[X] My Assistant

2. People have tendency to believe something do/don't exist, if given the choice. Rarely would people NOT to take side.

I've done some work on the subject. Anyone on the 'other side' have done anything?

What the 'other side' are supposed to represent or prove, again? (IMG:[invalid] style_emoticons/default/huh.gif)

This post has been edited by buktore: Jun 21 2012, 06:41

I've done some work one the subject. Anyone on the 'other side' have done anything?

What the 'other side' are supposed to represent or prove, again? (IMG:[invalid] style_emoticons/default/huh.gif)

If I ask you 'Prove yourself you haven't used bot', how are you going to answer?

Ei incumbit probatio qui dicit, non qui negat (innocent until proven guilty)

This post has been edited by varst: Jun 21 2012, 06:36

If I ask you 'Prove yourself you haven't used bot', how are you going to answer?

Ask Tenb. He have all the answer. (IMG:[invalid] style_emoticons/default/rolleyes.gif)


I better stop this; too much 'emotion' flying around here than I expected...

Unless you have an actual account name, it'll be difficult to convince Tenb, too many what-ifs.

If there are 20 people camping the bazaar at the given time you did the test, it can appear to be bots running.

ehwiki down?

No one is arguing it's not possible to get things "really fast," i.e.less than one second, or even near half a second.

(IMG:[invalid] style_emoticons/default/huh.gif)

So, even if you have sub-100 ms ping, it's impossible to buy something in 0.5 seconds after it has appeared without a hotkey macro (...)

(IMG:[invalid] style_emoticons/default/mellow.gif)

It's the specific crystals and material that appear and disappear in inhuman speeds that are causing problems (appear and disappear in 1 refresh; ~0.1s.) And it's not constant, it's intermittent, but it still happens frequently enough to be noticeable.

Before talking about inhuman speeds you should always consider that the item isn't necessarily avaiable the instant you see it on your screen. It's entirely possible it was avaiable just a couple of ms after your previous refresh and while you requested another one and received the page, someone else (who obviously isn't bound by your clock/"refresh rate") has already started his purchase procedure.

In this case (which is more likely to happen than you might think) you are essentially "two refreshes behind" from the start, not one. That's to say anywhere in the 300 ms range. Plenty for someone else to snatch the goods while you wait for yet another refresh.

So if you're unlucky you may easily end up being 450 ms late or more, which is actually enough to snatch certain items "by hand"(*). It's like getting a "red wave" on a street, while your opponent is getting a "green wave". Even if you both drive at the same speed, he'll leave you in the dust anyways.

(*) and we both know that some users openly admitted they're at least using macros in the bazaar.

In the end this fits very well the situation described as happening "intermittently" and not constantly.

I like how people keep insisting that they have deeper insight into what timings would indicate bot usage than I do, what with full log access and the ability to clock actual reaction time directly on the database and all.

Oh wait. No, I don't.

I seeded a few destruction bindings just to demonstrate my point.

CODE

3925189 Query  /* 1340279237.813 */ (seeded)
3868725 Query  /* 1340279243.247 */ (bought)
3925196 Query  /* 1340279263.625 */ (seeded)
3893910 Query  /* 1340279270.510 */ (bought)
3925189 Query  /* 1340279290.259 */ (seeded)
3920380 Query  /* 1340279293.066 */ (bought)

The first one was bought in 5.434 seconds, the second in 6.885 seconds and the third in 2.807 seconds. A total of six bindings (the first three which I didn't clock precisely) were bought by three different people. Those are not the millisecond reaction times you'd expect from a bot or even the sub-second times you'd expect from a macro.

And if the alleged bot wasn't running right now? Well, it's not as if I pre-announce when I do these checks, so the odds of it falling into one of the traps I place is very high.

It probably means you need to provide more crystals through bazaar to satisfy those people's need on bindings. (IMG:[invalid] style_emoticons/default/smile.gif)

I like how people keep insisting that they have deeper insight into what timings would indicate bot usage than I do, what with full log access and the ability to clock actual reaction time directly on the database and all.

Oh wait. No, I don't.

I seeded a few destruction bindings just to demonstrate my point.

