Jin’talks—Why is EVE Easy to Bot?

Jin'taan 2019-05-16

Art by Redline XIII

It’s an unfortunate fact that bots are a loathed part of every multiplayer game I’ve played. From idlebots in Team Fortess 2 farming items and ruining team balance, through leveling bots in League of Legends gumming up the 3v3 queue, to mining macros in Runescape. But it seems botting is something that uniquely effects EVE Online, and as a result of that it’s something that is rallied against at every turn, as we saw with the relentless finger-pointing over who harbors the most bots over the past few weeks.

This comes down to the fact that not only do bots generate an advantage for their user, they also serve to lessen the advantage other players get for the same amount of work, as both bots and players produce by and large the same thing. This means players have to deal with bots devaluing the work they themselves put effort into doing, which is understandably a frustrating thing to feel, as it can mean the difference between being able to play as an Omega instead of as an Alpha for players with limited time availability.

Something I’ve noticed in those other games about the bots however is that they were significantly worse than players. Idlebots were easy to kill after the game was over, LoL bots were free wins to the point it was boring, even Runescape bots could be tripped up if they were hogging one specific spot. This contrasts with EVE, where players typically complain that bots are nearly impossible to catch, which is something that clearly contributes to the perception of bots being a problem within the community.

Given that other bots perform so poorly vs players, and bots in EVE seem to perform so well, I am going to dive into a little bit about why I think this is the case. But to give a brief overview of my main conclusion – EVE doesn’t have a botting problem. It has a game design problem.

How Do Bots Make ISK?

There are a lot of bots for different purposes in EVE, and whilst Intel bots and DPlex bots are problems, they aren’t nearly as widespread or as economically impactful as the big four;

  • Mining bots
  • Market bots
  • Mission bots
  • Ratting bots

Unfortunately I don’t have a great deal of experience with the first three, and I hope other people can pitch in with their own expertise on those subjects, but for now I’m going to focus in on ratting as an income source. Ratting is also the main way in which bots add raw ISK to the game directly, which causes everyone else’s ISK to be worth less, so it’s also one which impacts every single player by effectively raising the price of PLEX.

Ratting bots are also a lot easier to find, as they have to spread over a wider range of systems than the other three, which can be centralised in one particular location. In fact I’d wager most people who have even simply roamed Nullsec believes that they’ve encountered one, regardless as to whether or not that is what happened, which is symptomatic of the overall problem.

Bot accounts can be trained on a large scale as alpha clones, or injected up to being at what the person running these bot accounts believe is an optimal skillset. These fresh accounts can then be applied to corps with access to Nullsec, either by purchasing rental space, or simply joining a corp/alliance that has existing access to Nullsec that’s good to rat in and an open doors policy.

These bot accounts are then placed in a ship such as a VNI or a Gila, as those are the most cost effective ways to make ISK in the game currently, especially considering that they can be piloted by Alphas and still rat using the same strategies as an Omega pilot. This is done by simply warping from site to site, dropping drones, and killing ships in a pre-determined order that matches the known spawnlists of said anomalies.

If an unknown pilot enters local – or a local that the bot has access to via a relay – it will immediately pull its drones in and warp off, then wait for a set amount of time, before warping back to the site and continuing.

Why Is This a Problem?

I don’t know about you – But that doesn’t sound any different to how I fly a VNI or Gila myself.

Bots are able to do everything that a player does in order to maximise their efficiency and safety whilst ratting, because almost all of it simply relies on the ability to press buttons in the right order, as fast as possible. Due to the mismatch in optimisation between PvE fits and PvP fits, there’s no reason for the PvE ship to engage a PvP ship (as it will lose), and there’s no reason for a PvP ship to engage in PvE (it will suck at it).

The static spawn lists that contain no tackle that can’t be easily dealt with or avoided means that there’s absolutely no thought required in running the content either, which means running anomalies boils down to a simple binary operation of ratting when local is clear, and getting safe as soon as possible when it is not. Money is even transferred directly into your wallet for every single rat you kill, meaning that even if you are caught, you retain all of the money you made so far. The cherry on top of this is that bots are always paying perfect attention to local, unlike players who can focus in on watching Netflix in their other screen for half a minute too long and end up getting caught, they’re able to warp off the tick you enter local every time.

Currently one of the few ways to deal with bots that exposes the sheer lack of decision making that it was necessary to program the bots with is using log-off bubble traps in their safespots, which I demonstrated in a video last year. As you can see, once you’re able to lay your hands on a ratting bot, it’s almost trivial to take out.

