Real-time Telegram alerts powered by institutional order flow detection. Two independent systems. Audited results. No tipster guesswork.
Performance shown Jan 25, 2026 – Apr 3, 2026. Flat 1u win-only staking. Past performance does not guarantee future results. All figures auditable from timestamped Telegram messages.
Our system continuously monitors order flow, depth imbalance, price momentum, volume surges, and spread dynamics across all live Betfair markets — processing thousands of data points every second.
Proprietary algorithms classify each runner's activity into institutional patterns: sustained accumulation campaigns, controlled price compression, and order book signatures that retail flow cannot replicate.
When a signal passes our 13-gate qualification pipeline — including campaign maturity, conviction scoring, and price zone validation — you receive a Telegram alert with full signal context and suggested staking.
Illustrative example only. Individual bets can win or lose. 30.4% of Pro Channel signals win at flat staking.
Flat 1u win-only. 11 weeks. Jan 25 – Apr 3, 2026. Green = profit, red = loss.
7 of 11 weeks profitable. Worst: -113.3u (W12). Best: +91.0u (W10). Peak equity: +258.0u. Current: +227.2u.
Flat 1u win-only. 11 weeks. Jan 23 – Apr 2, 2026. 9 of 11 weeks profitable.
All 8 complete weeks profitable. Best: +40.0u (W10, +72.6% POT). Only partial weeks (W03, W13) are negative. Sharpe ~8.3.
Pro Channel (green) + Tape Alerts (gold) = Combined (white). Both channels stacked.
● Pro: +227.2u ● Tape: +167.8u ● Combined: +395.0u
Peak combined equity: +413u (W11). Notice W12: Pro dropped -113u but Tape added +6u — genuine diversification buffering the worst drawdown. The gold (Tape) component grows steadily regardless of Pro's volatility.
Every alert is a Telegram message with a timestamp. Every race result is public record. Match the two and you have our complete P&L — verified by you, not us.
If a service won't show you their losses, they're hiding something. We publish every result — wins, losses, and drawdowns.
POT = Profit on Turnover. All figures from live, timestamped Telegram alerts. 25 Jan – 3 Apr 2026. View full performance dashboard →
Every signal must pass 13 sequential gates before it reaches subscribers. Each gate exists because data showed a specific pattern of loss without it. From 69,746 raw market signals, only 3,178 (4.6%) survive to become subscriber alerts.
652 signals with accumulation streak ≥7. The 95% bootstrap confidence interval for POT is entirely above zero.
ELITE is the system. The remaining 2,526 non-ELITE signals are flat. The 2x stake multiplier is backed by a CI entirely above zero.
505 selections with A+ microstructure grade. Strike rate holds at 56.0% against a 25.2% breakeven rate.
Wilson CI lower bound (50.7%) clears breakeven by +25.5 percentage points. The widest measured edge in the SDM system. Edge is strengthening in H2.
Where the edge concentrates. Green = profitable, red = value destroyer. This analysis drives our gate decisions.
| GRADE | $3-4 | $4-5 | $5-6 | $6-8 | $8-10 | $10-12 |
|---|---|---|---|---|---|---|
| Grade A | +18% | +12% | +65% | +19% | -61% | — |
| Grade B | +17% | +15% | +40% | +19% | -17% | +46% |
| Grade C | +6% | -5% | +1% | +16% | -40% | +52% |
The $8-10 column is red across all grades — this is the system's primary value destroyer (-120.8u total). CONFIDENT $8-10 alone leaked -129.6u at -65.5% POT. Our gates now block this zone for non-STRONG signals. The $5-8 range is the sweet spot: consistently positive across all grades.
A+ grade requires ALL five microstructure conditions to be true simultaneously. This is the most selective tape read in the system.
The IMB 60-80 band (strong back-side imbalance) delivers +50.2% POT with 66.7% SR — the system's highest-efficiency segment. All IMB bands are profitable. d_imb ≥ 5 is used as a subscriber quality filter, removing noise-zone signals (-16.1% POT) while preserving 96% of P&L.
21,000+ lines of Python running on FastAPI + WebSocket. Processes every runner in every AU/NZ thoroughbred race in real time.
Full methodology at /methodology. Performance data at /performance. The system does not use form data, trainer/jockey stats, or model-implied probability as decision criteria.
Placeholder testimonials — to be replaced with verified subscriber reviews
⚠️ Limited Subscriptions Available
We deliberately cap the number of active subscribers to protect signal edge and execution quality for existing members. When the market knows what we know, the edge disappears. Subscriptions may close without notice once the cap is reached.
All plans include a 7-day free trial. Cancel anytime. Prices exclude GST.
*ML Predictions channel currently in shadow mode. Elite subscribers receive access upon launch (expected April/May 2026). All prices exclude GST. Pricing subject to change.
No spam. No sharing. You'll receive setup instructions only.
We'll email you with access details and your onboarding guide shortly.