CODE

3925189 Query  /* 1340279237.813 */ (seeded)
3868725 Query  /* 1340279243.247 */ (bought)
3925196 Query  /* 1340279263.625 */ (seeded)
3893910 Query  /* 1340279270.510 */ (bought)
3925189 Query  /* 1340279290.259 */ (seeded)
3920380 Query  /* 1340279293.066 */ (bought)

The first one was bought in 5.434 seconds, the second in 6.885 seconds and the third in 2.807 seconds. A total of six bindings (the first three which I didn't clock precisely) were bought by three different people. Those are not the millisecond reaction times you'd expect from a bot or even the sub-second times you'd expect from a macro.

And if the alleged bot wasn't running right now? Well, it's not as if I pre-announce when I do these checks, so the odds of it falling into one of the traps I place is very high.

care to sell some more, i could use a few (IMG:[invalid] style_emoticons/default/smile.gif)

So if I were to write a script that mimics normal human reaction speeds and refresh rates then there's no chance of getting caught?

So if I were to write a script that mimics normal human reaction speeds and refresh rates then there's no chance of getting caught?

Correct way. Easy to do so, right? (IMG:[invalid] style_emoticons/default/smile.gif)

This post has been edited by varst: Jun 21 2012, 19:29

The first one was bought in 5.434 seconds, the second in 6.885 seconds and the third in 2.807 seconds. A total of six bindings (the first three which I didn't clock precisely) were bought by three different people. Those are not the millisecond reaction times you'd expect from a bot or even the sub-second times you'd expect from a macro.

And if the alleged bot wasn't running right now? Well, it's not as if I pre-announce when I do these checks, so the odds of it falling into one of the traps I place is very high.

You would definitely agree it's a bot if you saw it in action though. It buys in milliseconds, but it only appears intermittently. I don't "camp" the bazaar very often, but I saw it three times in a week, for 6-8 hour blocks. And I'm confident it's a bot, because I made a 1-click macro to try to combat it and never succeeded a single time.

And I might as well ask this now since you're paying attention. Are macros allowed? There's no logic in them, just a consolidation of key and mouse clicks, so it seems like a gray area.

It buys in milliseconds, but it only appears intermittently.

That's kind of a shitty bot.
I'm not saying it doesn't exist (or that it does), just that it sounds shitty.

Correct way. Easy to do so, right? (IMG:[invalid] style_emoticons/default/smile.gif)

It seems so. But I don't like the idea of that at all.
I can't think of any way to prevent it though.

I've got a fair share of destruction/slaughter bindings and have never felt there was a bot going for Bindings. Crystals is another issue though and I'm 100% with hito on this. You just know it's a bot if you see it happening yourself.

It seems so. But I don't like the idea of that at all.
I can't think of any way to prevent it though.

That's what I've said.

Also, think about it: which will cause more problem, a miss on those bots, or a hit on those innocent?

That's kind of a shitty bot.
I'm not saying it doesn't exist (or that it does), just that it sounds shitty.

Milliseconds is a generalization since I can't actually see how fast it is without a log...

You should put a counter on the bottom of the bazaar that shows the number of players currently on that page. (IMG:[invalid] style_emoticons/default/mellow.gif)

BTW, one of the solution against bots in bazaar (if there is one) is to apply a captha after a certain number of purchases (for example 5). Easy to implement, and won't screw most of the genuine bazaar campers. (assuming there's even chance those campers can obtain valuable materials from bazaar)

This post has been edited by varst: Jun 21 2012, 20:01

If there is or isn't a bot, you have to keep an open mind about it. Anything is possible with a computer nowadays, I'm sure if a person has the time and knowledge, it's possible. That's just my opinion anyways.

A ban on using that type of macros would be impossible to enforce the way the system works now, so everyone might as well use them to even the battlefield.

So if I were to write a script that mimics normal human reaction speeds and refresh rates then there's no chance of getting caught?

No, but this specific discussion was about how fast the bot reacts.

Incidentally, I do have a solution lined up that doesn't involve captchas and eliminates the need to camp altogether, but I have no idea when I'll get around to implement it.