10 years in crypto. 300+ domains. $20M+ wagered in crypto casinos. A $250K month and an $18K night on Polymarket. Now building trading bots with AI — without knowing how to code.
From today this account is a ledger. What I did with money, what happened, receipts attached. Wins and losses, same font.
@tematkachenko Всей мощи говноармии за 5 лет не хватило даже на это 🤷♂️
50+ баллистик по одному из мостов и куча по всем остальным - всё цело на древних 50-100 летних пролётах, это же ‘8е чудо света’ сложилось как хворост..
Mistake #3: −$91,050. LoL Worlds Play-In, Oct 17, FURIA vs Cloud9, best of five.
Late at night I bought 110,000 Cloud9 shares at 82–83¢ with limit orders. "I'll place the hedge in a minute."
I fell asleep. No second leg.
Cloud9 lost and went home first.
The stupid part: at 83¢ plenty of bettors wouldn't hedge at all. That would have been a decision. Mine was a nap.
My Polymarket profile says −$623,915. All-time, $22.9M volume, 1,697 markets.
I didn't lose all of it. About two thirds came back on the sportsbook side of the same bets. About a third is real.
Three mistakes alone cost $217K. Each one gets its own post 🧾
Mistake #2: −$76,887. LCK, Aug 23, Hanwha Life vs T1. Both sides at 1.90.
I had T1 on Polymarket, $54.7K. The hedge: $60K on Hanwha at Stake.
Same odds on both buttons. I put $20K on Hanwha. And $40K on T1, the team I already had.
Hanwha won 2–0.
My Polymarket profile says −$623,915. All-time, $22.9M volume, 1,697 markets.
I didn't lose all of it. About two thirds came back on the sportsbook side of the same bets. About a third is real.
Three mistakes alone cost $217K. Each one gets its own post 🧾
@OutcomeAhmed@Outcomexyz Odds only decent because it’s dead with no liquidity
Once (or if ever) people start using it - markets will become much more efficient and equal out
Half my feed: "this bot just buys cheap and holds to the end" or "it buys both sides for under $1, that's the whole trick".
If only. This is one 15-minute BTC window from two traders who actually make money on it. 408 trades between them. One-second candles.
I've gone through dozens of these windows frame by frame. I still can't fully write down their model.
If someone explains it in one tweet, they're farming your replies.
Mistake #1: −$49,080. World Cup, July 1, England vs DR Congo, who advances.
Before kickoff I had one side: $181K on England at Stake, around 1.12. The plan was to add the Congo side cheaper during the game.
England went a goal down. "Cheaper" became 35¢. I bought $78K of Congo on Polymarket to close the hole.
Right after my last order, England equalised. Then won 2–1.
Without the hedge: +$28.9K. With it: −$49,080.
My Polymarket profile says −$623,915. All-time, $22.9M volume, 1,697 markets.
I didn't lose all of it. About two thirds came back on the sportsbook side of the same bets. About a third is real.
Three mistakes alone cost $217K. Each one gets its own post 🧾
My Polymarket profile says −$623,915. All-time, $22.9M volume, 1,697 markets.
I didn't lose all of it. About two thirds came back on the sportsbook side of the same bets. About a third is real.
Three mistakes alone cost $217K. Each one gets its own post 🧾
@coleodds Fair. At this size it's noise both ways, and I can't call any of the 18 an edge yet.
The real damage on my Polymarket account wasn't the bots anyway. It was me. Posting that one next.
Since Oct 9 my AI-built bots have made 2,162 live trades on Polymarket. 18 strategies, 3 servers, $17.9K deployed.
Net: −$442.
Best strategy +$106. Worst −$177. Eleven of eighteen in the red.
Everyone shows you the green month. This is what "building a trading bot with AI" actually looks like 🧾
@GurrLabs No fixed bar yet. Paper always looks better than live: it doesn't see the delays, and speed is where things break. If paper looks fine, it goes live with $5-10 clips. Almost every time something new shows up, so I switch it off with a small loss and go back to studying.
@deltatraderXBT That I can’t figure out the majority of the strategies people use from A to Z 🤪 at a first glance - it’s all clear and simple, but when you dig a little bit… 😱
New security feature: Withdrawal address allowlisting
You can now lock withdrawals from your deposit wallet to addresses you’ve approved. Adding a new address requires email verification and a 48-hour time lock, and you can cancel it from the email if it wasn’t you. That means even if someone gets into your session or you sign a bad transaction, they can’t drain funds or transfer your positions to an address you don’t control. I highly recommend enabling it if you are on the latest wallet version.
Turn it on in Settings → Account. Passkey support is coming soon to strengthen security further.
For non-deposit wallet users, we are developing a migration tool to help you move everything to the new wallet type. If you don't want to wait, you can open a new account to access this feature immediately.
We are also working on adding more advanced built-in security protection, without requiring you to enable anything on your deposit wallet - will also be live soon.
My feed is full of "look at this Polymarket wallet, just copy it and get rich". Usually with a referral link under it.
I tried. 18 live strategies built with AI, 2,162 real trades, hundreds of hours. Some simply copied top wallets.
Result below: −$442. It doesn't work the way those posts tell you.
Next I'll take the strategies apart one by one. What the idea was and where it broke: latency, no infrastructure, no API key, bad fills. And a few honest "no idea yet, it just doesn't work" 😄
Don't pay for what I already paid for.
Since Oct 9 my AI-built bots have made 2,162 live trades on Polymarket. 18 strategies, 3 servers, $17.9K deployed.
Net: −$442.
Best strategy +$106. Worst −$177. Eleven of eighteen in the red.
Everyone shows you the green month. This is what "building a trading bot with AI" actually looks like 🧾