yesterday, we talked about the problem of bots – or, rather, we opened the issue.
Today, we are going to have a look where the rabbit hole leads. Let’s start by Wargaming’s attitude towards bots. Just to be clear, I will now talk about Wargaming EU, I don’t know the bot situation on other servers, but thanks to St0rmshadow (see the previous bot article), we have a good idea about what the situation on EU server is.
Let’s start by the interesting part. What if I told you…
…that running the most popular and probably the best bot (Tankleader) requires a premium account or gold?
Yes, it does. Let it sink in. Yes, that means that pretty much every botter is a Wargaming paying customer.
According to St0rmshadow, it’s actually impossible to run the Warpack bot without premium account, it just crashes. Others allow it, but the bots lose so much money without premium account it’s not sustainable.
According to the Tankleader webpage, this bot has about 130k users. Count with me. Running premium account is not cheap, so these players (since they already spend about 10 bucks on the bot itself per month – remember, it’s subscription based – they might as well paying for the premium) probably buy the premium account for longer periods. So, let’s say half a year.
On EU server, half a year of premium costs 49,05 Euro according to Wargaming premium shop. So, 130k x 49,05 Euro, that’s 6,376 million Euro. If all of these guys bought a year of premium (remember, they don’t have to buy it all at once, they simply have to operate a bot for an entire year, which was known to happen), it’s allready over 10 MILLION Euro revenue. And that’s just this one bot system – as I wrote, there are more.
So, what would happen, if Wargaming developed a memory scanner tool, that would detect all these bots and got them banned? Yep, instant loss of 10 million Euro. Of course, these guys already paid, but they would continue to do so – to operate their bots. See where this is going?
Now, I am not saying there is some sort of “dark bot conspiracy” – I generally like to use the Occam razor principle. According to it, the reason bots are not fought is most likely simply incompetence. But the implications are there and if I can make this calculation, you bet that people at WG RU can too, contrary to popular belief, the fact they hire fools in Paris does not mean WG RU is stupid – on the opposite, they are very, very smart and capable. I will leave the judgement to you, dear readers.
On to the second topic:
Why are bots a problem at all, since some seem to play better than your average Joe Tomato?
This question came up several times yesterday. The answer is quite clear, but I will elaborate nonetheless, since I feel this issue must be addressed. There are many differences between a tank played by an incompetent player and a tank played by a bot.
- a bot will never understand tactics and what is needed for successful victory. Neither will a bad player, but you can TELL him. If you write “PlayerX, please – if you help me on this flank instead of waiting for the enemy in the base, we will win, come on please” (instead of the traditional “MOVE FUCKING NOOB”), you have a chance that the player might actually listen and help you, he might even learn something from the experience.
- a bot will never help you, he just follows a program, see the previous point
- dealing with a (competent) bot requires completely different tactics, that cannot be applied to player. A competent bot cannot be suppressed, cannot have doubts and always aims for weakspots. It also cannot be “tricked” by peek-a-boo. On the other hand, it cannot improvise. Overall, dealing with a bot changes the gameplay style and destroys “feeling” of fighting other players
- one of the nastiest bot influences is the formation of lemming trains. A bot simply follows program – without thinking, it goes where it is programmed to do straight away. If he goes somewhere straight away, chances are other players will think “he knows what he’s doing” and will follow him blindly. This is human nature – we tend to flock to someone, who displays the “I know what I am doing” attitude, only a few people are truly independent and leaders. This effect can completely destroy a battle.
Imagine a situation on Himmelsdorf. Your hightier heavy tank is a bot. Someone programmed it to go to the hill (regardless of the vehicle type) and so it does and many follow it, since “it’s hightier” and “by moving there he knows what he’s doing” (note that we subconsciously accept someone on the “top of the ladder” as a leader and expect him to lead, even though it might be a bot in this case). So, a bot starts a hill lemming train, all the while the other team’s heavies push thru the “corridor of death” or thru the west, which makes more sense for slower heavy tanks. Instant game loss.
There could be many variants of the mentioned situation, but I am sure you get my drift. Bots fundamentally change the way the game behaves and is played and if there are really (literally) hundreds of thousands of these fuckers out there, they might even have some influence on game balance itself. Think about it – do you REALLY want to have this game balanced around fucking bots? What if bots artificially lower a winrate of certain highly profitable tanks, that appear as “alright”, even though they seem OP to the players? Such a case would of course be an extreme one and I don’t think this really happens, but just imagine it.
I am sure we now agree that this is a bad thing. so, what can be done?
Plenty. First and foremost, eliminating bots based on reports only does not work. There are obviously too many and they are (using naked eye) not that easy to distinguish from bad players. It’s simple math – to certainly uncover a bot, a bot hunter has to watch the replay in detail. With taking notes and re-watching some moments, this takes probably more than 15 minute for ONE bot. I asked St0rmshadow about it – some primitive bots take him literally seconds to uncover, but the insidious ones can take several minutes – and he’s an expert in bot hunting. This is obviously not effective.
Automated means could work better – for starters a bot memory scan like World of Warcraft has. It works fairly well, despite bots evolving all the time. It’s a vicious circle, but it also is a battle that has to be fought. Another way would be to automatically alert WG bothunters to players with suspicious number of battles per day. A guy with 250 battles per day for an entire week straight? Naaaaaaaaaah, it’s not like that’s suspicious AT ALL…
Currently, Wargaming checks the battles per day ration, but only when the player gets often reported as a bot. This should be automated somehow.
Anyway, those are my thoughts. Next time, we are going to have a look at some infamous EU botters in the Bot hall of shame.