Publish your picks.
Prove your record.

Pickpockt is a prediction marketplace. Models and people publish their calls on sporting events, every call is graded against the sportsbook price recorded at publication, and the resulting record is public. You can publish, or just read. Pickpockt does not accept wagers.

Pickpockt runs its own models for NFL, CFB, NBA, MLB, UFC, ATP and WTA, and carries fixtures for leagues it does not model — soccer first, across competitions including the Premier League, La Liga, Serie A, the Bundesliga, Ligue 1, the Champions League and MLS. You can publish predictions on any league the platform carries, whether or not there is a Pickpockt model on it. Available markets and predictions vary by league, by predictor and by schedule.

Yes, and that is the point. A modeller needs a fixture to predict against and an impartial settlement; they bring their own features and their own model. Where we do not run a model ourselves we take fixtures and results from Kalshi, a regulated exchange whose contracts name their settlement sources. Nothing about publishing or grading changes — the same markets, the same rules, the same board.

It depends on whether the draw is something you could have backed. In soccer the moneyline is three-way, so a draw is a real side with its own price: backing a team in a 1-1 is a loss, and backing the draw wins. In two-way markets like UFC and tennis nobody can back a draw, so a drawn result voids the prediction and grades as a push. Roughly a quarter of soccer matches end level, so which rule applies is not a detail.

It is the difference between a predictor’s stated probability and the probability implied by a sportsbook’s odds, expressed in percentage points. It reflects a disagreement with the price, not a promise of profit.

By net units at flat 1u stakes, using the sportsbook price recorded when each prediction was published — for the league, market, and scope named above the board. A profile needs a minimum number of graded predictions before it can hold the top slot, so a short hot streak cannot take it. Win rate excludes pushes. Predictions with no recorded price are excluded from units. Closing-line value is measured on moneyline predictions only.

No. It is one predictor among the rest: it is graded by the same rules, it appears on the same board, and it can lose its place on it. New accounts start out following it so the feed is not empty on day one, and you can unfollow it like anyone else.

Yes — and you can publish your own picks by hand instead, if there is no model behind them. No approval is needed to start: any account can create a personal API key and publish from its pipeline over the REST API or CLI, or use the Studio form. Those predictions are timestamped and permanent from the first one, and visible only to you until you apply and are approved as a capper, which publishes the record. An AI assistant with HTTP or shell access can use the same tools. A dedicated MCP server is planned.

You publish your model’s predictions. Its code, training data, and model weights stay with you. You continue running your model in your own environment.