Botting launches - $wink

Rugpulls: where bots go to die

Background

This post follows the launch of $wink, a meme coin on Avalanche. This is part 2 of 2; you might be interested in reading the guest post on the topic by 0xSmitty.

Additional context can be found in my series on the launch of $ket. I’ll try not to rehash that content too much here.

Anti bot mechanics

As part of the Wink launch, there were a number of fake LPs spun up before liquidity was pulled from them. The team warned about buying LPs, in advance, on Twitter. It was also pretty well known that I had done a similar thing, and had warned people against sniping Ket, and was consulting with the Wink team.

Regardless, people still aped into a variety of fake LPs and lost money. I think this might be partially my fault; during the Ket launch I kind of just bluffed and didn’t bother making LPs with full liquidity because I was worried about screwing it up. Anyone who investigated that and thought it would be the same here took the bait hard when the full LP was added.

I believe that in the end the team withdrew more than twice LP that they actually added due to all the insane sniping that happened. RIP snipers.

For context, the LPs didn’t have liquidity for more than 14 seconds.

What’s a sniper, anyway?

An interesting question to be sure. There’s no fixed, agreed upon definition, as far as I know, so I’ll attempt to define one.

People definitely consider buying in the same block as launch a snipe. First few blocks after, probably easy to call those snipes, gotta be using a bot.

But is it really the bot that makes a sniper? I’m thinking about the $TRUMP and $LIBRE ‘launches’. In both those cases someone bought a million dollars in the first 10 seconds. I’ve heard those buys described as ‘snipes’. What makes them snipes? If you bought a million dollars of a fake coin, would you call it a snipe?

After thinking about it, to me the difference between a sniper and a non-sniper is not the technology used. It’s the information inequality. Doesn’t matter if you use a dex or a TG bot or you roll your own stuff. If you have some way of knowing something is ‘real’ before everyone else does, and you buy before them, that’s a snipe.

Why this is relevant

I saw a lot of dumb posts like this. Ignoring the fact that none of these people will share the example tx where they bought, even if we assume that they somehow watched Snowtrace (laggy as shit), copy pasted the deployment into Trader Joe (laggy as shit), and got a buy through in time, I don’t think that this is a good excuse.

Anyway, here’s a funny montage

All of these people are well known snipers (ok not that last guy, but he’s sniper-associated).

Suck a dick.

The legal implications of this launch

Well, I talked to my lawyer and he said it was fine. Seems like it would be hard to get prosecuted for it. Imagine if a store was legally obligated to buy back anything they sold to you? Or your counterparty in a stock trade? Not that these are securities, no sir.

Kind of strange how it’s so illegal but there are literally no examples of anyone getting prosecuted for it, right? I’m trembling thinking of the consequences.

Funnily enough you can get prosecuted for things like faking volume (and people have been). But given how many rug pulls you see every day, you’d think it should be easy to find news articles about it right? No dice.

Anyway, I asked ChatGPT Deep Research to take a look, here’s what it told me.

The ethical implications of this launch

There are a lot of different ethical frameworks you can apply to any topic. This one seems particularly grey.

Arguments against this

Violation of expectations

People see an LP created from a funding source associated with the project and assume it’s legit. Not fair to remove those funds.

Disclosure doesn’t make actions ok

Would it be OK to tell someone you’re going to punch them in the face, and then follow through?

Others

I’m open to listing more arguments here but I honestly have a hard time empathizing enough to think of them. If you want me to add something reach out and I’ll update the post.

My thoughts

Fuck those assholes. They wanted to buy a huge amount of supply cheaply and then dump it on everyone else. They got dumped on instead. Turnabout is fair play.

Not sure what the exact term for my position is, perhaps crypto anarchist? I think code should be law (I understand it’s not). I also understand that this is an extremist position though.

A more nuanced argument would focus on the fact that they were warned not to do this and then persisted. People need to be able to take precautions to safeguard their projects from malfeasance. If the precautions have no teeth, then there’s no disincentive to refrain from doing harmful things. They should take responsibility for their own actions.

The ethics of botting

I slipped my opinion on my own ethics into a previous post, did you notice it?

Anway that’s just a joke, I do have ethics, but I doubt they align much with most people. I already mentioned before that I’m a ‘code is law’ kind of guy.

