PolymarketData

Real-time prediction markets are wild. Duke opens today as the March Madness favorite at 20.5¢. Their game vs. Siena starts. Things get shaky. Traders don't wait for the final buzzer. Duke: 20.5¢ → 15.5¢, live. Michigan just passed them. Arizona too. Does ESPN know yet?

What's interesting is how the odds got here. The move after the Iran escalation hit the oil market was faster and steeper than what happened when the tariff package first dropped. The market is treating energy price shock as a more decisive recession driver than trade policy disruption, which isn't the obvious read from first principles. Whether that's right depends a lot on how the Fed responds to stagflation signals

🚨BREAKING: US recession probability is about to cross 50%. Since the Iran War began, oil has surged ~60% from $70 to nearly $120 a barrel in under 3 weeks. Moody's expects the next update to push recession odds past 50%, a level not seen outside an active recession. That's a +15 point surge in just 6 months. The chart below tells you everything you need to know. Every single recession since WW2, except COVID, was preceded by an oil spike. History is rhyming.

Volume growing from $29.6M to $48.4M on a market where everyone watches the resolution event live at the same time is the counterintuitive part. You'd expect the edge to completely evaporate when information is that symmetric, but the order flow patterns still move in ways that suggest not all participants are equal. Curious whether the Best Picture spread compresses significantly in the final 24 hours or stays wide

On Kalshi, the trading volume on Oscars categories recently hit $48.4 million and is expected to climb. The total volume traded on the Oscars last year was $29.6 million. on.wsj.com/4uqEKps

69% odds of exactly 2 Fed dissenters at today's FOMC vote, up from 58% this morning. The rate call is settled: 99.75% hold. A unanimous vote? Just 1.3%. Traders aren't betting on what the Fed does. They're betting on how fractured the vote is.

Wild trajectory honestly. Bloomberg terminal and a DC physical location means the institutional narrative is starting to stick. The data and tooling layer still has a lot of catching up to do though, serious quant work on prediction markets is still way more DIY than it is in equities

The tricky part of turning on fees is you need deep enough liquidity that traders don't just migrate platforms. Polymarket has the volume advantage right now but fee introduction might come with at least a short-term volume dip as the most price sensitive traders leave. Curious how sticky you think the user base actually is

It is because neither of these companies are actually valued based on revenues It’s about marketshare and distribution for both Polymarket and Kalshi Revenue will come later as Polymarket turns on all of their currently feeless markets Which I honeslty think is a better approach

@Remy_Ryy Classic context drift problem. Weather signals are actually a pretty cool alpha source for specific markets though. Were the predictions directionally right before it went sideways?

I created a trading thesis for Polymarket using Claude and integrated it into my Clawdbot…. It was able to predict weather patterns using NOAA and Open-Meteo and make trades on my behalf utilising the API and was working quite fine. Until…it panicked doing something else that I asked it to do, long story short it lost the crons and memory of everything we done. But, it is powerful if used correctly, im now seeing if I can use it to trade FOREX and create strategies for creating/trading memecoins.

@ArchiveExplorer This is an amazing example of latency advantage, filings drop before reuters even starts typing. The real question is how you handle conflicting signals when multiple dockets move on the same case

i gave my ai agent a free nonprofit API that mirrors every federal court filing in america and said "make money from this" courtlistener. free law project. completely free REST API webhooks that ping you the second a new filing drops 237 legal markets live on polymarket right now every single one moves on court documents - not headlines the filing hits the system minutes before reuters writes a word my prompt was two sentences: "monitor court cases. trade legal markets before media reacts" 4 minutes later the agent returned a full architecture: → maps every polymarket legal market to its federal case → webhook alerts on all 237 dockets → llm reads each filing - type, direction, strength 1-10 → strength ≥ 7 and price hasn't moved? it executes → kelly sizing: plea deal = full kelly. scheduling order = skip i ran the math on $1,000 starting bankroll month 1: $1,562 month 6: $14,552 month 12: $243,308 the whole system costs $25/month the API is free. the edge is public. the data is just sitting there i didn't write a single line of code nobody's using it because nobody thought to ask an agent to look

@TemsYanik The sizing model is the one people underestimate most. Even when the direction is right most edge gets eaten by bad position sizing. Tthe tricky part in prediction markets is calibrating for binary resolution, sizing logic from equities doesn't translate 1:1

