Casatrick
@casatrick.bsky.social
Senior blockchain backend engineer (7+ yrs), event-driven infra for trading platforms & prediction markets. Shipped a prediction-market protocol w/ Pyth oracles, on-chain AI agents, and a low-latency trading terminal.
A Polymarket trading bot is more than strategy + API. - Market data - Order book - Strategy - Risk - Execution - Positions - Reconciliation - Monitoring - Recovery Momentum and market making can share the same infrastructure. The strategy should be replaceable. The infrastructure shouldn't.
1 week of running my bot against Polymarket's new TWAP settlement Quick honest number: [65, 999 trades, 55.6% win rate, +$3700.64]. The sniping exploit the update closed was worth $8.2M to ~821 wallets under the old system - that edge is just gone now. casatrick.substack.com/p/polymarket...
Also affects your strike price, not just settlement - both read from the same TWAP feed. And if you're reconciling your own TWAP, your drift tolerance is now miscalibrated too, since a longer window has a higher natural noise floor. Full breakdown: x.com/casatrick/st...
Casatrick (@casatrick) on X
Polymarket 5-Min TWAP Doubles to 60s: The Feed Swap Most Bots Will Miss
x.com
5-min Polymarket TWAP moves from 30s to 60s tomorrow (Aug 14). Most people will treat it as a config edit. It's not - Chainlink ships these as separate feed IDs, not one feed with an adjustable window. If you hardcoded a window variable, you're still pulling the old feed.
14 of the top 20 wallets on Polymarket's leaderboard are fully automated trading bots. Not assisted-fully bot-run. The edge isn't better prediction. Research estimates ~$40M was extracted from Polymarket in a year through pure execution-speed arbitrage.
Most Polymarket trading bot writeups stop at the pricing model. The part that actually decides profitability on thin-liquidity markets is execution timing - the gap between calculating an edge and the order actually landing. A stale book means you fill at a price your model never saw.
Been updating my Polymarket trading bot for the new TWAP settlement rollout. Most of the work isn't just changing one formula-it's revisiting assumptions around execution, settlement windows, and late-entry logic. dev.to/casatrick/ho... #Polymarket #TWAP #TradingBot #RustLang #AlgoTrading
How to Update a Polymarket Trading Bot for TWAP Resolution (Live August 7)
Polymarket is switching its crypto up/down markets from single-price-snapshot resolution to...
dev.to
Wrote the follow-up to my Polymarket TWAP piece - this one's the actual Python script. Chainlink doesn't publish the exact TWAP sampling/weighting for the feed, so you can't independently derive it. But you can still catch drift and feed gaps before they matter for a live position.
Polymarket is closing a real exploit in its crypto markets. From Aug 7, resolution moves from a single price snapshot to a time-weighted average (TWAP) - closing a loophole tied to ~$7.6M in reported manipulation losses. Wrote up the mechanism + what it breaks for bots: x.com/casatrick/st...
Adapted a multi-timeframe trend-alignment idea as a directional signal for Polymarket's BTC 5m/15m/1h markets. Most of the original strategy - DCA layers, trailing stops - doesn't survive the trip to an expiring binary market. casatrick.substack.com/p/a-trend-al...
A Trend-Alignment Signal for Polymarket's BTC Markets
Borrowing a multi-timeframe momentum idea from futures trading, minus the parts that don't survive contact with an expiring binary market.
casatrick.substack.com
A polymarket price isn't a price, it's a probability the crowd agreed on a moment ago. The second real news hits, that number is stale until everyone finishes the same math a bot can run in milliseconds. The formula: posterior odds = prior odds × likelihood ratio.
Polymarket Trading Bot Strategies: Bayesian Updating for Real Edge paragraph.com/@casatrick/p...
Polymarket Trading Bot Strategies: Bayesian Updating for Real Edge
A prediction-market price is not a price. It's a probability the crowd agreed on a moment ago. The second real news arrives, that number is wrong - not forever, just for the few minutes it takes every...
paragraph.com
Pulled a Polymarket wallet's full trade history (~65k trades) and tried to reverse-engineer their strategy. One global rule didn't hold - but segmenting by price and time-to-close revealed two distinct, consistent modes. Writeup: dev.to/casatrick/i-...
