The Rise of the ISK Printing Machines: A Case Study on Bots.

Many of you are familiar with the controversy surrounding botting in EVE Online. Some of you
have read this website’s previous coverage of the issue. I’ve lived in eastern nullsec for about a year, first in Insmother and now in Vale, and in that time I’d gone on many a roam through the drone regions. People had mentioned that it was generally a waste of time to do so, since the local reds were macros and would safe up as soon as you entered local. After going on a few such roams, some lasting hours, I got the message: these roams are not fun, you’ll never get a fight, don’t bother. I forgot about it and busied myself with other things, in-game and out.

I Notice Some Bots
The recent botting controversy piqued my curiosity. I decided I would train for a nullified,
cloaky Loki with a probe launcher to go hunt some bots. When I got one, I looked at the starmap, colored the stars by population, and noticed one nearby constellation where there were 2-4 people in every system. Then, I travelled there, spamming directional scan. In each system, there were ravens named 11 and exequrors named 1111 or 1113, or sometimes 2121 or 1616. Some systems had two ravens and two exequrors, and some only had one of each. Strange, I thought. Worse, the ravens disappeared from d-scan within seconds of my arriving in the system.

I warped to random belts to find them perfectly chained. A few of the belts had partially looted and salvaged wrecks. I attempted to probe down one of the exequrors to find it safed up at a POS. These, I thought, were surely bots, and they were impossible to catch. Just for grins, I cloaked up and just hung out 70km off of a belt. I may not be able to kill them, but at least I can deny them revenue! Then I went to the gym, had some lunch, and did a bit of shopping. When I got back several hours later, they were still safed up, and hadn’t left local.

Bored, I wandered around a bit more – there’s a lot of space out there! Most of the systems I
encountered were totally empty. Not only that, most of the systems in the Drone Regions were
totally empty – it’s a tough place to live! So how could I go about efficiently finding some bots, and how could I prove them when found?

Excel: The Bot to Find All Bots
Being a spreadsheet-minded fellow, I copy-pasted and massaged the data from Dotlan Eve Maps for the drone regions into Excel. Then, I went constellation by constellation on the Eve star-map, coloring stars by system, and took a snapshot census at 08:00. This took a couple hours, but I found it more interesting and challenging than shooting at an IHUB. Then, I went to bed. The
next day, I had a great breakfast, hopped back on at 21:00 and took another snapshot census. I ignored constellations that were empty on the first go round, so it only took about a half an hour.

What I found was that most of the systems were empty. Another good chunk of systems
displayed the kind of variation that suggests normal player activity; they had 3 players at 08:00, and 7 players at 21:00 or first 3 players and then none. However, several constellations had systems with the same number of players as they had 13 hours earlier.

Obviously, the lack of variation could also be random chance – maybe it was two different
players, or maybe two players ratted a bit, then left, and two more players took their place. But I noticed a pattern in the systems which had the same # of players 13 hours later. Many of these systems had truesec below -.60. I’m not sure why that is – my guess is that battleship rats don’t spawn in sufficient quantity in higher truesec systems. Further, many of these systems had a lot of belts. In fact, the number of players in the system was directly proportional to the number of belts. A system with 8-15 or so belts would have two players. A system with 20 or so belts would have four, and a couple systems with 25-30+ belts would have six.

Next, I consulted dotlan for the suspicious systems. A system with bots would display a consistent NPC kill count. Normally, players come into a system, rat for a while, and then
leave. Or maybe they’ll run a plex or a CA, kill a bunch of stuff in a short time, and then go do something else. It is relatively unlikely that a human would have the patience to chain belts for 13 consecutive hours and produce a smooth, even NPC kill count with low volatility.

Part Three: Destination: Botville
After that, I had my list of a handful of suspicious constellations to go and visit. I hopped in my nullified Loki and took a trip.

My experience visiting the suspicious constellations was very similar to that of visiting the first
constellation where I discovered the raven/exequror bots. In fact, that was the dominant bot type I encountered. The ravens were usually named “-=-” and the exequrors were commonly named “=”, though sometimes they had the default names (Bob’s Raven) or just Raven. Sometimes the ravens POSed up instead of cloaking. I encountered a handful of systems where tengus were used instead of ravens, and one or two with badgers instead of exequrors. I also noticed exactly two systems where zealots were used in combination with exequrors. Some systems had a weird mix of ships – an abbadon and a dominix ratting together, or a vargur, golem, and vagabond seemingly just hanging out. I assumed those weren’t bots and were perhaps plexing, POS’d up, or otherwise just chilling out. The only bots I confirmed were ravens (mostly), tengus (a few), and zealots (two) with exequrors (mostly) or badger mk2s (a few).

I took down the name of every bot character that I noticed. Clicking on their info, I noticed that bots in the same system would almost always belong to the same corp. Sometimes, the same corp would own all the bots in the constellation. Further, looking at the characters, almost all of them had little-to-no corp history: they joined their current corp directly from an NPC corp. Lastly, they very often all had the same character born-on date, or very close ones. For example, there would be two raven pilots created on December 20th and two exequror pilots created on January 26th.

