Botting 24/7 doubles ban risk, and other findings

[JB.AIOBot]

Hi ElitePVPers,

I just ran a statistical analysis on the voter data from of AIO Bot users. The results should be generalizable to other bots. For example, Honorbuddy that showed a 37% overall ban rate, although they didn't collect as much data on other risk factors. AIO Bot's ban rate so far is 30%. Our poll is still open; right now it's at 63 votes.

Here's what I found:

1. Botting 24/7 doubles the risk of getting banned.
2. Using additional non-AIO Bot hacks or bots increased ban risk by 1/3

The data suggests that to be safe, bot anywhere between 4-16 hours per day. And don't use additional hacks or tools: 40% of those who used them were banned, vs only 30% who didn't.

Details:

• 43% of 24-7 botters have been banned.
• 21% of 8-16hr botters have been banned.
• 18% of 4-8 hour botters have been banned
• 50% of 1-4 hour botters have been banned (there were only 4 votes in this category)

Why does 24/7 botting double ban risk? Here are my guesses:

• User babysitting of the bot fails above 16hrs botting per day, leading to more obvious bottish behavior and more getting caught, AND/OR
• Blizzard takes into account extremely high play times when searching for botters and evaluating flagged accounts.

The sample size for the 1-4hr botters is too small to conclude anything, but it may mean that a minimum level of user engagement and competency is needed to prevent errors leading to a ban.

Note on Method:
Around half of the survey votes are incomplete or self-contradictory. And there are massive self selection problems, both in who took the survey and what questions they decided to fill out. So these numbers, though informative, are just a rough guide.

I would be remiss if I failed to point out that correlation does not equal causation. 24/7 botting or use of additional external tools may be simply proxies for some other variable like "botter hardcoreness" that is driving the bans.

EDIT: Linking to the probotters domain is apparently verboten, so you can see the poll by going to our forum and clicking on the "informational polls" section. It's open to the public.

[Pexus]

I stopped reading after about 20% of the post. This is bullsh*t.

I doubt that you can compare the results with ANY poll EVER made.
You would have to force EVERY user to vote.

• Imagine you're a "leecher" botter who never posts on forums and gives a f**k about polls and stuff. You wouldn't vote.
• And now imagine you're a angry botter who has been banned recently. Ofc yo would vote. They should see that You've been banned to improve any anti-detection systems.

That are just two examples. There are plenty more. So please stop announcing that AIO is more safe than HB. (Don't say that's not your goal. It is...obviously.)
Furthermore getting banned is not only about bot security. It's also about dumb users who are not smart enough to choose a secure place to bot.
I'll end my post like I started it.
I doubt that you can compare the results with ANY poll EVER made.

[JB.AIOBot]

Hi Pexus,

I put my note on methodology at the end, so I guess you didn't read it. I agree with most of the points you've made. Self selection is a huge problem with the survey, and the 50% inaccurate/incomplete responses is another, so the numbers shouldn't be taken in an absolute sense.

As a comparison, though, and for showing what types of behavior increase risk, I think the numbers are informative.

First the poll comparability issue. I think the AIO Bot poll is directly comparable to the Honorbuddy poll. Both measure hours spent botting in the same way, and both ask whether the user has been banned. The AIO poll goes into greater detail on risk factors, but this shouldn't damage the comparability of the results.

I don't think the numbers say anything about warden detectability for either bot. The biggest takeaway is that going from 16 hours/day botting to 24 hours/day botting doubles your risk. That's not a Warden issue, but it might well be a behavior profiling issue.

The difference between AIO and HB's overall ban rate could simply be that HB is more popular and so has more unsophisticated users getting themselves into trouble, as you said.

[floha]

I agree with Pexus. Either the bot is detected or not. Not vote ever can show how risky or detectable a bot or any other hack program is. In my opinion it only depends on the user who uses the program. For example: Since 4.0.1 i'm botting with AIO 24/7 - no ban yet. Before 4.0.1 was released i tried nearly all hacks published on this forum - no ban.

