Football Predictions & Data-Driven Picks2026-04-30

Daily selections based on a four-algorithm system (Today's Research Selections), combining different modelling approaches to identify more reliable outcomes. Also Blog posts about the models and the experience of using AI coding agents in their building and updating.

Video

No video for today yet

How apps work

What each product does. Live apps link to the App Store; coming-soon items are described here only.

Live on the App Store

  • StatStrike

    A daily selection of Over 2.5 and Under 2.5 football forecasts. The statistical criteria used is listed with most forecasts and the algorithm's confidence in its work. App includes Track Record for performance transparency, filters to tighten focus on big fixture list days, aggregate market odds when available before KO, and a Best Performing category that only includes forecasts where the model has a minimum of 70% league accuracy historically.

  • GoalLab

    GoalLab forecasts the majority of published global fixtures daily. The algorithm uses an 11 criteria model to forecast Over 2.5 and Under 2.5 football goal bands. It will forecast with all 11 criteria, if available for the fixture, or whatever it can get - forecast confidence is reflected in the volume of criteria available for any given fixture. This doesn't mean a lower confidence forecast is necessarily less accurate than one with more criteria - it depends on the criteria mix and how they interact. Historical win rates of every confidence level is listed as tracked by a rich and growing archive.

Coming soon

  • PopGoals

    Three golden balls every day with the top slice of Over and Under 2.5 selections. Not on the App Store yet.

  • ProphIt Coming Soon!

    A new service! Have a theory for predicting goal band outcomes?

    This service lets you test your approach using real data, live execution, and transparent tracking — so you can see how it actually performs.

Coming Soon!

PopGoals app icon

PopGoals

Coming soon

iOS app in development. App Store listing and preview copy will follow.

ProphIt — Test Your Own Prediction Ideas in a living app!

Coming Soon!

Have a theory for predicting goal band outcomes?

This service lets you test your approach using real data, live execution, and transparent tracking — so you can see how it actually performs.

How the research service works

  1. You define your ideaDescribe your logic — from simple rules to more detailed concepts.
  2. I build your modelYour idea is translated into a working forecasting algorithm.
  3. We run it liveYour model is executed against real matches over a fixed research period.
  4. You track the resultsYou get access to a dedicated app/dashboard showing:
    • Predictions
    • Results (W/L)
    • Performance over time

What you get

  • A working version of your idea as a live model
  • A private dashboard to track performance
  • Real-world validation (not just backtested theory)
  • Clear insight into whether your idea has an edge

After the research period

When the initial research period ends, you can:

  • Extend testing for an additional fee, or
  • Have your algorithm deployed in a dedicated app for your personal use, for a one-off fixed cost

For ongoing use, you will need a low-cost API subscription for match data. You can connect your own API key, and the system will run your model automatically.

If preferred, managed data access can be provided for a small monthly fee.

Important

  • This is a research and testing service, not financial advice
  • No outcomes or profitability are guaranteed
  • Most ideas do not perform well — that is the purpose of testing
  • Your model is treated as confidential
  • Similar outcomes to existing models may occur independently

Pricing

Flat fee depending on complexity and duration.

No ongoing commitment required.

Submit your ideaClick to expandTap to expand

Please provide a clear outline of your approach. The more specific you are, the faster I can assess whether it's suitable for testing.

Your details

Your idea

(What factors determine your prediction? Be as specific as possible.)

(e.g. form, odds, player stats, historical results, etc.)

(e.g. match winner, over/under, both teams to score, etc.)

4. How often should predictions be generated?

Practical details

6. Do you already have an API for match data?
7. Preferred research duration

Expectations

9. Have you tested this idea before?

Final step

Optional

I will review your submission and confirm whether your idea is suitable for implementation, along with next steps and pricing.

Prefer email? jmclarenscripts@gmail.com