The way in which I was able to do it also demonstrates how poor bots can be when compared to players in terms of decision making. In that particular video I had logged my Sabre off in front of those same VNIs as they sat in the PoS, then simply waited for the NPC/h deaths to go back up on Dotlan, logged in and killed the one I was able to catch. A player would almost certainly never do that, and would at the very least consider changing systems/safespots.

This shows the main advantage players should be able to leverage over bots, their ability to adapt and make better decisions based on the information they’ve been given. However, as we explained above, the best way to keep yourself safe whilst ratting is simply to not get caught in the first place.

What Can CCP Do About It?

Not as much as I’d like. I doubt CCP is going to entirely strip and replace the anomaly system in the foreseeable future, so I wanted to take a look at a fix that I think could be done with CCP’s existing structure and technical limitations. I also think it’s something that would be of benefit to the health of the game in general, regardless of it’s impact on bots, which I think is an important thing to consider—Penalising regular players to own the bots isn’t a good strategy long-term in my opinion.

But that’s only one option. There are others, and they have their downsides as well.

Eliminating Renting

This has been bandied about, but there’s a problem: it’s basically impossible. There are just too many ways to transfer value in EVE. Eliminating rental fees just means you move the payout to market fees and fees to join the ‘rental’ group. Or to get onto the ACL. Or any one of a number of other ways to pay. Yes, all of these things can be tracked, but at the same time, they can take enough forms that anything that includes ‘you do X and we don’t kill you’ can be the de facto rental agreement.

To give a recent example of just how hard it can be to draw the line on renting, many reading this article will remember how Sort Dragon was mocked as a ‘renter’ after paying the Imperium for an end to his last war with them. Whilst that was not entirely serious, can we expect Team Security to understand the nuance of a large amount of ISK being transferred not as rent, but as part of diplomatic tribute – Or conversely, that the pomp of something like this wouldn’t merely be used to cover up the now ‘banned’ renting practice.

Make the Alliances Do It

This runs into problems, too. For this, we’ll just go through some points:

As recently noted by Elo Knight, for many years the leader of Black Legion’s various forms, Alliance leaders do not have tools to monitor for botting activity that Team Security has. In addition, most bots do not rat 23/7. They’re not that obvious. As such, all accusations will have to be done based on hearsay and suspicion. So rather than reporting this to CCP, Alliance leaders are now forced to immediately kick upon suspicion. This is because, as CCP Peligo’s reddit post indicates, the wallet impacted is the main Alliance bill wallet. If the wallet is empty when a bill become due, all Sov will drop.

This, in turn, means an end to open door recruiting policies, realistically. People who wish to rent will set up alternate ‘client’ alliances (ala B0T/PIBC) which are not run by the same people as the parent alliance to protect their sov. And that means that this actually achieves nothing: the ‘alliance leaders’ being given the responsibility aren’t actually anything of the sort. As such, this mainly impacts large alliances, without impacting alliances which contribute more bots overall to the ecosystem.

And then there’s the metagame: weaponised botting. Using VPN, groups can put ‘rental’ corps into their enemies renters, then bot up a storm. CCP then punishes the targets of the meta-scam. CCP gets meta’d. The initial community reaction from the masses will be great—most players only look at immediate intentions and don’t think of the bigger picture. But in the wake of the Brisc Rubal episode, does CCP really want to step into that pitfall when organized groups use Team Security to wage their wars for them?

Another Idea:

The TL;DR is simple – Remove a significant percentage of bounty payouts from all NS anomalies, but add a guaranteed spawn at the end of each anomaly, which holds the equivalent ISK in Overseer Effects.

I’m always hesitant to add numbers to ideas this early on, so let me know what you think the breakpoint of percentage would be there, but adding a physical component to the rewards that ratting provides would have a number of positive effects;

Firstly, it will give an immediate option of a reward for players who are able to push a PvE player off of a site, regardless of their ability to catch the player. As someone who has hunted nullsec ratters (both botting and not), there’s nothing more frustrating than seeing a VNI enter warp just as you land, knowing your work has been for nothing. Adding a potential reward for the intruding pilot if the site is near completion, by allowing them to get a reward that the PvE pilot helped them work towards. Currently the total reward for a Sanctum is only 40m, so this reward is unlikely to be a huge motivating factor for older players, but it will provide a way for players who specialise in hunting bots to gain an income even if they fail to secure a kill.

