Bot Count, The Estates Strike Back: 50%

For two days I surveyed the estate sims for bots and campers. I chose randomly, by map location. On the second day of surveying I collected data during hours I did not collect on the previous, so this data fairly uniformly covers 24 hours, including both high and low concurrency times. The raw data has timestamps, so I could break it into low and high concurrency, but I haven't done that. Once again, I'd expect the low-concurrency percentage to be higher than the number I'm reporting and the high-concurrency percentage to be lower.

The results are as follows:

524 sims
953 avatars, 481 of them bots and campers
50% bots and campers, overall

For sims with 5 or fewer avatars there was an average of 13% bots and campers (42 bots out of 325 avatars). For sims with more than five avatars there was an average of 70% (439 bots out of 628 avatars). The overall percentage of bots was 50%.

Some observations:

While visiting it really felt like there was going to be far fewer bots. I thought I'd be able to report less than 40%, but the large bot farms just dominate the numbers. Even if you remove the largest of them (79) from this data set, the overall percentage is still 46%. And the 79-bot farm isn't even the largest I found. One of the 50-bot farms in this survey was adjacent to a 94-bot farm and a 98-bot farm (those weren't included in this survey).

There's an even more dramatic division between low- and high-occupancy sims in the percentage of bots (13% vs 70%) than there is on the mainland (~20% vs ~62%).

More than half the estates I chose (63%) were empty.

I could not get to some estates because they do not have public access. Assuming those are not commercial they'd likely have few or no bots. I assumed any avatars in non-public sims were human. This only amounted to perhaps 20 avatars.

Data collection was cut short by about two hours because of grid borkage. I'd have to wait a whole day to resume collection, but...screw it. Close enough.

Data is here...

http://greendots.typepad.com/green_dots/2008/11/bot-survey-the-estates-strike-back.html

The mainland surveys are here...

/327/cc/293027/1.html
/327/65/293199/1.html

Another round of fascinating numbers Anya, thanks for all your work here.

I don't know enough about statistics to form an opinion on how statistically significant your samples are, or how rigorous your methods (that said you certainly seem to be erring in favor of a higher "presumed responsive" count yet finding loads of bots/campers all the same), but it's difficult not to be struck by your results.

I'm sure the usual suspects will be along soon enough to tell me why your numbers are worthless/wrong

TBH I can't be arsed.

Thanks very much for continuing to study, Anya, and for posting the results.

This was quite unexpected. I honestly would have expected the island bot/camper total to be lower than this.

Are you going to be continuing these surveys?

I think a % of sims surveyed with 10+ bots might be interesting here. Remember, estates can inflate their occupancy from 40 (mainland) to 100. So, someone who wants to run a bot farm on an island can have far MORE bots, and inflate the overall number (and % of population on estates).

I haven't looked at the numbers yet to see if 10+ botfarms are more or less common on estates yet, but I would expect the ability for a single person to run 70-100 bots to dramatically increase the overall # of bots on estates as a whole, even if there are fewer sims with a large number of bots on them.

Brilliant surveys, keep the data coming.

Here's another data point for you - Caledon, 44 regions, no bots.

Well, save for the bots that port in and leave that we can't stop, and I *rarely* run a map bot myself for like... ten minutes every coupla months. And we were thinking of one bot for assisting new residents with groups, and things like that someday, but never got around to it.

This may be a bit of a harsh distinction, but bot and camper crowds can kinda separate classy SL from... the rest. I'll leave it as an exercise for the reader to determine which side is which.

Me neither.

I have allways assumed there would be more bots on the Islands compared with the mainland. With a larger initial investment and tier there is more pressure for a commercial venture to generate a profit so more motivation to game the search system imho.

Same for Ryder. Most of it is residential so bots there are pointless, of course, but the businesses that ARE there aren't allowed to run bots, IIRC. My own store is probably the largest store, and has never had a single bot. I don't even have any alts at the moment, I forgot the password to the one I had

Personally, I see bot usage as kinda pathetic. It means the business doesn't have enough real traffic for the owner to feel like their place is worthwhile. My business has always been in search, but I couldn't care less how it ranks in traffic listings. People find it when they want to find it, and that's all I want. If I ran bots, I'd probably end up with a lot of visitors who wouldn't be interested in the store at all, so it would be pointless.

