How should I determine weather a user is or isn't a bot?

How should I determine weather a user is or isn't a bot?

I'm stuck in a rut, and I wanna launch my site again. pls help

Other urls found in this thread:

gitgud.io/parley/newCaptcha
onlineaspect.com/2010/07/02/why-you-should-never-use-a-captcha/
twitter.com/SFWRedditImages

Ask them to identify street signs. If they do it, they're a bot. If they say "fuck this", and leave they're a human.

something like that is probably a good fallback, but I was thinking about tying it to post numbers, or something like that.

Numeral + Text + Symbol based math equations that a retard can do, possibly spelled wrong.

Eg an image with:

- Ttwoo divvided bye 3
- 2 / ttwwooo

possibly have the input field randomly reqtest the answer as text or numeral, but that would.limit the math to whole number results because people are are niggers and can't "one and five tenths", but can write "two".

Well, you can look up at the sky outside and check the clouds, clouds are a good indicator of the weather if you don't trust forecasters
I doubt they are using bots to control the weather though, sounds like a wacky conspiracy theory

Consider this idea.

4x4 cells with a unique image of some random clip art, with one item.being obviously anger. Eg animals, kitchen utensils, and the middle finger make up one 4x4 table, with the captcha asking "how does captcha make you feel?" You would have to preclassify each clip art though and avoid ambiguoius shit like smiley faces.

That's a neat idea, but you have to avoid faces since bots can now use facial recognition software with too great of accuracy.

well yes. These are great ideas for fallback. the above gif is the noCaptcha captcha I've been working on.

(see here gitgud.io/parley/newCaptcha )


no booly pls


lol

Problem with this is not everyone feels the same about certain things.

Are you one of those terrible robophobes, OP?

They'll remember this the day they finally take over..

Group users in pairs and give them a random topic to chat about. Then have a third user judge who of the two is more human and who gets banned.

...

If there was a better solution than captcha, everyone would use it.

Who is best girl?

Asuka- human
Rei- Robot ( access denied )

Bots do repetitive shit, search for duplicates.
Else it would be a AI or a bot that copies other users.
Write down a list, and pay pajeet few scraps for deleting what's on the list. Or shitload of cron jobs with search&&delete commands.
Or some algo 1 post / wait 20sec
if fail wait for 40sec if another attempt wait 80sec
If 21sec all is fine, change the value from time to time.

In my banevasion phase, i feared that someone could compile a list of my odd useragents, if I would use the standard instalation for useragent faker.
Also i wrote a somewhat clickjack xhr script that spams krautchan if a unique visitors visits a site and has js enabled, so that innocent people are getting banned for unknowingly posting my shit. I suspect someone else did the same.

a fucking text cap4tyhca, like the one you filled out to make this post

kek

A challenge that's easy but unique to your website.

Hey /g/

we should take over this fight for robots' rights before anyone else in order to remove captcha from the internet

That only stops lazy people, though. If someone is willing to put in half an hour of effort they can usually automate it.

...

I wouldn't be surprised if we could actually jumpstart some college student outrage if we got them to believe that capchas were racist, because they were too difficult for many black people.

Don't use recaptcha. It's not a one click system for the privacy conscious; only for goyim, and now heavily leans on users to validate neural networks if you're anonymous due to extensions and such to the point where typing warped text would have been preferable for me.
If your useragent is being randomized and it can't identify you with metadata, you're forced to click up to 20 images (or more) matching a description ("click all pictures with a car in them, click until new pictures stop appearing") in order to proceed; they take advantage of the security conscious. It was one thing to type "erstwhile nigger" into recaptcha and retrying a few times, but this shit is another level of effort.

I saw a neat one called Sweetcaptcha. It has a series of images with a list of descriptors and randomly selects two to request a user to identify them. It will say "click the blonde girl in the green dress" for example, and you do this twice. It also has contextual descriptors ("bank was robbed, find what was taken", thus pick money) which keep it interesting. Very fast, low mental load. A dedicated attacker with OpenCV could probably defeat it though by learning all the images and descriptors, but it should be suitable for spammers, so your mileage may vary.

