FunkyIQ builds proprietary predictive AI systems that synthesise complex, high-dimensional data into clear, high-confidence decisions — faster and smarter than conventional models.
Live Products
Prediction Engines. Running Now.
Live
🏉
NRL Footy Predictions
Match-by-match win probabilities for every NRL game. Multi-signal model synthesising team form, head-to-head history, venue, and injury data into a single ranked prediction.
Race-by-race win probabilities for Australian thoroughbred racing. Multi-signal model synthesising form, track conditions, jockey and trainer data into ranked selections.
Suburb-level price movement forecasts and property valuation signals. Combining macroeconomic data, sales history, and local supply metrics into actionable market intelligence.
Explore The Engines
A Closer Look
FunkyIQ scans every Australian thoroughbred race, every day, and processes millions of data points to rank each field and surface the races where the model holds a genuine edge over the market. One engine drives five distinct views, each tuned for a different style of play.
Track record — Sharp picks on the industry 5-tier scale
Our Sharp picks graded against the industry five-tier scale, Poor to Elite. The marker shows where FunkyIQ currently sits against the blind, public, and market-favourite benchmarks.
Our predictive models sit at the high end of the industry standard for horse racing. High win and placement rates don't eliminate risk — every race can go either way. Figures are provided for context; betting decisions are the user's own responsibility.
Disclaimer
Past performance is not a reliable indicator of future results. Horse racing betting involves inherent financial risk. Our service provides statistical data modelling for informational purposes and does not guarantee net profitability. If you require assistance, please call GambleAware on 1800 858 858. Think! Before you bet.
All statistics on a flat, single-unit stake on a single horse (rank-1 Sharp pick) per race. Sample: only genuine pre-race predictions (source="live", enriched before race jump). Backfilled and post-race reconstructions are excluded to prevent look-ahead bias.
Five views, one engine
The Lounge — every Australian race today, ranked top to bottom with a win and place percentage for each runner. The foundation the other views drill into.
The Edge — high-confidence picks compared directly against market odds, highlighting where the model most disagrees with the bookmakers.
Hot Seat — a live race-card view sorted by jump time, built for tracking what's worth backing right now.
Playbook — the day's lower-variance shortlist, distilled into a handful of concrete plays.
Labs — a paper-trading sandbox that grades exotic strategies against real dividends, with a full ledger and ROI tracking.
Confidence tiers
The model's top pick in each race is graded by how much win probability it carries. Premium is the rarest: strong model conviction meeting a market that's underpricing the horse. Hot covers the model's highest-conviction races, High Confidence the next band down, and Strong the broadest tier where there's a clear top pick with meaningful separation from the rest of the field.
FunkyIQ produces a match-by-match win probability for every NRL game. The engine weighs both sides of a fixture and returns a calibrated probability for each team, alongside the model's ranked call on the result.
Season track record
Head-to-head tips graded against the industry five-tier scale for tipping accuracy, from a coin-flip floor to elite.
74%
121 / 164 match tips correct this season
5.5 / 7
average correct tips per round
Tipping Accuracy
👑 FunkyIQ 74%
Poor
Avg
Good
Great
Elite
50%58%63%68%73%80%+
📊 Expert Tipster (High Bar)📰 Media / Public Avg🎯 Coin Flip (Floor)
📊 Season Record to Date
Match Tips Correct121 / 164🏉 74% accuracy
Average Per Round5.5 / 7🥇 Elite consistency
Disclaimer
Past performance is not a reliable indicator of future results. Predictions are provided for informational purposes only and do not guarantee outcomes. If gambling is causing you harm, call GambleAware on 1800 858 858. Think! Before you bet.
Our multi-layered model architecture captures non-linear relationships and hidden signals that standard statistical approaches miss entirely.
⚡
Real-Time Inference
Models update continuously as new data arrives — predictions always reflect the most current state of the world, not a snapshot from hours ago.
📈
Self-Improving Models
Every outcome feeds back into the system. The engine recalibrates continuously, compounding its accuracy advantage over time.
🔍
Deep Feature Synthesis
Thousands of raw signals are distilled into a compact set of high-signal features — removing noise, retaining predictive power.
🎯
Calibrated Confidence
Every prediction carries a calibrated probability score — so you always know how certain the model is, and when to act on it.