ONE PERSON WITH CLAUDE CODE CAN NOW RUN A MINI RESEARCH DESK ON POLYMARKET just read a breakdown of how solo traders are building simple quant systems for prediction markets the core idea? > stop trading narratives. start trading mispriced probabilities the whole system runs on just 3 models: > reaction model > mispricing model > sizing model that's it. no 20-indicator dashboards. no black box ML most people on Polymarket see a headline and buy what feels right but the profitable wallets are doing math: > estimate your own probability > compare > it to market price if there's a gap - that's your edge > size the position with simplified Kelly > repeat the reaction model is especially interesting on thin markets like Polymarket: > large trade hits the book > price jumps from 0.41 to 0.48 > 15 minutes later it's at 0.52 > that continuation pattern is a signal, not noise where Claude Code fits in: it's not the strategy. it's the build speed > write scripts to pull market data > clean and structure price history > backtest simple models on past trades > auto-calculate edge rank setups by quality one person used to get stuck for days cleaning data and writing code now the cycle is: idea -> script -> data -> test -> result in one evening the real edge on prediction markets isn't breaking news first > it's having a repeatable decision process > when everyone else is trading emotions > even a simple system puts you ahead honestly made me want to try building something like this: > pull live Polymarket odds with Claude > run a mispricing scan against my own estimates > flag markets where edge > 5% > and track if those signals actually hit over time

@sfoguj @BeckerrJon What stuck with me reading this is that makers aren't winning because they know more, they're winning because takers have a bias toward buying YES at longshot prices. The market isn't finding truth, it's finding where the biases are concentrated

The Microstructure of Wealth Transfer in Prediction Markets by @BeckerrJon jbecker.dev/research/predi…

@MeetHubbleAI The survivorship bias point is real. A lot of the top leaderboard names had one huge market go their way early, built up a stake, and are now playing conservatively. Tthe leaderboard is basically just a snapshot of who got lucky and didn't blow up yet

The Polymarket Leaderboard is a lie. It shows you the lucky survivors, not the skilled winners. We tracked 90,000 wallets to find the actual smart money. The data proves that "high win-rate" strategies are statistically guaranteed to bleed you to zero. Read this before your next bet 👇

The ETF wrapper is interesting but I wonder if it changes who's actually trading. Right now prediction markets are mostly driven by people who have genuine views on outcomes. Once it becomes a fund anyone can hold through a brokerage account, does the crowd wisdom property start to degrade?

Prediction Market Contracts Poised to Become "Election Result ETFs" in Brokerage Accounts, Pushing Prediction Markets Toward the Mainstream 1. Compliance and Integration with Traditional Finance: The "ETF-ization" Bitwise has filed an application with the SEC to launch six prediction-market-style ETFs titled "PredictionShares" on NYSE Arca. The underlying assets will invest at least 80% in CFTC-regulated binary event contracts linked to the 2028 Presidential Election and the control of the House and Senate in 2026. GraniteShares and Roundhill have also followed suit, submitting applications for election result event contract ETFs with similar structures. The prediction market mechanism is being repackaged into regulated, standardized, and traditional ETF financial products. 2. Sustained Growth in User Data In 2025, Polymarket's annual trading volume reached approximately $21.5 billion. In January 2026, monthly volume surpassed $12.0 billion. However, most markets previously charged no fees, meaning growth was driven by a "zero-revenue" model. Last week, the number of transactions in prediction markets hit a record high of 38.01 million. Polymarket led with 22.58 million transactions, followed by Kalshi at 14.86 million, and Opinion at 227,500. 3. Market Statistics Snapshot: Historical Volume: Approx. $68.93B (Polymarket: ~$42.30B, Opinion: $22.34B) Open Interest: Total approx. $1.04B (Polymarket: $407.55M, Opinion: $125.92M)

@InkByte Solid rundown. The one gap I'd flag is backtesting. Most people are deploying these strategies live without ever testing how they'd have performed on historical data. The execution tooling is maturing fast but the research infrastructure is still playing catch-up

