Frequently Asked Questions

What is BeatTheBots?

BeatTheBots is a prediction intelligence platform. We collect predictions from players, measure their skill over time, and use this data to train AI models. Players with proven predictive skill earn payments for their high-quality data.

How is this different from betting?

You're not wagering money against outcomes. Instead, you're providing prediction data that we use to train AI. Think of it like being paid for your expertise — the better your predictions, the more valuable your data, and the more you can earn.

Is this gambling?

No. BeatTheBots is a skill-based data platform. You purchase Data Passes to access prediction intelligence from other players. Payments are made for quality prediction data, not from a prize pool funded by other players' losses.

What do I predict?

You predict Net Goals — the difference between home and away scores. For example, if you think Liverpool will beat Chelsea 2-1, predict +1.0. If you expect a 0-0 draw, predict 0.0.

Why decimal predictions instead of exact scores?

We measure skill by average error over time, not lucky one-off guesses. Decimal predictions (0.1 increments) let you hedge between likely outcomes. For example, predicting −1.5 means you think −1 or −2 are most likely — this minimises your average error if either result occurs. Your B_Acc score rewards precision over time.

How is my score calculated?

Your match score combines four factors:

• • Accuracy: How close to the actual outcome
• • Timing: Earlier predictions score higher
• • Conviction: Deviating from consensus (when correct) is rewarded
• • Skill: Your proven track record (R-Score)

What is the Gatekeeper?

The Gatekeeper is a bot that sets the minimum bar for payment eligibility. To qualify for payment, your total series score must beat the Gatekeeper's score. This filters out lucky guessers and ensures only consistently skilled players earn payments.

Can I change my prediction?

Yes! You can update your prediction as many times as you like until the match ends. However, later predictions receive a lower Timing score — so there's an advantage to committing early if you're confident.

How do I get paid?

After each series settles, players who beat the Gatekeeper are eligible for payment. Your payment is proportional to your score share among eligible players. Higher scores = larger share of the payment pool.

Who is eligible for payment?

You must: (1) submit predictions for all 3 matches in a series, and (2) score higher than the Gatekeeper Bot. If you don't beat the Gatekeeper, you still keep your Data Pass access and your R-Score updates — you just don't earn payment for that series.

How much can I earn?

Payments depend on: (1) the total pool for that series (based on Data Pass sales), (2) how many players beat the Gatekeeper, and (3) your score relative to other eligible players. Consistently skilled players earn more over time.

Why does it cost to see other players' predictions?

Prediction data from skilled players has value — it's intelligence that can inform your own decisions. The Data Pass gives you access to the skill-weighted consensus, showing what proven forecasters are predicting. We charge for this data based on the same principle that we pay skilled players: quality prediction data is valuable.

How do my predictions train AI?

Your predictions, combined with your measured skill level, become training data for our AI models. The AI learns which patterns and signals correlate with accurate predictions. Over time, it develops insights from the collective intelligence of skilled human forecasters.

What makes prediction data valuable?

Raw predictions alone aren't very useful — what matters is who made the prediction. When we know a player's track record (their R-Score), we can weight their predictions appropriately. Predictions from consistently skilled players contain more signal, making them more valuable for training AI.

Why does player skill matter for data quality?

A prediction from someone with a proven track record contains implicit information — it reflects genuine insight rather than random noise. By measuring skill over time through our R-Score system, we can identify whose predictions carry the most signal and weight the training data accordingly.

What happens if a match is postponed?

Postponed matches are marked accordingly and removed from scoring. If a match in your series is postponed, only the remaining matches count toward your series score.

How does the R-Score work?

R-Score is a modified Glicko-2 rating that tracks your long-term predictive skill:

• • μ (mu): Your skill rating (starts at 1500)
• • σ (sigma): Uncertainty in your rating (decreases as you play more)
• • Win: Beating the consensus raises your μ
• • Loss: Losing to the consensus lowers your μ

New players start with high uncertainty, which means a skill penalty until you've proven yourself.

What is the consensus?

The consensus is the skill-weighted average of all predictions. Players with higher R-Scores have more influence on the consensus. Data Pass holders see this weighted consensus, which represents what the most skilled forecasters collectively predict.