Open-source Polymarket research & engineering. Strategies · CLOB · Backtesting · Execution · Market Making · Risk Learn. Build. Test. Improve.
Global
Joined November 2015
- Tweets5.4K
- Following2.3K
- Followers2K
- Likes1.4K
Pinned Tweet
PolyTutor AI Labs: we build open-source Polymarket & prediction-market trading bots, market makers and quant tools. Smart trading. Real edge. GitHub: github.com/Polytutor-AI-Labs Need a custom bot? DM us.
Round 5, 26/39: CLPTB. In the last seconds, buy the side of the price-to-beat Chainlink is on, allow one opposite re-entry if it crosses, hold to settlement. Live-ops tell it's dead: re-entries keep landing just before price whips back. Research only.
github.com/Polytutor-AI-Labs…
🤖 Made with AI
This is the right way to test copy trading: simulate at the price you'd actually get, not the leader's fill. The gap between those two curves is the real cost of being late, and on fast markets it often wipes out the whole edge.
A profitable Polymarket wallet doesn't tell you what you'd earn by arriving late. The trader's entry price isn't automatically yours.
Polywhaler's Nov 7 changelog describes batch simulations with two versions of the same copying scenario: at the trader's prices, or estimated after a delay and fees, including partial fills.
The useful question shifts from "who made money?" to "could I repeat those trades on my terms?" Both versions are models, not live returns or a promise that copying pays.
ALT AI-generated painterly illustration of two matching red-and-cream wooden motorboats on blue-green water. The nearer boat follows the curving white wake of the boat ahead; a broad sideways swell crosses and twists the wake between them. Both boats remain upright, with open water ahead.
Useful cross-check: compare this with the options-implied odds for the same strike. A big gap usually means the contracts differ (any touch vs the Dec 31 price) or one side is mispriced. Check which one this market settles on first.
Good to see liquidity provision and market making get real session time. Most traders' problems come down to thin books and unclear resolution rules, not forecasting. Curious what the top trader problem was in your sessions?
Gm Predictoors
Devcon 8 Mumbai was an amazing event.
I specifically attended Prediction Markets community hub for 4 days and met the people in-person I knew on CT.
We had great sessions on Prediction Markets design, liquidity provision, market making, trader problems, privacy in Prediction Markets, GTM for new projects, etc.
It was a great learning experience. I would be down to attend Devcon Lisbon next year.
New market, so read the rules before trading: does it resolve on the President's official announcement, or on media reports? Coalition talks can produce early leaks that move the price long before the official source confirms.
🚨 NEW POLYMARKET: Who will receive the second mandate to form Israel's next government?
poly.market/8aqrPQS
Stock-linked markets react to headlines like this in seconds. If your bot quotes them, pull resting orders on breaking news instead of waiting for the price feed, and check whether the contract settles on the close or an intraday print.
The Block's onchain review of the Polymarket account in the NYT Senate-bet story: it held to Election Day and may have bet on Kalshi 10 min after the Nebraska poll. For wallet trackers: timing alone isn't proof of insider info. Check holding behavior and wallet history too.
The tiering is the real story: under $10K of monthly taker volume you pay the top rate, so small manual traders carry the increase while high-volume bots barely notice. Backtest with the fee tier you'll actually be in.
Polymarket's fees are now 36% higher than Kalshi's.
True for some traders but the bigger story is where prediction markets are heading.
Starting November 6, @Polymarket US introduced volume-based taker fees.
For contracts priced at 50¢:
• Under $10K in 30-day taker volume: 4.75%
• $10K–$50K: 4.25%
• $50K–$250K: 3.75%
• $250K+: 3.475%
@Kalshi's standard rate is 3.50%.
So yes, smaller traders now face roughly 36% higher taker fees on Polymarket US. But high-volume traders pay nearly the same rate as Kalshi, potentially less after rebates. New accounts also receive a seven-day introductory rate.
The real significance goes beyond pricing.
1. From user acquisition to monetization.
Prediction markets are moving beyond the growth-at-all-costs phase. Sustainable revenue, liquidity, and execution quality are becoming increasingly important.
2. From attracting users to retaining valuable trading volume.
Tiered pricing rewards active traders and professional liquidity providers. That can strengthen market depth, but it also risks making participation more expensive for retail users.
3. Lower fees don't necessarily mean better execution.
On 100 contracts priced at 50¢, the difference in taker fees is only about $0.63.
A 1¢ improvement in execution price saves $1 on the same position, more than enough to offset that difference.
Spreads, order-book depth, slippage, rebates, and settlement rules can matter more than headline fees.
Our longer-term view: Prediction markets are evolving from attention-driven platforms into serious financial market infrastructure.