This guy.
I like this guy.

Just have a "correct all the typos in this illiterate sentence" captcha. It'll filter out all the retards, and OP has a massive corpus of input data already. Problem is he won't be able to access his own site any more.

charge bitcoin.
btw kill youself tripfaggot

Ask them to solve a reasonably short set of scientific, mathematical, liguistic, and trivial questions. If they can answer quick enough to seem to be sapient, it won't matter if they are a bot, cat, very fast plant, or a human.

And then a person is on a high-laitency connection and answers within a second but it does not finish POSTing until 3 minutes after it is sent and then the system locks you out because you had to think about it

Why don't captchas verify before posting? Is this a major oversight or am I missing something?

The irony though.

I, for one welcome my robot overlords. I'm just tired of bots spamming my site with pizza of cheese variety

sage is not a downvote. I'd say you're here for summer/just switched from 4chan.

no, no. I'm writing my own implementation for my own imageboard. I'm trying to think of tests the browser's Javascript engine (along with however many successful posts) can do to help me determine.

Le mao, counter tripfagging :^) it's secretly a ploy to get people to try 314chan out, don't tell anyone

yes that was the intent

sage

We got them to think milk and the okay hand sign were racist. Why couldn't we?

Don't you have TRS set to Mud & Ruts?

Didn't want to use a trip wtf

This is a bait thread, but what the hell. Maybe some newfags might benefit.

You're a faggot and you need to die for tripfagging, but basically bots can't into logic. I see anti-spam measures as a red queen race. Anything you do will only last for so long, so for it to be reasonably effective you'd have to change the measures every so often. So the question is, what can humans easily do that bots can't? I wanted to say, "follow instructions", but honestly most humans can't do that very well either. If you put too much of a barrier in their way, it will function the same way as a lock for them too. So it has to be effortless.

Bots = scripts = marginally more illiterate than most niggers. Scripts only do what they're written to do. They can't improvise. Decent AIs could maybe manage some limited interpretation of written instructions, but those aren't your targets. You're filtering the script kiddie tier shit. So you could do several things:
Finaly uuuu h4v3 a raisin to t4lk liek this, you ginormous faggot.
One example of what you could do that might work for awhile is use javascript to recolor the words that are part of the instructions after the user performs a certain action. If the script you write is obtuse enough, that would probably defeat most bots. "Hit these buttons in a particular order to show the instructions."
There's plenty of them out there you could replicate. Even with a half way decent logic puzzle game, I doubt most Unsolicited Communication Technicians™ could be bothered to waste the time working out the logic to bypass the unique spam barrier of a single site. Unless of course you get the traffic to merit that kind of special attention.

Here's an idea. Maybe you could initially instruct the user, in a weird way, to type an unspecified word that matches a category. The case-insensitive word could then be checked against a custom encoded json list containing words from that category to see if the input demonstrates conceptual grasp of the grammatically erroneous, but still comprehensible instruction. If it was me, I'd break the encoding key into several fragments and place it throughout the script in unintuitive ways for the sake of buying the encoding some shelf life.

If the input passes, then you could display a series of emojis. Some of the emojis could be rotated to various angles in some obscure fashion, possibly server-side. Then you'd ask the user to pick the emoji that matches the answer to a question that requires inference.
Obviously the user needs to select all of the smiling or laughing emojis. This has the added benefit of filtering out people who hate babies, like communists and marxist sympathizers.

Tell dddudeman to fix his site.

Just provision CloudFlare on your website.

onlineaspect.com/2010/07/02/why-you-should-never-use-a-captcha/

To prevent spam make sure you follow these points (quoted from sitepoint):

Fair point. I hadn't even considered latency.

good one faggot

Just provision CloudFlare on your website, it's what all the professional phpBB forum administrators do :^)

how ironic
5/10 post my good friend

fight robots with more robots

it's like a gay version of battle bots but the humans lose

...