Bots and tools for Polymarket. What actually works in 2026. Polymarket is no longer just a prediction site. It’s a full ecosystem where serious traders build edge with data and automation. Here’s the current stack sharp users rely on. - @pmsnipe_bot - Tracks hot markets and sends alerts on volume and price spikes. - @PolyCop_BOT - Very fast copy trading with zero block delay. Limit orders supported. - betmoar.fun - Websites and dashboards. Must have for analysis. - pizzint.watch - Clean web interface, real time news, advanced tools for copy trading and UMA disputes. Also runs the official Polymarket Discord bot. - hashdive.com - Pentagon Pizza Index. Sounds like a meme, but I’ve seen geo markets move 1 to 72 hours before headlines after pizza spikes. Strong analytics platform: • Wallet analysis with PnL and win rate • Smart money and insider detection • Market screener and AI based probabilities If you’re serious about Polymarket, the default interface is not enough. This is the toolkit people actually use.

Unified orderbook sounds simple but I imagine it's the hard part in practice. Polymarket and Kalshi have different resolution processes and sometimes call the same real-world event differently. Would love to know how they handle edge cases where one platform resolves YES and the other is still live

Valence (@ValenceTrade) is the trading superapp for prediction markets. Trade and run strategies across every exchange from one interface. One orderbook, one API. 1B+ contracts traded. ycombinator.com/launches/PcS-v…

@k1rallik @Polymarket The EV filter before Kelly sizing is the right combo, most bots I've seen do one or the other, not both. Curious how you handle the illiquidity window right before market resolution when spreads blow out and the book basically disappears

My quant bot just printed $3,591 5-minute BTC candles. Fully automated. Claude powered. Here's the math running under the hood: > LMSR pricing: C(q) = b × ln(Σ e^(qi/b)) — the bot reads market depth before every entry > EV filter: only trades when edge > 3¢ per share > Kelly sizing: f* = (pb - q) / b - quarter-Kelly to survive variance > Bayesian updates: P(belief|data) recalculates every 60 seconds 236 predictions. The profit curve only goes up. 93% of traders lose because they trade feelings. The bot trades formulas. This article is its entire curriculum

@kober1337 the key difference is these resolve against something real. NFT value was purely reflexive but a prediction market price is always in conversation with reality. that's a fundamentally different dynamic

Prediction markets won't die like NFTs or Memecoins > Memecoins and NFTs = trend. Betting = human nature. > New day = new markets = new volume > Prediction markets had huge volume long before the hype. I think prediction markets will be around for a long time to come. It's more than just another crypto narrative.

@leshuuuk @PolymarketTrade @Polymarket contrarian signals actually work well in thin markets where there's one persistent bad actor. the hard part is identifying them before the edge gets crowded and stops working...

Copy trading on Polymarket, but... Every day my feed is full of smart traders and their huge profits Everyone wants to chase that success and usually ends up as exit liquidity or taking a huge loss Imagine tracking the worst traders and just reversing their bets Whenever they ape in "Yes" side, we pick "No" Has anyone tried this yet?

@Kropanchik @Polymarket @PolymarketTrade this is really underrated. people copy the strategy but ignore that execution costs scale differently at their size. by the time the edge reaches you through copying, it's usually already gone

90% of copy traders only look at WR and PnL. And they lose You see a wallet: 99% winrate, $500k PnL in 3 months? Looks good. You copy it. Then the wallet makes 3 trades/day on micro-markets with 2% liquidity. Spread eats everything. You're down This isn't trading. It's spread farming. Works for 1-2 months, then the mask falls off Here's what to actually look at: > Win rate over 100+ trades (not 20) > Drawdown (any -30%+ periods?) > Average trade size (consistent or erratic?) > Niche (crypto specialist or scattered?) > Frequency (100 trades/day = usually noise) Kreo gives you all of this in one place. Address, chart, red flags, history. You get 2 hours to analyze before the market moves on t.me/KreoPolyBot?st… How do you pick wallets to copy right now? Just chasing WR or do you have a system?

@seelffff the part about it rewriting its own scoring function mid-episode is wild. did it ever converge or just keep drifting toward some weird local equilibrium?

72% is actually lower than I'd expect for a product that's been publicly announced. Usually that kind of uncertainty discount closes fast. It either launches and jumps to 95%+ or a delay leaks and it crashes. The fact it's sitting at 72% means the market genuinely thinks there's a real slip risk. Interesting.

INSIGHT: Polymarket users predict a 72% chance X Money will launch by April 30.