Real-time odds, breaking headlines, and data-driven insights that cut through the noise. Newsletter coming soon. Livestreams every Tuesday, Thursday & Sunday.
New Accuracy Research on Polymarket
TLDR:
Polymarket:
- has a Brier score (stats method of measuring forecasting accuracy) of 0.0565, on par with the best predictoors in existence.
- is getting more accurate over time as more liquidity comes online
- volume boosts from one-off events like elections spread out into other markets after the big event ends
- as expected, there is a linear relationship between volume and accuracy: more volume = more accurate
I just posted a new @Dune board with new insights into @Polymarket's accuracy.
This board goes more in-depth, analyzing:
🎯 The accuracy of PM's odds
⏳ How overall accuracy is trending over time
💰 How accuracy is impacted by total liquidity
(Link in next tweet)
1/11
Day 303: One of the easiest ways to print on Polymarket is by sniping new markets with wide spreads.
Most people ape into the obvious side, but the real edge is setting stink bids while liquidity is still thin.
For example right here I got filled for 1000 shares at 51c (510$). What I usually do is try to sell these instantly for 72c which would lead to a profit of 210$.
The trick is to monitor fresh, low-liquidity markets, drop stink bids with an expiration date, and let impatient traders hand you free money.