Polymarket arbitrage bots don't fail at finding the gap - every bot sees the same mispricing. They fail at capturing it before it closes (2-3 sec windows). Wrote up why execution speed matters more than detection logic here: casatrick.substack.com/p/polymarket-arbitrage-bot-execution-speed
The Real Bottleneck in Polymarket Arbitrage Isn't Finding the Edge - It's Keeping It
Polymarket arbitrage bots don't fail at finding mispricing - they fail at capturing it before the window closes. Here's why.
casatrick.substack.com
Most trading bot writeups cover the edge calculation and stop there. Position sizing is the part that actually determines whether a real edge turns into real returns. Kelly sizing looks great in theory. In practice, it amplifies your model's own error. #tradingbots #polymarket #fintech
latency matters. A book can shift between last poll and order landing when top-of-book depth is shallow. calculate real edge, then fill against a price that no longer reflects it. pre-submit re-check - diff the live book against your pricing snapshot, abort if drift exceeds tolerance.
Most Polymarket bots place a market order and call it done. Fine on liquid markets - costly on thin, long-tail ones (5-10 cents of slippage per fill). Wrote up the real risk logic: depth checks, edge vs. implied probability, fractional-Kelly sizing. Full write up: github.com/casatrick/po...
Building a Polymarket Trading Bot: Order Book Monitoring, Slippage Handling, and Position Sizing paragraph.com/@casatrick/b...
Building a Polymarket Trading Bot: Order Book Monitoring, Slippage Handling, and Position Sizing
Most first-attempt Polymarket bots place a market order and call it done. That works fine on liquid, high-volume markets, but Polymarket has thousands of long-tail markets with thin order books and wi...
paragraph.com
Why Polymarket Bots Need Different Risk Logic Than Casino RNG Games paragraph.com/@casatrick/w...
Why Polymarket Bots Need Different Risk Logic Than Casino RNG Games
If you're building a Polymarket trading bot, don't copy the risk logic from a casino or RNG betting bot. It's a category error that will bite you the first time a market gets volatile. Casino odds are...
paragraph.com
What I Learned Running a Trading Bot on BTC 5-Minute Markets This Week casatrick-newsletter.beehiiv.com/p/what-i-lea...
casatrick-newsletter.beehiiv.com
My new Polymarket Trading Strategy paragraph.com/@casatrick/m...
My new Polymarket Trading Strategy
So far I just am poking around btc 5m, 15m markets so completed my own strategy for btc market. I noticed that there's a trader on Polymarket quietly printing money on ETH 5-minute candles. Not throug...
paragraph.com
started at $500 for 2 day it reached $5090 can't reveal all figures but here is some still in test process of strategy(just named it Binary Parity Arb)
I built a Polymarket arbitrage bot - here's what the live data actually shows casatrick-newsletter.beehiiv.com/p/i-built-a-...
I built a Polymarket arbitrage bot - here's what the live data actually shows
Five strategies, three live accounts, and one bot losing money in real time right now
casatrick-newsletter.beehiiv.com
Still here, still building. paragraph.com/@casatrick/s...
Still here, still building.
I have been building a market maker bot on Polymarket for the past few weeks and honestly it has been one of the most interesting developer projects I have worked on. Most people think trading bots ar...
paragraph.com
Built a trading bot for Polymarket in Rust. Decision loop went from 48ms (python prototype) to ~800µs after the rewrite. Few months in, mostly breaking things and fixing what I broke. github.com/casatrick/po...
GitHub - casatrick/polymarket-arbitrage-bot: Polymarket Arbitrage Bot: Automated trading bot for risk-free arbitrage, YES/NO spreads and crypto 5m/15m Up/Down markets. Real-time detection + execution.
Polymarket Arbitrage Bot: Automated trading bot for risk-free arbitrage, YES/NO spreads and crypto 5m/15m Up/Down markets. Real-time detection + execution. - casatrick/polymarket-arbitrage-bot
github.com
