Tom Liptay

We built a general-purpose AI forecaster. It uses deep web research and proven forecasting best practices to produce transparent, well-sourced probability estimates. We hope it is valuable to investors, policy makers, and prediction market traders alike.

Forecasts: #gid=2132174818" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Trades/portfolio: #gid=1769292422" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Forecaster: colab.research.google.com/drive/1Rh9iuGD… Trader: #scrollTo=pWsOw-2SSece" target="_blank" rel="nofollow noopener">colab.research.google.com/drive/1KWOSz4B…

I updated our play money AI Kalshi trader. Summary: Feb 26: $100,000 Mar 13 after buying: $101,486 Mar 19 after buying:$97,901 I made a major upgrade to the trader. TLDR: Previously it tried to buy equal positions. Now it buys the highest returns. 4% max size.

@ClarkSquareCap Most recently forecasting Kalshi markets. Currently compounding at 50%+ after 2 weeks, ~70 positions. We'll see if it is sustainable. Before that Stockfisher. Made a 10+ year DCF for every company in the S&P 500. Sharpe ratio ~ 1.5 on diversified portfolio.

What are some of the best ways you (or somebody you've seen) are using agents/LLMs in their investment process? [I want to make this a weekly/bi-weekly running thread]. Appreciate a RT for visibility + more responses. Thx!

@SleepwellCap Gravel bike.

what’s a material thing between $1000-$5000 that you’ve bought that actually changed your life? mine is the Eames lounge chair, no questions.

Screenshots:

This is not investment advice. Forecasts: #gid=197136818" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Portfolio: #gid=70820025" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Forecaster: colab.research.google.com/drive/1r0Pf1It… Trader: colab.research.google.com/drive/12XUH6iF…

I updated our play money AI Kalshi trader. Summary: Feb 26: $100,000 Mar 5 after selling: $101, 044 Mar 13 after selling: $102,122 Mar 13 after buying: $101,486 For updates we: 1) Forecast. 2) Handle market resolutions. 3) Sell over-priced positions. 4) Set target buy positions. 5) Buy using the order book. This past week saw: - 7 resolutions (we had 3 winners, 4 losers) - 4 sales - 19 buys - Ending cash $24k

@MMikeMMa There are a bunch of assumptions required that I'm glossing over with regards to the trading algo. I just tried to make it simple. The main point is to test if our forecasts are more accurate that the prediction market in a rigorous way.

@MMikeMMa Sorry, it is a bit more involved. I should have written more. We forecast the probabilities on all positions. Then, we sell if it is negative EV. And, then we buy the markets that are highest return - if our forecasts are correct.

I updated our play money AI Kalshi trader. Summary: Feb 26: $100,000 Mar 5 after selling: $101, 044 Mar 5 after buying: $99,446 We first sell positions. Then when we buy we pay a higher price than the mid-price, since we use the order book.

This is not investment advice! Forecasts: #gid=326143494" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Portfolio & trades: #gid=1436244269" target="_blank" rel="nofollow noopener">docs.google.com/spreadsheets/d… Colab forecaster: colab.research.google.com/drive/1M5k5AbR… Colab trader notebook: #scrollTo=_qJWMdFupUyN" target="_blank" rel="nofollow noopener">colab.research.google.com/drive/1QpieIXH… Posting screenshots of the trades.

Partly I'm posting this on twitter with screenshots to serve as a public record.

The full case study includes links to notebooks that you can run yourself. everyrow.io/blog/kalshi-tr…

We built a general-purpose AI forecaster. It uses deep web research and proven forecasting best practices to produce transparent, well-sourced probability estimates. We hope it is valuable to investors, policy makers, and prediction market traders alike.