Every number on this site comes from a defined, consistent process. Here's exactly how tips are logged, settled, and counted.
A tip qualifies for the public record only if the model confidence is ≥ 65%. Tips below this threshold are not published and not counted. This threshold has been fixed since day one and has not been changed retroactively.
There is no selective logging. If a tip qualifies (≥ 65% confidence) it goes into the public record regardless of result. Losers are not deleted. The record is not pruned. You can download the full CSV at any time and count the rows yourself.
All P&L is measured in units, not dollars. One unit = 1% of a notional starting bank.
BANKER (85%+ confidence) = 3 units staked
STRONG (75–84% confidence) = 2 units staked
VALUE (65–74% confidence) = 1 unit staked
Stakes are fixed at time of tip publication and not changed after the fact.
P&L is calculated against the Betfair back price available at the time the tip was generated (typically the morning market price). Exchange and bookmaker prices can differ, and commission, liquidity and availability vary by market and timing. Return on a winning bet = (odds − 1) × stake. Return on a losing bet = −stake.
Results are pulled from the sports database after match completion. An automated settlement verifier cross-checks tip outcomes against actual match scores. Open real-money Betfair bets are shown separately on the record page until they settle — they are not pre-counted as wins.
Strike rate, ROI, units P&L, and monthly breakdowns are all derived fresh from the underlying tip database on every sync cycle (approximately every 15 minutes). There are no manually edited headline numbers.
Settled results only. Void/abandoned matches are excluded. A 55% strike rate means 55 out of every 100 settled tips were correct.
Cumulative profit/loss in units. A result of +201.9u means that following every qualifying tip at the specified stake would have returned 201.9 units of profit over the tracked period.
Return on investment as a percentage of total stake. Total units staked = 3u × BANKER tips + 2u × STRONG tips + 1u × VALUE tips.
Delete losing tips from the public record
Cherry-pick which tips to include in stats
Change the confidence threshold retroactively to improve past stats
Count open unresolved bets as wins before they settle
Manually adjust headline stats
Use best-possible odds (e.g. opening lines) to inflate returns
Odds are captured from the Betfair Exchange morning price, typically around 7am AEST, when the tip card is generated.
If odds move significantly before match start, the recorded price may differ from the price available to a user at execution time.
Void, push or abandoned events are excluded from strike rate and ROI calculations, but remain visible in the CSV record for transparency.
A typical 5% commission on winning exchange bets reduces actual user return compared with displayed gross P&L.
Execution timing, odds movement, commission, market liquidity and bookmaker-versus-exchange pricing can all change real user results.
Tips are deduplicated by event, market, selection and publication context so the same bet is not counted multiple times.
If a tip is republished because of a rain delay, fixture delay or similar operational issue, only the original qualifying tip is counted in the public record.
The model can be wrong. Some sports have smaller sample sizes. Variance over short periods can make results look much better or worse than the long-term record.
| Date | Change | Reason |
|---|---|---|
| 2026-03-01 | Initial methodology established | Launch |
The full tip database is available as a public CSV download. Every row, every result.
Past performance is not indicative of future results. Tips are for informational purposes only.