bot count: 44%

The survey was done on a lot less that 1% of the sims, and the population distribution is lumpy through the grid - not homogenous. Also the 10% LL figure is something different - it's 10 to 15% of *user hours* - not logged-in avatars (unless there's another statement that I'm not aware of).

And that's the point. With most of world awake, the 44% figure means that all but ~1000 of logged in avatars at low concurrency are bots. Common sense dictates that it isn't true.

At the end of the day, people owning or renting land pay the way for the bot resource usage. When our rates go up for resource usage, you better believe there's a problem.

I don't care if it's 44% or 0.44% bots, if my rates are going up astronomically due to resource usage issues, it's time to start asking the hard questions. That includes non-bot use such as people camping, too.

It's not a paid/unpaid member situation; there are a lot of renters who aren't payment on file accounts. I'm cool with free visitors so long as the usage isn't egregious, severe, or milking me dry.

But I got a huge problem with the bots when they are abused like this. Hey, I got more than a few spare computers and lots of weekend bandwidth, should I get on the gravy train too? There's tons of $L to be made with bots, and just about any one of us could launch a few hundred.

We can either fix it by policy and paying attention now, or we can just be sort of dismissive until it's too late and the hammer falls, as we have seen in cases elsewhere.

Bot runners should be the *first* people to validate legitimate usage before someone comes down with a ridiculously overwhelming anti-bot policy to fix the situation.

I understand your point about land per account but I don't believe the social aspect of it. People are not limited to one place in SL. If a place is worth going they will go there. There is just a lack or worth while places to go and come back to. I have been to many nice places but have no desire to return. Been there done that type of thing. I have only a couple places that I choose to return to over and over. Most of SL seems residential anyway so of course everyone is spread out. I don't think most people want to be stuck in a crowd. Not like people come here to actually be social with everyone. That is what separates a 3D world from a text chat room. In a 3D world there is more to do than sit in one place and try to talk to everyone. In a chat room that is all there is to do.

I think you are thinking people come here just to socialize with everyone? I only socialize until I find 2-3 friends to hang with. Anymore and there is not enough time to spend with them. Most of my time is spent with a couple people in a private sim. I prefer it that way and so does most the people I met.

You misunderstood what I said, Desmond. I said that it's debatable whether or not not promoting the number (44%) is harmful. I said nothing about the bots themselves.

The sky is falling...the bots are taking over...yawn.

`



Soon it will be a Virtual Virtual World.

hey!

today i visited 1 private sim

result:

1 sim
1 person, x_amount of them bots and campers

a sim with 5 or fewer persons = 18% bots/ campers

Maybe traffic isn't the motivator for new bots anymore. They could all be BuffyBots (http://www.your2ndplace.com/node/1331).

LOL i now got me a BuffyBot thanks for posting the link, interesting piece as well

The survey you propose would cost a fortune. It took ten hours to collect the data that I got. You're proposing about 57 workdays of effort.

The problem is that a raw survey of locations isn't quite adequate. If, say, the 100 most popular teleport locations are 99% bots, it doesn't particularly matter that there may be many locations that are all real people; the majority of actual users playing SL will be experiencing the presence of bots most of the time.

I think it's a mistake to conflate use of bots by shops with the damage caused by their usage in the social situation, though. No one is going to Prim Savers, for example, expecting to socialise.. but the clubs that use bots or campers, people are visiting for that reason.

How fun! Now I have a bot too!!! Protecting my customers from vampire attacks is a good task for my Puppet Army Infantry.

I don't count all AFK people as bots. I occasionally make Turning tests, but that's just to satisfy my curiosity. It doesn't affect the numbers.

Bots aren't hard to spot. You're going to miss some, but so what? All that means is the number is probably higher than I report. Nobody dies.

It's not worth watching avatars in a private home to determine if they're bots. Why would there be bots there, and if I miss a couple by not looking, so what? I've checked the airspace for stores and my experience says that doing so rarely convinces me that people in private homes are bots, so I don't bother anymore.