This also has a number of knock on effects to the way PvE plays out. It adds an effective “upper limit” to how fast you can clear sites whilst still making sure that you have 100% ISK retention, as you’ll need to stop to collect the Overseer’s Effect in each anomaly unless you wish to use MTUs or alts to pick it up. This in turn then makes defending these systems and stopping hostiles from getting inside them more valuable, as it allows you to better make ISK if there’s an active defence force keeping hostiles away from your system, as them entering the system will leave any MTUs or unlooted Overseer wrecks easy to be probed and looted, taking a percentage of your hard earned ratting ISK for themselves—If they can get it back to Hisec!

I’m curious to see what you think of this suggestion, and with a wider lens the problems more generally outlined, as I want to be able to give CCP direction and feedback on how they can let us – EVE players – do what we do best;

Exploit predictable behaviour for our own gain.

Let your voice be heard! Submit your own article to Imperium News here!

Would you like to join the Imperium News staff? Find out how!

Comments

  • Eve Talaminada

    There is one basic and efficient counter-botting tactic against ratting bots, a tactic that probably is not triggering any EULA breach.

    In theory, as ratting bots are triggered by a new pilot entering local, the simplest way to counter-bot with a bot is to infiltrate an alpha pod in a system, then to use a bot script to start the game for that account/pilot at random intervals of time, and disconnect the account shortly after it was connected, creating a random pattern of incoming in local that triggers the warp to safety of the bots.

    Bots spending their time warping in and out of safety do not rat.

    The fact that the script applies to starting and stopping the client rather than using macroed actions within the client is possibly shielding counter-bots from any EULA infringement.

    This counter-botting is designed around one of the guerilla tactic our tribal warriors use, as part of their cloaking campaigns, around the constellation where we have deployed our R.I.G.

    May 16, 2019 at 12:07 pm
    • Arrendis Eve Talaminada

      Pretty sure using automated tools to interface with the client in any way is a EULA/TOS violation.

      May 18, 2019 at 7:11 am
    • Carvj94 Eve Talaminada

      Yea that’s a EULA breach.

      May 19, 2019 at 2:11 am
  • Moomin Amatin

    Bots are a symptom of something else and not the cause. Any of the suggested changes will only lead to the development of more advanced bots, or for people to simply adapt in other ways.

    So as much as botting solutions may seem inventive and robust they are most certainly not. False positives are likely to feature and that is when CCP start penalising money paying customers.

    Just go and have a look at tools such as TensorFlow or OpenAI and how bots are impacting games like StarCraft or Dota to name a couple.

    May 16, 2019 at 4:05 pm
  • Guilford Australis

    This article has convinced me to run for CSM on a platform of eliminating all PVE from EVE. The best and surest way you, my supporters, can make this dream reality is to run a simple program that creates and logs in millions of bogus accounts to cast your votes automatically. Thank you and Spodbless.

    May 16, 2019 at 5:30 pm
    • Moomin Amatin Guilford Australis

      One of the better CSM platforms I have seen of late. You would get my vote ;-).

      May 16, 2019 at 7:11 pm
  • Rovinia

    Drone mechanics seems to be one of the main problems in botting, isn’t it?

    I mean, all the botting ships are either VNI’s, Gilas, carriers and supers… Perhaps the combination of removal of auto-aggro / drone mechanics revamp and an increase of tackling rats in anomalies would at least ease the problem?

    May 16, 2019 at 10:37 pm
    • Carvj94 Rovinia

      That’s cause they don’t use ammo and damage type can be whatever you want. Not so much the underlying mechanics.

      May 19, 2019 at 2:09 am
  • Perhaps this is just nitpicking, but this was a curious statement…

    “…so it’s also one which impacts every single player by effectively raising the price of PLEX.”

    The last I heard, which admittedly was before the introduction of alpha accounts, less than 25% of players PLEXed their accounts. I’d really like to hear how the rising ISK price of PLEX/falling real world currency price of ISK affects those who don’t deal with PLEX.

    May 17, 2019 at 8:28 am
    • Arrendis NoizyGamer

      Now that PLEX has been changed from OldPLEX (1 month = 1 PLEX) to NuPLEX (1 month = 500 PLEX, and PLEX is the currency used to buy SKINs and other items (like Skill Extractors) from the New Eden Store), a lot more people deal in PLEX then just the people using it to PLEX their accounts.

      May 18, 2019 at 7:08 am
  • Carvj94

    Since we know most bots are triggered via local CCP could just add ghost pilots to randomly show up in the local of null systems that have ratting going on. Implemented so that they aren’t visible in the client but would be seen by a botting program looking at the data. With a bit of work they could set it up to automatically change up the value that defines the pilot as a ghost at maintenance every day. That way botting programs would need to be updated daily to recognize the new value else they’ll be less effective due to auto docking. Would also make it ridiculously easy to spot a bot as they’d be docking to avoid an otherwise invisible pilot.

