Three horses, 19/08/2026 at Ipswich, and what the model saw
On Wednesday the 19th of August a card ran at Ipswich. I opened three of the runners in the model, the same way I would for any owner, and the three of them together show what the model is actually for. None of this is a tip. It is the read on each horse, its class, what it needs, and where the result might be lying.
Moscow Mule, the genuine grade horse
The model rates Moscow Mule in the top 19 percent of Queensland horses on class, and class is the part that matters. Class here is not the clock. It is whether a horse beats the rivals it actually meets, and this one does. The model calls it a grade specialist, a horse that wins its grade without posting flashy times. Its last three runs read second, second, third, all of them close, and the model says it keeps grinding late in every single one. It wants about 1400 metres on soft ground. Put it there and it is honest every time. The one caution the model raises is that it is sitting below its own best right now, so it is a watch, not a blind follow.
Kadesky, the class horse to look at twice
Kadesky is also a genuine class horse, top 22 percent, another grade specialist that beats rivals rather than the clock. This is where the model earns its keep. It flags a stride watch on Kadesky, a real drop in its stride once you adjust for the going, and it reads the horse as falling below its best, 37 points under its peak. So the plain read is this. Good class, but showing a physical warning sign and out of form, so treat it carefully until the movement steadies. That is the sort of thing a bare result on a Wednesday never tells you. It wants around 1350 metres on a track with a long straight, because the model reads it as a late-burning finisher that needs room to wind up.
Veomas, fast on the clock, not yet proven on class
Veomas is the interesting one, because it shows the two lenses pulling apart. The model tags it a fast horse, quick on the clock, but its class only sits at midfield, the 46th percentile. Fast figures with modest class is the profile the model treats with care, because the number can flatter the horse. It has only two runs on file, so the model is upfront that the read is provisional. What stands out is the setup. Its best internal run by a distance was over 1350 metres at Doomben, not the short trips. So a horse the market might chase on raw speed is one the model says wants a middle distance to be seen at its best.
What the three have in common
Three horses, three very different animals, and the finish line on any one day would not tell you which is which. One is a proven grade horse. One is a class horse with a small red flag showing. One is a speed type whose class has not caught up yet. That is the whole idea. The model does not tell you who wins on Wednesday. It tells you what each horse is, what it needs, and where a plain result might be lying. What you do with that is yours.
You can open this same read on any Queensland horse, at horsetruth.com.au. Type in a name and you get the full file. The class, the type, the trip and track it wants, and whether its last run was better than it looked. Not a tip. The read.