So if I understand you correctly

Botting with AIO has a 30% ban chance.
24/7 botting doubles it so my chance to be banned is 60%.
I used other hacks, so +1/3 ban chance -> 80% ban chance for me.
Let's take another few whispers while afk -> nearly 100% ban.

But till now i'm not banned.

So, just be careful what you use and how you use it.

[JB.AIOBot]

A-Jay,

Actually, your group has a 40% ban rate from the survey data.

I agree that it all comes down to the user. Unfortunately I couldn't figure out a poll question for "do you do dumb things likely to get you banned."

The other interesting thing - and I haven't run the numbers fully yet - it didn't seem like buying or selling gold had much impact on ban rate.

EDIT: I tried to look at intersections of non-AIO tool use and time botting, but it seems the effects don't stack. Also the sample sizes were getting too small due to the incomplete answers.

[hathoor]

Fail

User babysitting of the bot fails above 16hrs botting per day

Some Bots dont need to be babysitted...

ah and

yet - it didn't seem like buying or selling gold had much impact on ban rate.

I suggest to research that again because thats not really the chase

[JB.AIOBot]

Originally Posted by hathoor

Fail

Some Bots dont need to be babysitted...

ah and I suggest to research that again because thats not really the chase

Depends on the activity. AIO's questing and solo instance farming require more supervision than its grinding, gathering and PVP. Other bots can't quest or solo instance farm, so babysitting them is irrelevant.

Per your suggestion, I analyzed the gold buyer/seller ban numbers. The numbers will surprise you...

Gold buy/sellers had a 16% ban rate (4/25)
Non gold buy/sellers had a 33% ban rate (9/27)

Botters who buy/sell gold were banned half as often as botters who don't.

Furthermore, every banned gold buy/seller was either using non-AIO tools, botting 24/7, or both.

My only explanation is that gold sellers are pros who know what they're doing, and that it's mainly the noobs misusing the bot who get themselves banned. Pros get banned only if they get too aggressive.

[Pexus]

Originally Posted by JB.AIOBot

My only explanation is that gold sellers are pros who know what they're doing, and that it's mainly the noobs misusing the bot who get themselves banned. Pros get banned only if they get too aggressive.

True.
That's what I was going to post..

[hathoor]

Thanks for calling many Users Noobs..

[JB.AIOBot]

Originally Posted by hathoor

Thanks for calling many Users Noobs..

I should have said amateurs, not noobs.

Someone asked me about the standard error and sample sizes for the various conclusions. Turns out that the gold-seller conclusion was the weakest one. Here's the entire data set.

The standard error for Honorbuddy's 37% ban rate vs. AIO's 27.2% ban rate was .01%. I didn't get the decimal wrong, that's one hundredth of one percent.


| | count | bans | ban rate | std error |
|---------------------------+-------+------+----------+-----------|
| AIO safer than Honorbuddy | | | | |
| AIO | 92 | 25 | 27.17% | 0.01% |
| HB | 199 | 74 | 37.00% | 0.00% |
|---------------------------+-------+------+----------+-----------|
| non aio tools | | | | |
| other | 15 | 6 | 40.00% | 0.27% |
| aio only | 23 | 7 | 30.43% | 0.13% |
|---------------------------+-------+------+----------+-----------|
| time botting | | | | |
| 24/7 | 14 | 6 | 42.86% | 0.29% |
| 4-16 | 30 | 6 | 20.00% | 0.09% |
|---------------------------+-------+------+----------+-----------|
| gold | | | | |
| non seller | 27 | 9 | 33.33% | 9.07% |
| seller | 25 | 4 | 16.00% | 7.33% |

And here it is as an image in case the table is hard to read:

[hathoor]

I should have said amateurs, not noobs.

These Noobs pay your Bot and others