    May 19, 2019 at 2:17 am
  • dryson

    Maybe the alliances should declare a universe wide ‘Bot Hunting Event’ where Capsuleers pay 100 mil ISK to freely hunt bots in the belts and at combat sites. The Bot Hunting Permit would be either NRDS or BDSI (Blue Don’t Shoot It) for the number of days that the permit is valid, meaning that if any member of the issuing alliance shoots the Bot Hunter down, the alliance must fully repay the Capsuleer for the ship, module, ammo and implant, along with any other losses involved in the kill.

    The Capsuleer will also not engage in any PvP fights for or against the alliance issuing the permit. If the Bot Hunter does engage in PvP for or against the alliance issuing the permit, then punishment is up too the alliance that issued the permit.

    Maybe the alliances could even set up certain times during the year where hostilities between all sides ceases to allow for Bot Hunting as coordinated effort.

    With Bot’s being a nuisance as they consume resources for the alliance, I’m surprised that alliances haven’t started using the Bots against other alliances to drain the enemy alliance of much needed ISK from mining and ratting the belts.

    May 27, 2019 at 5:21 am
  • dryson

    “Would you care to assist me in performing surgery on a torpedo?”

    The first is too establish an interface with the arrogance of the human bot master, this would involve looking for key word clues on the forums.

    The cherry on top of this is that **bots are always paying perfect attention to local**, unlike players who can focus in on watching Netflix in their other screen for half a minute too long and end up getting caught, they’re able to **warp** **off** the **tick** you **enter local** every time.

    **bots are always paying perfect attention to local, warp,off,x location,tick,enter local**

    In this case the word **tick** is being focused on, or the machine learning program using an algorithm of rate my ticks. My ticks being the interactions between the Bot and Object ID’s in the same system of the Bot.

    The definition of **rate** is; the degree of something measured per unit of something else that creates a change over time that can be analyzed . In this case the rate of ticks of Capsuleer ships entering into a system occupied by Bots is measured by the ratting and mining bots as not being NPC ship ID’s.

    Each Capsuleer ship has a unique ID assigned to it that is assigned to the pilot flying it. NPC’s ships only have unique ID’s assigned to them as being NPC’s ships and not Capsuleer ships. So in this sentence alone we have four different **ticks** that are rated by the Bot machine learning program to determine if the unique ID is an actual Capsuleer ship or NPC ship.

    If the Bot reads the ID as being two ticks or the ships own unique ID + the Capsuleers unique ID then the bots warp off too a safe location. If the scan returns only an NPC ID then the Bot remains active. But since NPC’s do not use Warp Gates jumping between systems, then another tick is the gate activation itself which also has its own ID. An activation ID that can be read using a program that reads how many ships have used a Warp Gate in each system that is taken from relative data from the **Map**. Since the Capsuleer’s ship activates the gate before entering the system with the Bot in it, the Bot has at least six seconds to warp off to a safe spot. When you come across a ship that is at a safe spot and not a station, then most likely that ship is a bot, especially when you encounter them in Low and Null.

    The machine learning program also recognizes other Bots in the system and would not warp off right away when the bot ship enters the system thus creating at least six **ticks** that the bot is using to identify NPC from Capsuleer, Bot from Bot and Gate activation.

    …but

    The main aspect that cannot be avoided by the Bot program is the physical interaction that a Capsuleer uses to interface with the game. Such interactions typing on the keyboard, using the keyboard for purposes such as locking a target or the mouse itself that is an extension of the keyboard to perform keyboard based functions normally assigned to the keyboard, all create electronic foot prints on the server to complete the action. Very fast interactions to keep the process as close to real time as possible.

    Such electronic impulses, although very fast, still take a small amount of time to complete.

    The bot program, unless very well built, does not use keyboards or a mouse like the Capsuleer does that would completely erase the interaction footprint between the Bot and Server but would still exist as a point to point interaction that would take place instantly and would not see any time lapse between the Bot, the command function, or a key being pressed or mouse click and the Server itself.

    If the machine learning program is able to mimic the keyboard function that is translated as an actual physical interaction taking place between a keyboard or mouse function such as selecting a point to warp too and then clicking or pressing the warp function, then Eve Online might be the least of our problems.

    May 27, 2019 at 10:18 pm