It's not worth looking in a busy club, because as you say there's no way to tell who's a bot and who's just AFK. Some of the large groups of humans in my numbers are clubs. I have re-visited clubs to see if the same staff is there, but this has never convinced me they were bots.

It's not worth watching the avatar on the pose stand against a photo backdrop, or the one with particle effects hitting an under-construction boot or house, or the one pointing at a sign in a store, or the one walking or flying or driving a vehicle, or...

When you eliminate all the ones that probably aren't bots, and the ones that obviously are bots (e.g. twenty unrezzed avatars in an empty box over a store, eight stuffed in the roof peak, four on a bench with text in their lap) there are very few that actually warrant investigation. You write a note and you investigate hours later. In this way you eliminate suspects and confirm others. The numbers of these questionable cases is small, so even if I get it wrong it doesn't change the numbers substantially. I think in ten hours of looking I re-investigated about 20 parcels, and after eliminating more avatars from consideration these questionable cases amount to perhaps 10 avatars in the 312 bot number I reported.

An example would be security officers sitting at desks in otherwise empty offices or clubs. I found perhaps five of them. I always re-investigate, because it might be the owner in IM or something. Likewise, a DJ in an empty club warrants a second look. I only found one of those, and eliminated him on a second visit because though he was still there he'd moved and was no longer DJ'ing.

That is, I'm conservative to avoid false positives, so the actual number is likely higher, not lower.

Rather than surveying one continent (which might be easier to implement), a true random sample would be to sort all the sims on the grid by some method (alphabetically for example), and choose every 50th one to investigate. That would be a sample size of 600 or so, which should produce an expected error (one sigma) of 4%.

I would visit at the low and high points of the day, to get two data points, record names in each sim, and then assume if an avatar is there at both times they are "not an active player" (or need to get some sleep . The best way is to do a subsample where you investigate each avatar to classify it as active or not, and figure out what "features" will give the most accurate estimate with the least work.

Why it matters, as Desmond Shang pointed out, is if they put a load on the whole grid which makes the game more expensive for the real players. If an island owner only affects his own island, well, she is paying for it, so it would not matter.

The other case for why its a bad idea, is losing new players. If they go to a place they think will be full of real people, and find its not, they can decide "this game sucks", and not come back again.

In reference to the Linden figure of 10-15% of *user hours*, that would be an average concurrency type of number, since user hours is just the concurrency summed over a day or a month. Since average concurrency is around 55,000, that implies 5500 to 7750. So a simple way to disprove that figure is visit popular places and count bots. If you go well over 8000, then the number is disproved on inspection.

What I wonder, isn't visiting 219 sims on one day very straining on resources? Much more then running 20 bots in a skybox

On a more serious note: I cannot even start to understand why anyone would spend an entire day on counting bots. Really. But hey, that's me.

What is pretty certain, is that the results are not by far representing anything. The OP does not like to hear that, something I can understand. But it is true: the 44% says nothing to anyone knowing remotely anything about statistics. It says that at a given time span, 44% of the avatars in sims visited by an avatar, were probably bots.

When tonight I visit the estate I live on, and Desmond' s Caledon, and I find 0 bots, is that a valid number as well?

…and some people get their kicks making virtual beds that no one actually sleeps in!

True, but at least they sell, so if the building is not the kick, the selling can be

then one has to wonder why somebody would give away a script they spent hours on for others to use freely in their beds and whatnot then, yes?

What I find spooky is when someone/thing I know is a bot, just know it, turns to me and starts talking to me. There is a momentary "oh darn" (different word used than darn) when you realize. I for one see them as furniture or the like, something to let my guard down around - as in I am not super conscious that they are there as I am when I know others are around. But then I'm odd and have problems with crowds.

So to have the furniture turn and start talking with you, to the point you realize they actually are real and there . . . while you had your guard down. eek

The survey I propose could be got from LL's own server logs. Just they can't be bothered to do the work.