I expected estates to be more-bottish than the mainland.

If an estate owner wants to park 50 ruthbots in orbit over their store, LL isn't going to care. On the mainland, you can't even get that main avatars in a region and you can bet the neighbors will complain if you try..

I thought these numbers were a bit suspicious so I've been working on my own automated system. I use a couple of bots that teleport around scanning sims plus a network of parcels with 96m scanners on them and have the results here: http://ninjaland.net/botdetect/

I swear I didn't do it deliberately but it so far shows 10% of the avatars are bots.

Since you didn't personally visit each sim, how did you determine who was a bot? No, I am not going to your website.

I wanna know;
How many computers do you need to run 50 bots?

Interesting. I've seen pretty detailed case studies with error analysis that were showing 14% back in the day - I have no real idea what levels of bots and campers there are now, other than 'too many.'

One thing is for sure, there are certainly a few various categories of use and a few cultural divides regarding bots, campers and the like. In the grid I'm familiar with, I almost never come across any bots or campers at all - not even far overhead - other than maybe scanners.

Barely even one if you know what you are doing.

He has a record of an avatar being at the same co-ordinates once an hour for x amount of hours.

So he detects them at 2pm, 3pm, 4pm, 5pm, 6pm, 7pm, 8pm, 9pm at exactly the same co-ordinates.

Certainly not conclusive proof but definitely very telling. Seven times are the lowest number of times he detected one to be counted and they go as high as being detected 24, which is definitely a bot because I'm pretty sure that one is Elanthius' bot! The name rings a bell that's for sure!

Q1: Are you counting campers, or bots disguised as campers?

Q2: Are you counting only avatars found within 96m of ground level? Most bots I see are hidden in skyboxes, so I doubt you'd find many on the ground.

Q for Anya: Are you finding more of the bots in skyboxes than on the ground, and is it more common to see bot boxes or bot campers?

No bots here! All real pixels thankyouverymuch!

Inc

Perhaps you need to feed you bot.

In Gauri you found six avatars over a 24-hour period and called one of them a bot. Well, there are nine camping positions, and you can apparently double up on them. I found ten campers and an eleventh standing nearby. A few minutes later she, too, was camping.

In Alleni you found five avatars and called none of them bots. There are four mannequins at a store at the landing point.

In Abaddon you found eight avatars and identified no bots. I found three filled camp positions.

In Acrocorinth you found 1 avatar and identified no bots. It looks like there are three avatars sitting in an empty club, and someone in a skybox above. I'd mark all four for later investigation, and I think I'd find the three are bots.

Those are just the ones I checked.

There are a lot of stores on platforms and in boxes, so even if the bots and campers are in plain sight, that often means they are at altitude. I'm finding more *locations* that are near stores, but more bots in boxes away from the store. That is, the hidden boxes are swamping the numbers despite being less common than other forms of traffic botting, because they tend to have more bots in them.

On the actual ground? Not a lot of bots, but lots of camping at ground level. Of course some of them, maybe the majority, are bots.

I thought it was lower even as I was doing the survey. The numbers surprised me. What's happening is the large bot farms are swamping the numbers. The estate bot farms are larger.

As I wandered the estates it's easy to see why people don't think there are many bots. Most sims have none (89% of all sims, 71% of occupied sims), and the bots tend to be hidden, and they tend to be in sims that are uninteresting, so you'd likely not encounter many.



No. I'm done with this for now.

Edit:

I corrected the percentages of sims with no bots. I had mistakenly posted mainland numbers (which are 74% of all sims, 58% of occupied sims).

http://www.google.com/search?hl=en&q=slbot&btnG=Google+Search&aq=f&oq=

1st link is interesting.

and they made fun of me when saying there are 30,000 bots heh

I'm counting neither bots nor campers. I'm counting people who move suspiciously rarely.

I have two counting methods one of which gets every avatar in the sim and another gets every avatar within a 96m radius of a point on the ground.