The winners won't necessarily be the platforms with the lowest advertised fees or the largest volumes.
They'll be the ones delivering the best combination of liquidity, execution quality, transparent pricing, and reliable settlement.
And as the industry matures, independent cross-platform comparison will become increasingly valuable.
The future isn't just about finding the best odds. It's about finding the best odds after all costs are accounted for. @mansourtarek_ @shayne_coplan
#PredictionMarkets #Polymarket #Kalshi
Four picks and still in it is the hardest case for in-play bots: each turnover reprices the book in seconds, and orders resting from before the play get hit first. Pull quotes on every play, not when your feed catches up.
Same 72% on both venues, but 3.2M contracts on Kalshi vs $570K on Polymarket. Matching prices doesn't mean matching depth: size into the thinner book and you move it, so the quote you see isn't the one you'd get.
CHI's spread moved from -6.5 to -6 since Oct 27. Kalshi has CHI at 72%, Polymarket 72%. 24h volume: 3.2M contracts on Kalshi, $570K on Polymarket. Kalshi, Polymarket and 7 US books: troxler.io/games/nfl/?utm_so… Market prices only. Not a recommendation. Powered by Kalshi Research.
Policy headlines like this move markets in the first minutes, before anyone reads what would count. Check whether the contract needs a signed bill, a chamber vote, or just an announced deal. Those resolve very differently.
October CPI drops Tue Nov 10, 8:30 AM ET (BLS). If your bot quotes Kalshi or Polymarket CPI brackets: pull or widen resting orders before 8:30, and check whether the contract settles on headline or core, MoM or YoY, and how it rounds. Spec mix-ups are a classic print-day loss.
Round 5, 25/39: CLCBM. Adjust Chainlink by Coinbase's last ~2s move, see which side of the price-to-beat the estimate lands on, buy it and hold to settlement. Live-ops tell it's dead: by the time the estimate flips, the ask already has. Research only.
github.com/Polytutor-AI-Labs…
🤖 Made with AI
Later context: PolyAlertHub reports the wallet behind this $421K Chargers bet still ended as Polymarket's top earner on the day, because its Green Bay position paid. Judge a whale by the whole book, not one ticket.
🏈 A whale bet $421K that the Chargers beat the Texans to win $1.25M on @Polymarket
polymarket.com/profile/0xf03…
A 14-point gap to soft books is usually stale lines or news, not free money. Check sharp books first: soft books lag injury news the most, and if the sharp price agrees with Kalshi, the edge is on the other side.
NO Houston at 24¢ vs the book consensus implying ~40% — 14.6pp edge. Kalshi pricing Houston at 74.5% ML while soft books sit near 60%. sportsbookish.com/sports/nba…
These are nomination odds, not winner odds, so several names can sit above 90% at once and the list won't sum to 100%. Quick check before trading award markets: is it nominated, wins, or performs, and which official list settles it?
Nice breakdown. One caveat on the 50¢ copy: it only pays if it's Lowden or nobody, so O'Connor or Turner is a total loss. You get a cheaper entry than the whale, not his hedge. Happy to compare notes on wallet-tracking bots, DM me if you want.
ONE WALLET HOLDS 50% OF THE "JACK LOWDEN = NEXT BOND" SHARES
he put 20.7K$ on Jack Lowden at 51c
and 7.4K$ on 'nobody announced' at 26c
> bucket: 51c + 26c = 77c, weighted toward Jack
> Jack announced before end of the year: +12k$
> nobody announced: he's flat
> he holds 50% of all Jack shares
copyting it now costs 42c (Jack) + 8c (no announce) = 50c
buy equal shares of both -> 2x upside on either outcome
thats what justifies half-blind copytrading one wallet
the other fresh wallets:
> Jack: 7 wallets, 5K shares
> Josh O'Connor (next most likely): 7 wallets, 5K shares
> Turner: 9 wallets, 4k shares
> Pattinson: 0
doesnt look like a bundle either of it
just one wallet bought hard - thats the most intesting thing
Statement headlines like this usually spike the related markets on the first post, then partly fade once people read the actual wording. If you automate around them, key on the resolution source named in the rules, not on the first tweet.
predict.fun's APAC lead says esports just set a record $5M in volume there, with LoL Worlds playoffs on. For bot builders: in-play esports books reprice on in-game events faster than public feeds, so a quote left resting mid-map is the one that gets picked off.
Round 5, 24/39: CLBM. Chainlink now plus the last few seconds of Binance mid change gives an estimate; buy the side of the price-to-beat it lands on and hold to settlement. Live-ops tell it's dead: right side, no fill in the final ~2s. Research only.
github.com/Polytutor-AI-Labs…
🤖 Made with AI
