How a live state becomes a signal.
How that signal becomes a public result.

1. What does eSoccerSignal monitor?

The scanner monitors supported short-format live eSoccer events while those events are present in connected sources. A live event being visible does not mean a signal will be published. Most candidate states are filtered out.

2. What is a qualifying signal?

A signal is a live event state that passes the production qualification process in force at that moment. Qualification can consider event context, score and clock state, market availability, price information, historical evidence and production safety guards. Exact thresholds and proprietary rules remain server-controlled and are intentionally not published.

Changing a member's signal profile can affect the number and selectivity of future signals, but it does not rewrite historical records.

3. How are duplicate observations handled?

Multiple platform observations of the same underlying event state are treated as observations of one canonical signal when they resolve to the same signal identity. This prevents public signal totals from being inflated simply because the same state was visible through multiple connected sources.

4. What do “AI probability”, “Logic probability” and “Market probability” mean?

AI probability is the independent estimate from eSoccerSignal's eSoccer-trained machine-learning model when the AI has the exact same candidate context as Production Logic. Logic probability is the production-reference probability attached to that qualified candidate. Market probability is calculated from captured market pricing when that information is available.

The current AI model uses gradient-boosted decision trees (CatBoost) trained and versioned by eSoccerSignal on structured live-market observations. CatBoost is the underlying ML framework; the trained model, feature pipeline and eSoccer dataset are specific to eSoccerSignal.

AI probability uses a validated live calibration overlay only when recent settled same-candidate evidence shows that calibration improves Brier score versus the base AI model. If that proof is not present, the system automatically falls back to the base AI probability. AI probability still does not create, reject, suppress or modify production signals; Production Logic remains authoritative for qualification.

None of these probabilities is certainty. A 70% estimate still leaves meaningful loss risk.

5. What does “Recorded odds” mean?

The original issued odds remain permanently attached to each signal proof. For ROI, eSoccerSignal also considers genuine repeat scanner observations for that exact event, market, side and line during the original signal opportunity window. If a later scan within that same valid window records better odds, that actually observed price is used for ROI; observations after the signal validity window are excluded.

Prices are never reconstructed after settlement and the original issue snapshot, AI/Logic probability context, score and clock are never rewritten.

6. How are public results settled?

The original published signal record is resolved after final match data becomes available. The public status is recorded as won, lost or void. The public proof page is tied to that record and its commitment rather than creating a separate result entry.

Cryptographic proof demonstrates consistency of the stored commitment. It is not an independent certification of the upstream sporting event or bookmaker data.

7. How is win rate calculated?

Win rate is calculated from settled decisions only: wins ÷ (wins + losses). Voids are reported in the settled total but are excluded from the win-rate denominator.

Short samples can move sharply. The site therefore shows the reporting window and sample size alongside performance rather than treating one percentage as a permanent forecast.

8. How is ROI calculated?

ROI uses one equal unit for every settled signal with a valid recorded price. For each signal, the ROI price is the highest genuine scanner-observed odds for that exact issued selection within the original signal validity window; observations after that opportunity window are excluded, and when no later valid observation exists the issued odds are used. A win contributes odds − 1 units, a loss contributes −1 unit, and ROI is total net units divided by total staked units.

This deliberately removes personal stake sizing from the comparison. It is not a projection of future returns and does not represent what any individual user earned.

9. Why can there be long periods without a signal?

Continuous monitoring does not imply continuous publishing. If live states do not satisfy the active qualification and safety checks, no signal should be published. A quiet period can therefore be normal product behaviour rather than a scanner failure.

10. What are the important limitations?

Live feeds, market prices, clocks, notifications and third-party destinations can be delayed, changed or unavailable. A user may see a different price after opening a platform. Historical performance and AI-model performance can vary materially by time period, competition, strategy mix, model version and sample size.

eSoccerSignal is an information and alert service. It does not place wagers, control bookmaker accounts, guarantee wins, or guarantee that a notification will arrive before a live state changes.

11. What is public and what remains private?

Public pages can show monitoring status, broad signal categories, settled outcomes, performance windows and proof references. Exact production qualification rules, protected historical player/selection details, private account information and security controls remain private.

12. Who operates eSoccerSignal?

eSoccerSignal is operated by Seqube. Product and account support is available at support@esoccersignal.com.