Sport Forecasters

  • Data

  • How We Measure
    • Documentation

  • Roster

  • FAQ

  • About

    • Terms Of Use

    • Privacy Policy

  • Contact Us

Sport Forecasters


  • Data

  • How We Measure
    • Documentation

  • Roster

  • FAQ

  • About

    • Terms Of Use

    • Privacy Policy

  • Contact Us

Frequently Asked Questions

  1. How Long Have We Been Keeping Track Of The Data?

    • NFL and College Football predictions across all pertinent shows have been tracked back to 8/8/2025. NBA predictions have been tracked back to 10/10/2025.

  2. What Qualifies As A Betting Prediction?

    • Whenever an analyst predicts a game result or outcome in relation to a money line, point spread, or a teaser, it qualifies as a betting prediction. Essentially, this category seeks to measure how accurate analysts and media personalities are when they advise their audiences to make a bet. In some cases, an analyst can state that Team A is the underdog by 3 points, but he believes in the team so he will pick the team to cover and win the game. In this scenario, the analyst is attributed two predictions:

      • A betting prediction where Team A has to win or lose by 3 points or less
      • A regular, game prediction where Team A wins.

    • If Team A ends up losing the game but by only 2 points, they get credit for the betting prediction but are decisively marked incorrect for the regular game prediction. If someone mentions parlays, each part of the parlay gets their own data point/credit as we want to judge analysts. If you want more clarity on what defines a prediction overall, please click here.

  3. How Does Quality Differentiate From Accuracy?

    • Accuracy is the number of correct predictions/number of total predictions. Quality is defined as riskiness ( analyst score/number of predictions) multiplied by accuracy. As a reminder, the analyst score is the number of days separating the day a prediction is made from the day that prediction comes true. Essentially, we use quality to recognize analysts who take more risks (making predictions way out in advance) as it is easier to make successful predictions closer to gameday. For example, if an analyst makes a successful game prediction a day before the game, that should not get the same value as an analyst predicting a Super Bowl winner at the beginning of the season. Quality seeks to reward risk taking as well as offer weight to predictions that seem less likely to occur at the time it was made.

  4. Why does a prediction have to be affirmative?

    • All predictions that are tracked have to be positive statements of what a team will achieve rather than what they wont. Otherwise, it would be easy for analysts to take advantage of the system. For example, predicting a team to not make the Super Bowl at the beginning of the year has a good chance of occurring as opposed to predicting a team to make the Super Bowl. Giving equal credit to both predictions would distort the results, so we are limiting predictions to those that are accomplishments. The one exception we make is if a negative prediction takes place at the beginning of the season and has more equitable odds of occuring. For example, stating team A will not make the playoffs (rather than the Super Bowl) before the season begins is a riskier take and therefore those predictions are counted.

Sport Forecasters
  • @spoforecaster
  • Instagram