plan) like my comment reasoned about this topic.
LikeLike