๐Ÿ”ž 18+ Only  |  Gambling can be addictive โ€” please gamble responsibly  |  Gambling Help: 1800 858 858  |  GambleAware

How It Works

No black box. Here's exactly how we pick our tips.

1. Data Collection

Every day our scrapers pull fresh data: fixtures, recent form, head-to-head records, home/away splits, player injury reports and live odds from major bookmakers. All data is stored and normalised before model input.

2. Feature Engineering โ€” What Goes Into Every Tip

Raw data is transformed into meaningful predictive features. Every tip accounts for:

  • ๐ŸŸ๏ธ Venue & Home Ground โ€” quantified per venue from historical data. The Gabba, GMHBA Stadium and Optus Stadium each carry different home advantage weightings.
  • ๐Ÿฅ Injuries & Player Outs โ€” team injury reports scraped daily. Key player absences adjust team strength before the model generates predictions.
  • ๐Ÿ“Š Rolling Form โ€” win rates and score differentials over last 3, 5 and 10 games, weighted by recency.
  • โš”๏ธ Head-to-Head Records โ€” historical H2H at the specific venue. Some matchups are structurally one-sided regardless of current form.
  • ๐Ÿ˜ด Fatigue & Scheduling โ€” back-to-back games, road trips and short turnarounds all reduce performance in measurable ways.
  • ๐Ÿ“ˆ Live Odds & Market Pricing โ€” cross-referenced against model probability to identify where bookmakers undervalue a team.
  • ๐Ÿ” Win Streaks & Momentum โ€” teams on winning runs outperform their season average more than statistical regression would predict.
  • ๐ŸŒฆ๏ธ Season Context โ€” finals implications, dead rubbers, player resting โ€” all factored into the model's confidence output.

3. XGBoost Models

We use XGBoost (gradient boosted trees) โ€” the same family of models used by winning teams in major data science competitions. Each sport has its own model trained on multi-season historical data. Models are retrained weekly with new results to stay current.

4. Confidence Threshold

Only predictions above 65% model confidence make the cut. Below that, the model outputs NO BET. Higher confidence = larger recommended stake:

  • ๐Ÿ’ฐ VALUE (1u) โ€” 65โ€“74% model confidence
  • ๐Ÿ’ช STRONG (2u) โ€” 75โ€“84% model confidence
  • ๐Ÿ† BANKER (3u) โ€” 85%+ model confidence

5. Where to Place Your Bets โ€” We Recommend Betfair

We strongly recommend Betfair for placing your bets. Here's why:

  • No bookmaker margin. Betfair is a betting exchange โ€” you bet against other punters, not against a bookmaker. This means no built-in overround eating into your returns.
  • Better prices. Exchange prices on AFL, NRL and NBA are consistently 5โ€“15% better than traditional bookmakers on the same selection.
  • No account restrictions. Unlike Sportsbet or TAB, Betfair doesn't limit winning customers. You can keep betting at full stake as long as there's liquidity in the market.
  • Commission instead of margin. Betfair charges a small commission on winnings (typically 5%) โ€” far less than the 6โ€“8% margin baked into traditional bookmaker odds.
Example: A tip priced at $1.90 on Sportsbet might be available at $2.05โ€“$2.10 on Betfair. Over a full season of 100+ tips, that difference in price compounds into significantly better returns.

6. Tip Delivery

Qualifying tips are formatted and delivered to your Telegram DMs at 7am AEST every morning. Results are posted at 6pm AEST. Free channel members get the single highest-confidence tip; Members get the full card.

7. Staking Guide

Set your unit size at 1โ€“2% of your total betting bankroll. That means if you have a $500 bankroll, 1 unit = $5โ€“$10. Never bet more than 3 units on any single tip. Stay consistent โ€” the edge only works over volume.

Example: $500 bankroll, 1u = $5.
A BANKER tip (3u) = $15 bet. A VALUE tip (1u) = $5 bet.
Over 100 tips at 70% strike rate, that's sustainable long-term profit.