# realtime.tips

> Horse racing analysis for UK and Irish racing: free machine-learning tips, member-only large language model (LLM) selections, public model results, and a free System Builder for testing selection rules on historical results.

Free machine-learning tips and member LLM selections are separate products. Today's LLM selections require membership; public results and methodology can be read without a subscription.

Performance figures belong to the model, date range, selection set and price basis stated on each source page. Win strike rate is not profitability. All-picks results, filtered subsets and historical backtests describe different records. Figures labelled gross of exchange commission exclude that cost. Past results do not guarantee future returns.

## Products

- [Home](https://www.realtime.tips/): Overview of the service and links to daily racing analysis.
- [Free daily tips](https://www.realtime.tips/tips): Machine-learning selections for UK and Irish racing.
- [LLM tips](https://www.realtime.tips/advanced): Overview of member selections from OpenAI, Anthropic and xAI models, with links to public results.
- [System Builder](https://www.realtime.tips/system-builder): Free tool for building and backtesting horse racing selection rules.

## Results and evidence

- [Public model record](https://www.realtime.tips/advanced/proof): Settled model results, including wins and losses, with the reporting period and price basis.
- [How accurate are AI horse racing picks?](https://www.realtime.tips/blog/how-accurate-are-ai-horse-racing-picks/): Published samples, dated results and the distinction between accuracy and profit.
- [Which AI picks the most horse racing winners?](https://www.realtime.tips/blog/which-ai-picks-the-most-horse-racing-winners/): Model comparisons with sample sizes and performance context.

## Methodology and policies

- [Results policy](https://www.realtime.tips/blog/results-policy/): How results, price bases, complete records, filtered subsets and backtests are presented.
- [Editorial methodology](https://www.realtime.tips/blog/editorial-methodology/): Evidence standards, sourcing and corrections for published articles.