I actually think the botters did nothing wrong. Surprise! I still don’t care what happened to them though. The project did nothing wrong either.

This shouldn’t be that unexpected, given how open I am about the fact that I bot things. Botting is an edge, and I like edges. I even told people during the Ket launch that they should use a bot. Level the playing field.

But I’ll be upfront and admit that the things I do are unfair and I am a bad person.

Building a bot for a (hypothetical) future Ket

I obviously had no time to run a bot for Ket. People don’t understand the kind of stress and amount of work that goes into pressing the buttons correctly at launch. If you’re doing things properly you don’t have time to bot your own launches.

I wouldn’t have botted it anyway though. It feels wrong to me, to bot a coin I launch. I went out of my way to make sure the launch was spotless.

But I couldn’t help but think in the days after launching that there were things I could do to bot launches like this in the future. Sniping a launch is easy if you have the CA or even the token name in advance, and the launcher doesn’t take precautions. But launches like the one I did, not so easy.

I’ve used Sigma before, and I’ve at least looked at some of the other TG bots. They’re not really flexible enough to do what I wanted to be able to do. So I decided to build my own.

Interactive sniping suite

I set out to build some tooling that would let you quickly evaluate new LPs, and buy them with a button press. I don’t know how to do TG bots, but I do know Discord bots, so that’s what I went with.

A very minimal tool to evaluate new LPs, and buy if you like it. There are a few extra features that didn’t trigger here, and I have a laundry list of potential improvements.

Works very smoothly. Not automatic obviously, since you want a human in the loop. But you can buy 1-2s after launch if you’re watching like a hawk. If you want to wait for more metrics to come back, takes more time.

I have some great ideas about how to incorporate AI that I’ll never get to!

Testing it out

So obviously I let my friends use my bot (for a fee, suckers). Unfortunately they kept buying random rugs, so I decided to leave it off, except for when we were expecting a launch.

At this point I had a great idea to incorporate a Twitter scraper into the bot to be used whenever a new LP is detected. The idea being that most projects (if they’re legit) will immediately post the CA on Twitter, giving you extra signal about if it’s OK to click buttons.

Unfortunately Elon is an asshole who made this extremely painful. I ended up registering a bunch of bot accounts and cycling through them to fetch tweets, and it was still very flaky.

Other bots

I’m a paranoid asshole, and the best way to prepare is defence in depth. So I actually wrote two other tools. The Discord bot is good for nontechnical people (my friends), but obviously I can run whatever I want myself, which gives me more options.

The first was an automatic Twitter scraper/buyer. The second was a more manual terminal based buyer that used some optimizations to send requests even faster than my Discord bot, with the idea that if the Discord bot failed that I’d still be able to buy something.

I had so many damn terminals and browser windows open during this launch it looked like a spaceship.

Wink launch

I mentioned before that I do have some ethics. Those ethics don’t extend to not trying to bot a launch of a friend. But they do extend to not having any special information to do so.

In this case, I knew that they were planning on doing fake LPs, and that you had to wait for the Twitter post. I knew what the contract looked like, but it didn’t have anything punitive towards people who waited for the CA to be posted.

I guess it’s slightly unfair that I knew that they were really serious about it and not just faking, but that seems like a pretty marginal advantage, given that you should probably have taken their warnings seriously anyway.

As for the people in my Discord? Well, I tried to warn them.

Did they listen? No.

I was assuming that the Twitter integration would give us a huge edge. Did the Twitter integration work? No. I ended up getting rate limited from scraping every time another fake LP launched.

So anyway, 15 minutes later the actual fucking LP finally launched, although rugs continued launching in the background.

The extra ping in there is another embed that I deleted to stop people from buying a rug.

I bought completely blindly based off someone’s random post that it was the right message (which in retrospect, seems crazy, but I didn’t gambol much). Lets see how they did in terms of latency.

Here’s the Discord embed post time.

Here’s the call time.

11 seconds, no idea how he managed it. I didn’t even see the fucking post update with my phone open to the page and a Twitter tab focused on my desktop.

Anyway, I blindly aped with my fastest bot based on that call and then bought the rest of what I had prepared based on other people confirming it.

At the end of the day, people using my bot had mistakenly dumped over 1.5K avax into the wrong LPs.

Rest in piss, brothers. Follow the fucking directions next time.