Polymarket Tennis Live Score Edge

This strategy uses the livetennisapi/polymarket-tennis toolkit as a data layer to discover tennis markets on Polymarket's Gamma API and match them to live ATP/WTA/Challenger matches. The core edge is that Polymarket tennis prices often lag real-time score events such as a break point, set point, or service break. By ingesting live score feeds alongside market prices, the bot can detect when implied probability diverges from the true win probability given the current score state.

Entry logic typically fires when the market-implied probability for a player is more than N cents below the live-score-implied probability (computed from a Markov or historical base-rate model conditioned on score, server, and set). Exit happens when the price converges toward fair value or the match state changes materially (e.g., opponent breaks back). Position sizing should be fractional-Kelly because tennis score states are highly volatile.

The repo is observe-only and MIT-licensed, with 317 stars and 276 forks — a high fork-to-star ratio suggests active derivative development, though the star count also carries a POSSIBLE_STAR_FARMING flag. The toolkit handles Gamma API discovery, match matching, and live score streaming, so it is best used as an infrastructure component rather than a full trading bot.

Key risks: live score feed latency, API rate limits during Grand Slams, and the fact that Polymarket liquidity on individual tennis markets is often thin, causing large slippage on entries. Also, sharp books may price in break points faster than the data feed arrives, so speed of execution matters.

Implementation notes: pair with a Polymarket CLOB client (py-clob-client) for order placement, cache match IDs to avoid re-matching every tick, and backtest against historical score timelines before going live.