pull down to refresh

@chrisliss is working on some things.

reply

yes, building a survivor tool, several different prediction markets betting tools.

reply

I’ve used Claude to analyze betting results. It produced some python script(s) to do it, and I iterate on it frequently

reply

optimized for anything specific?

reply

Yes! I am trying to learn and get into +EV betting, so I have been tracking the bets I place and then reviewing their Closing Line Value (CLV) to see how I am doing at picking lines.

The analysis is targeted at identifying my aggregate CLV, my rate of beating the close, etc. both in general, and broken down by facets like:

  1. sports league
  2. market type
  3. odds ranges
  4. sportsbook
  5. etc.

My goal is go have it help me learn where I am finding the most consistent edge to know where to deploy more capital, where to avoid, and so on.

reply

Vibecoding sports models via LLMs is really effective for data pipelining. You can have it write quick Python scrapers to pull historical box scores, feed them into Pandas/Polars, and backtest Poisson distributions for goal/points modeling in minutes. The key is keeping the data clean before letting the model calculate expected value.

reply

Not yet. I have a project in mind though.

I'll try to get something done by the time NBA tips off. The idea wouldn't work for NFL.

reply