🔗
Multi-Source Data Fusion
Structured data, unstructured text, and real-time feeds are unified into a single coherent model input — no signal left behind.
Methodology
How Our AI Works
Every FunkyIQ prediction engine is built on the same principle: no single number tells the whole story, so we don't rely on one. Each engine synthesises dozens of independent signals per outcome — and weights each signal by how much predictive power it has actually demonstrated on historical results, not by intuition. As real-world outcomes arrive, those weights are recalibrated, so the model that makes today's predictions is a slightly sharper version of yesterday's.
Model versus market
A prediction is only interesting if it disagrees with the crowd. For every selection, the model calculates its own implied probability and compares it to the probability baked into the current market price. When the two line up, there's no edge to act on. When the model rates an outcome materially higher than the market does, that gap — what we call the Edge — is surfaced explicitly, so you can see not just what the model likes, but where it thinks the market is wrong.
Calibrated confidence
A 70% prediction should be right about 70% of the time. That property — calibration — is what separates a probability you can act on from a number that merely looks confident. FunkyIQ grades every prediction with a calibrated confidence score and sorts outcomes into clear tiers, so you always know how much conviction sits behind a pick and when it's worth backing.
Always learning
The engines never stop running. Behind the scenes they continuously ingest fresh results, re-fit their patterns, and re-check their own calibration. The predictions you see reflect everything the system learned from the last outcome — not a snapshot frozen weeks ago.
Reference
Key Terms, Explained
Win Probability
The model's estimate of how likely an outcome is, expressed as a percentage. A field's win probabilities always sum to 100%.
Calibration
The quality that makes a probability trustworthy: over many predictions, outcomes rated 40% should happen about 40% of the time.
Market-Implied Probability
The chance a price implies, calculated as 1 divided by the decimal odds. A $4 runner implies a 25% chance.
Edge (Overlay)
The gap between the model's probability and the market-implied probability. A positive edge means the model rates the outcome higher than the price does.
Place
A finish in the paying positions — typically the top three in larger fields, top two in smaller ones. A place bet wins if the runner places.
Starting Price (SP)
The odds on a runner at the moment the race jumps — the standard reference price for settling and comparison.
Barrier Draw
The starting gate a horse is allocated. Inside and outside draws can materially help or hurt depending on distance and track.
Track Condition (Going)
The state of the racing surface, rated from Firm through Good, Soft and Heavy. It changes how a race is likely to be run.
Quinella
An exotic bet requiring you to pick the first two runners home in either order.
Trifecta
An exotic bet requiring the first three runners home in the correct order.
First Four
An exotic bet requiring the first four runners home in the correct order.
Head-to-Head
In footy, the market on which of the two teams wins the match outright — the simplest form of the bet.
Confidence Tier
FunkyIQ's grade for how much conviction sits behind a top pick, from Strong up to Premium.
Questions
Frequently Asked
What is FunkyIQ?
FunkyIQ builds proprietary AI prediction engines. Two are live today — NRL footy and Australian horse racing — each turning large, messy datasets into ranked, calibrated win probabilities you can read at a glance.
How accurate are the predictions?
The engines are designed to be well-calibrated, meaning a stated probability should hold up over a large sample of predictions. No model is a crystal ball — sport and racing are inherently uncertain, and any single result can go either way. We track performance openly on the product sites.
Is this betting or financial advice?
No. FunkyIQ is an informational tool that provides statistical modelling and probabilities. It does not guarantee outcomes or profitability, and nothing here is financial advice. Any betting decision is your own.
How often do the predictions update?
Continuously. Models re-run as new data arrives — form, results, prices and team news — so the predictions reflect the latest available information rather than a fixed snapshot.
Where do I see the predictions?
Each engine has its own site: the horse racing predictions and NRL footy predictions links above take you straight to the live product.
Play Safe
Responsible Gambling
FunkyIQ provides statistical predictions for informational purposes only. Past model performance is not a reliable indicator of future results, and no prediction eliminates the inherent risk in betting.
Betting is for adults 18 and over. Only ever stake what you can afford to lose. If gambling is causing you harm, free and confidential help is available.
Gambler's Help / GambleAware — 1800 858 858. Think! Before you bet.