It took me two trips to visit all of the botting hotspots, travelling in long loops visiting
constellations and systems in all 8 regions.

Adding It Up
All told, I discovered 82 toons that I was pretty certain were bots. For reference, there were
447 players online in the 8 Drone Regions when I took the first snapshot, which means that
approximately 18% of the players online in the DRs were bots. They were largely concentrated
in six constellations in 4 different regions, although some were in isolated systems that were
particularly juicy from a #belts/truesec standpoint.

I added all 82 bots to my contacts list, and watched, and waited. They’re all online, all the time. They’re online the next day, and the next day. Judging from their kill counts on Dotlan, a pair of bots (raven+exequror for loot/salvage) can kill between 1200 and 1500 ships in 24 hours. They appear to be chaining battleships, which I’m given to understand have a loot value of around 700k a piece. That adds up to between 840 and 1,050 million isk per day, not including salvage; the tengus are probably even more efficient.

One particularly industrious system (ETO-OT in The Spire) is home to four ravens and two
badgers that produce 4200 kills in a 24 hour period, which is about 3 billion isk per day.

Ed. Note: Is there some revenue source I’m missing? I don’t rat; never lived in the drone
regions. Are there officer drops? Escalations? Anything else besides alloys and salvage to take
into account?

As an aside, I want to point out that, as far as I can tell, none of the botting appears to occur in areas of the DRs that are rented out to alliances. IRC, B, FLAME, BREW, C0RE, SPIRE, and
VIP do not, as far as I can tell using the above methodology, appear to have any botting going
on in their space. In fact, no botting appears to occur in constellations where the renters have sov in a single system, even if the other systems in that constellation are unrented and are otherwise suitable for bots (truesec <-.6, belts >10). None of the systems where I’ve found bots are station systems, which makes sense because you’d have to be pretty stupid to run a bot in plain sight. In fact, a significant fraction of the systems with botting are unclaimed. I also found very little evidence of any other sort of botting requiring multiple characters in the DRs, such as macro ore or ice mining or macro-ratting in CAs.

CCP’s Response
Subsequently, I submitted a petition under the Exploits section. It was somewhat lengthy, as you might now believe, and I described my methodology and listed all of the reasons I had to believe that the characters in questions were bots, and I listed the character names of all the bots. I told them that I disbelieved that CCP took botting nearly as seriously as they claimed to, since it would be trivially easy for them to realize that no human could rat for 23 hours straight, but that I was giving CCP the benefit of the doubt and that I was trying to do the right thing.

The next day I logged in to find that all of the bots in Kalevala — 20 of them, were offline. The constellation they were farming was now empty of players. I went to the constellation to see the POSes that they used to store their loot and to safe up were still up, online and running, so I gather that they weren’t deleted/forfeited.

The following day I had a nasty bout of insomnia, probably related to this awful cold with
which I was afflicted. Point being, I stayed up until downtime, when I realized I would have the
opportunity to witness the bot’s morning login behavior. My contacts list started to light up at a seizure-inducing pace. Within three minutes of the servers coming back up, all 12 of the bots in the DA0V constellation in Cobalt edge were logged in and ready to go. All 12 bots in 3TS-12, Malpais were online within twelve minutes of the servers coming up. Within 25 minutes of the server being up, 30 of the 31 bots in The Spire were online.

Worst of all, 18 of the 20 bots from Kalevala were back online. From the looks of it, they
received one day bans. They are now back to producing over 12 billion isk per day.

Soapbox
First of all, I’m not trying to pick on people who live in the Drone Regions and play honestly,
which apparently includes all of the smaller alliances that live out there – it’s a really crappy
and difficult place to live, and it’s one of many parts of this game that has been neglected for
years and is un-fun. I also don’t claim that the NC or any other part of space is free of bots; I just haven’t noticed them. The drone region bots are the ones that I noticed, I figured out the pattern, and I thought people might find it interesting to read about.

While I found this automated behavior interesting to study (especially compared to, say, running the Damsel in Distress for the 85th time), I am displeased that it is necessary.

Apparently, accounts can engage in a behavior pattern displaying *all* of the following
characteristics while setting off precisely zero flags at CCP:

- Log in within minutes of the servers going online and remain online for 23 straight hours
- Remain in the same system, in the same ship, doing the same thing the entire time
- Log back in the next day after downtime and go back to doing the exact same thing in the exact same place
- Always cloak or POS up within seconds of a non-blue entering local
- Have ships, fittings, ship names, skills, and character born-on dates that are identical to other players in the same system and adjoining systems
- Have zero corp history aside from your current corp

CCP also has a good deal of data that I do not, such as IP addresses, account records including referral data that likely connects many of these accounts, detailed real-time client data, the ability to modify the client if necessary, the ability to visit suspicious systems without spending 2 hours to travel 80 jumps, and superior monitoring capabilities should they decide to sit and watch for a jump freighter to come pick up 250km3 of drone alloys.

