Paul Daily - ratings2win.com.au

ALT Platform update: The Next Axis Development Cycle. We've begun rolling out improvements to the methodology behind the platform: refining calculation standards, including how race and sectional times are broken down, and strengthening performance assessment. We've recalibrated modelling layers, combining 50+ years of team experience with the latest AI models to go deeper, test faster, and uncover new analytical possibilities. Our work has shaped industry direction for 15 years, with a focus on predictive performance, not just presentation. Axis's development has always covered the platform and technology, and, more importantly, the data, methodology and processing systems beneath it. Each cycle renews both, including groundwork for future API and AI integration. We build for serious, active, privacy-conscious users, and rely on it ourselves for real wagering decisions. That's what Axis stands for: numbers you can trust with real money, not just ones that look good in a headline.

ALT Ratings2Win Indicators, 14 to 16 October 2026. Friday 16 Oct: 27 winners from 55 races, including Cranbourne 6 from 8. Itsukushima $13, Smokin Rohie $13, Family Of League $9. Thursday 15 Oct: 22 from 30, a 73% strike rate, including Ipswich 6 from 7 and Warrnambool 6 from 8. Sonnig $31, Mystacheeno $11, Costa Del Darlo $10, Eagles Fire $10. Wednesday 14 Oct: 21 from 45, including Happy Valley 7 from 9 and Caulfield 5 from 8. Hamadryad $14, Johnich $14, Elegant Life $11, Charming Babe $11, Cho Oyu $8. Four-day total: 148 winners from 259 races, a 57% strike rate.

ALT Ratings2Win Indicators results, Saturday 17 October 2026: 78 winners from 130 races, a 60% strike rate (usual is about 55%). Caulfield 7 from 10 races, Doomben 8 from 10, Randwick 6 winners. Feature winners: Caulfield R1 King Snitz $16, Caulfield R3 Gentle Steel $21, Caulfield R9 Kingswood $26, Randwick R7 Tempted $6.50, Toowoomba R3 Araminta $16, Ascot R6 More Than Enuff $16, Gold Coast R6 Canadian Fling $13, Morphettville R6 Artistic Genius $11, Avoca R3 Secret Crossing $11, Newcastle R3 Lady Talea $9, Bendigo R1 Quynh $9.

ALT This is the final piece of our data overhaul: the Intelligence Rating (IR), R2W's premier predictive rating and the culmination of the platform's predictive automation. Highlighted stat box: IR Strike Rate 31.4%. Top-selection strike rate for the revised IR, a 6.4% improvement on the previous version. Comparable gains extend through the Top 3 and Top 5, across both wins and places. Bullet points: IR brings together the depth of our modelling with the domain knowledge and racing-specific angles built up over decades. The rating also continues to identify winners the market prices well outside its leading chances, including a $31 winner and hundreds of winners at $8 or better. Closing line: A higher strike rate and continued value away from the obvious market selections are what set the Intelligence Rating apart.

ALT SR, Improved. Text: Continuing our data overhaul from Pars, Base and PIR, the SR has now been updated. SR stands for Speed Rating, one of the original predictive foundations of Axis. Highlighted stat box: SR ~6.8%. Top-3 SR predictions found the winner 6.8% more often, with comparable gains in the Top pick and Top-5. Bullet points: The underlying domain logic hasn't changed. The modelling beneath it has, now drawing more efficiently on the full body of rated form for each horse. Loss on turnover across an assumed backing of all horses/races substantially reduced, without sacrificing the model's ability to identify $10+ winners. The gains come from more mature modelling and from stronger foundations now shared across the platform. The rating scale and the way it's used remain unchanged.

ALT Data Overhaul: PkD, Improved. Continuing our data overhaul from Pars, Base and PIR, the PkD has now been updated. PkD stands for Peak Distance Rankings. Stat: PkD, approximately 1.6%. Top-1 accuracy gained 1.6%, with Top-3 and Top-5 up 1.25% and 0.6%. The strongest improvements were in provincial areas, up 2.7%, and rural areas, up 4.3% from a smaller sample size. The PkD model shows consistent improvements across all metrics. The provincial areas saw the strongest improvements. In metro regions, only the Top-3 improvement was statistically significant.

ALT Data Overhaul: PIR, Improved. Continuing our data overhaul from Pars and Base, the PIR has now been updated. PIR is our prediction of a horse's expected position in run. Stat: PIR, approximately 4%. Settling calls are around 4% more accurate overall, with the biggest gains in large fields and for horses that settle back. Lead prediction has also improved. We've improved how we predict where each horse settles during a race, with a number of additional variables now feeding the model. Predicted Race Speeds have also improved.

ALT Data Overhaul: Par Times and Base, Rebuilt. Our new generation of par times is now in production, the benchmark standard for every track, distance, class and age, feeding into Speed Ratings, sectional ratings, class conversions and the wider ratings chain. Stat: Pars and Base, approximately 2%. Across 18 months and more than 25,500 races, the top selection from Speed Ratings, built from the new pars and the track variant, found the winner about 2% more often. Every model was rebuilt and required to prove a measurable gain in blind testing before acceptance, and the gain came before any separate refinement of the Speed Rating itself. Place accuracy and top-five rankings also improved, and loss on turnover across an assumed backing of all horses and races was substantially reduced, with the scale and presentation of the ratings unchanged. Splits and Sectionals were also updated across races and trials, with the biggest gains at country and provincial tracks and unusual distances.