Thanks NEXUS for inviting me to TOKEN2049 Singapore.
Come try Pop — swipe YES or NO on what you think will happen next and see how your predictions resolve.
We’ll also have a POP airdrop for anyone trying it there!
I’ve mostly just been following my intuition, building as I go, and trying to look at everything from the user’s side — what would actually be better for them.
One thing led to another, and here we are.
https://t.co/1QlqArBHcC via @LinkedIn
@web3_sumon@SSRN Opened the research repo behind The Friction Ledger.
Inspect it, check the arithmetic, and tell me where it breaks😊
https://t.co/FZAAhha5aQ
Published my first paper on SSRN.
I analyzed 1.34M Polymarket BUY-side fills across 439 settled markets. In this sample, modeled fees were 10.4× the gross trading edge. @SSRN https://t.co/WI4bIjhL7C
Prediction-market fee curves do something strange: they make uncertainty expensive and near-certainty volume cheap.
I tested a one-parameter fix — add a floor, keep static revenue neutral, and make the result falsifiable.
https://t.co/TWhAGetUDM via @LinkedIn
1/6 A small gross edge. A fee bill 10.4× larger.
I reconciled $419.4M of Polymarket BUY-side flow: +0.143% before modelled fees, −1.345% after.
1.34M fills. 439 settled markets. Dec 2025–Aug 2026. All BUY fills treated as takers.
5/6 Limits are part of the result:
Top 500 active markets; newest 6,000 fills per market. Unsettled markets excluded. All-BUY-as-taker fee assumption.
Frozen aggregates let you check arithmetic. They do not reproduce the raw-tape analysis.