I cannot fathom what makes CCP believe a one-day temporary ban of disposable macro raven
toons is an effective deterrent. Do they imagine that this stern warning will cause the owner of 20 macro ravens to straighten up, fly right, grow 19 additional heads, eyes, and pairs of arms, and start playing honestly with 20 two-month-old raven accounts? The mind boggles. A bot army of this size will require substantial logistics: POSes in every botted system, frequent runs by freighters or jump freighters, the tacit consent or willful blindness of at least a few senior people in major alliances, and perhaps a corresponding army of middlemen and builders. I have not observed any consequences for any of these parties.

Now, these are only 80 bots on an average server population of 40,000. However, these 80 bots are enough to produce minerals and isk for 2-3 supercarriers per day. A relatively devoted honest player might farm isk for 10 or 15 hours per week, so a bot will outfarm even a relatively active player by a factor of 10x to 15x. I have only explored and studied 8 of EVE’s 64 regions. They are among the 8 most sparsely populated, which makes them the easiest to study and observe. There is no way of extrapolating from this sample how widespread botting is in the other 56 regions. I certainly don’t think that 18% of online characters throughout the game universe are bots; given that they’re at least 10x more productive than honest players, the game would become unplayable. However, given that CCP does not appear to take even the most basic steps of checking up on accounts that log in 23 hours per day for several days in a row, it seems easily plausible that 3-5% are, which would be sufficient to introduce very large distortions into the game.

CCP says that they take this sort of thing seriously. Perhaps in their heart of hearts, some of them do. Maybe it just never occurred to them to run a query that asks: which accounts have been logged in and undocked for >45 of the past 48 hours. It’s possible that some bureaucrat has tied the hands of the GMs to improve PCUs/subscriber numbers/PLEX prices, or burdened GMs with three pages of hand-written forms to fill out with a 15 step process per banned account. Who can say?

You can read GM Grimmi’s dev blog, and you’ve now read this. You can add the bots I’ve listed to your contacts list and watch them yourself. You can hop in a shuttle or a covops or a t3 and go visit the constellations and systems I’ve written about. Check my numbers, please. Come to your own conclusions.

PS – If you’re particularly good at extracting data from the API like dotlan does, I’d be interested to hear from you.

-Fromage

[spoiler effect="slide" show="Extras"]

I don’t think CCP wants to take care of botters. Think about it. In your list there are 82 individual accounts (there may be 83 to 85, but I couldn’t tell the separation). Each count is paying 14.99 (or stimulating the EvE Economy through time card purchases). That is $1,228.18 a month income they would lose from those 82 accounts, or $17,750.16 a year. You can bet there are just say there are 3 x that many botters in all of EvE, 246. That is $3,687.54 a month or $44,250.48 a year of loss revenue. Seems like chump change, but in the grand scheme of things, everything adds up to thousands more.
Vortura

List of Known Bots:

Outer Passage:
Kitsunya
BiJim
Ko0kKi Mega Mujik Valnor Bane

Perrigen Falls
Balzat HunterNsk

More Spire Bots

Beant Jester, EG Zentensuken, Billy Gatesey, Dolce Gabana Jack Freestylo Lena Kanto Vinyl Chloryde Samwise thin

——————————–

The Spire
Dj Vasyliy, Toma Tilde, Xxx Homie
Bealune Rageham, Cisireb, Dolis Sharp, Ritestond
Dalaan Fearles, Faniquid, Manakelv Chillray, Tkacrope
DarkShadowXL, Golden Ravenn
Prudekek, ierbivoru89, Electric Fin
CybershadowX, DoomPilot1, Tegar Whitefury, Whisper Bringer
Arandetta, Atretori, Brich’Es, Ketoch, Swiftycobra

Kalevala
Artiom Cerkas, Bitea Tiomnii
Stas Pieha, kalyan1, Gena Darkovic, calacolcik1
Galustian, goretzz, Kristina Balaban, Malaya11
Banea Damso, bazazaa, Felip Flamenko, syllyaa
bazucca, Bella Be, Karolina Dimkovic, Medofia
Misa Arcaii, sharlotte joness

Malpais
Antuan Calisto, Sivlak Sidilak
Arab BamBam, Aram Danu, dominion940, FlameStrike1, Kirdik Maru, SciMass
FlameW1nd1, lies hrum, niiel para, ThunderGuy

Cobalt Edge
DoriGrinn, Mirgu Krea
Klem Almer, FelHariel
Masha Monstr, Taran Treg, KaginN, TrueWeaver
AdronnMari, Ardailas, Lulu Mega, Misha Vase4kin

[/spoiler]

Send us Intel/Corrections via dropbox or shoot us an e-mail