The Fynbos Hypothesis:

 

In our post dated August 30th, we dissected the surprising failure of Indian Derby winner Fynbos in a comparatively lower grade race, despite running amongst a field that consisted of 70% of her own stablemates. At the time, we put forward a distinct opinion: Fynbos’s actual dosage traits—which limit her effective stamina reach to 2000 metres—were starting to catch up with her after a period of rest.

While her earlier Derby triumph could be largely attributed to raw muscle power, her post-rest performances suggest a shift where lung power and stamina limitations have become the dominant factors.

The Pune St. Leger Test

This hypothesis was put to the ultimate test today in the St. Leger Cup in Pune. In a setup remarkably similar to her previous failure—once again surrounded by three of her stablemates—Fynbos was surprisingly, and hotly, fancied in the betting rings. Punters chose to trust champion trainer P. Shroff’s instincts, heavily backing the filly over an unproven and grueling 2800-metre trip. The fact DUKE OF TUSCANY  her detractor in last run was beaten overwhelming in President of India Gold Cup at Malakpet yesterday had a little bearing on Punter's confidence.

The result? She failed again, ultimately being beaten by a stablemate.

Dissecting the Race Video

A close review of the race footage reveals exactly where the wheels came off:

• The Mile Marker: After traveling a mile, Fynbos became visibly keen to run free.
• The Bend: Despite her jockey's best efforts to hold her back and conserve energy, she began moving up places rapidly.
• The Straight: The horse took the lead immediately after the bend and briefly shaped like a winner.
• The Finish: The stamina deficit struck hard. Her strides shortened drastically in the final stages, and by the end, she looked as though she was literally stopping on the track.

The Bottom Line

Fynbos has now failed twice in significantly easier grades than her historic Derby win. Watching her struggle to see out the distance makes it safe to conclude that her limits are firmly exposed. Expect her to face uphill tasks in the near future if she continues to race anywhere near or beyond the 2000-metre mark.

Horse Racing Data: Too Much Noise, Less Sense?

Essence is mine, language and syntax is AI !!

 I have been compiling horse racing data for the last 15 years. During the COVID-19 pandemic, I even expanded my database to include races from Singapore and Hong Kong. With 160,000 lines of India Horse Racing data sitting in front of me, I recently arrived at a staggering realization: horse racing data is often more noise than sense.

My journey to this conclusion was built on years of manual grind, shifting philosophies, and a sudden realization that sometimes, less truly is more.

The Manual Grind: Building a Database from Scratch

As someone with absolutely zero programming or automation knowledge, I had to build everything by hand.

The Excel Struggle: Initially, I manually entered every single variable—distance, timings, current and previous weights, penetrometer readings, and track centers.
The Formula Nightmare: I spent endless hours just figuring out how to convert race timings format (mins:secs:secs) in Excel.
The Monster Worksheet: Over time, I built massive, complex formulas to pull data automatically just by typing in a horse’s name. Looking back at those nested formulas today, I genuinely wonder how I managed to write them with such limited technical knowledge.

Eventually, to reduce my manual burden, I expanded to compile data across all racing centers. The database grew into a massive data monolith.

Enter AI: Trying to Tame the Chaos

When my database hit 160,000 lines, I decided to level up. I turned to Advanced Artificial Intelligence, deploying high-end decision tree techniques to uncover hidden patterns, anomalies, and winning formulas. I thought the sheer volume of data would finally reveal the sport's hidden code.

The result? The AI failed.

Horse racing possesses an overwhelming number of variables—jockey changes, track bias, weather, gate positions, and human intent. It is near impossible to derive a single, workable automated method. The machine learning models simply choked on the noise. I realized that a horse is not a machine, and a race track is not a laboratory; there are elements of intent and animal psychology that a spreadsheet can simply never capture.

The Contrast: Dynamic Ratings vs. Merit Handicapping

This AI failure forced me to look back at where I started, revealing a stark contrast in my handicapping journey:

Feature	The Old Way: Dynamic Ratings (DR)	The New Way: Merit-Based Handicapping
Complexity	Bare minimum data entry; simple and lean.	Massive database; 160k lines of complex variables.
Core Logic	Calculated base ratings + checked if top-rated was active in trials.	Complex statistical modeling and AI decision trees.
Success	Regularly caught 20/1 longshot winners with ease., sometimes even 100/1	Lower satisfaction; drowned out by data noise.
Current State	Still shows a high strike rate at great odds today.	Hard to find consistent, satisfactory patterns.

The Illusion of Total Control

In hindsight, I fell into a trap that many data enthusiasts face: the illusion of control. We mistakenly believe that if we just add one more variable—the wind speed, the jockey’s recent form, the exact depth of the turf—the picture will become clear. In reality, every new variable we add introduces new chaos. We trade our gut instinct and sharp intuition for an endless maze of statistics, effectively blinding ourselves to what is happening right in front of us on the track.

Life Comes Full Circle

I often wonder: What if I had just stuck to my original Dynamic Ratings method? It required a fraction of the work, yet it yielded incredible satisfaction and massive payouts. Even today, when I look back at the old DR system, the top-rated horses still boast a phenomenal strike rate at highly profitable odds.It proves that a streamlined, focused strategy often beats a hyper-complex algorithm.

But life has a funny way of coming full circle. Once you train your brain to look at a massive web of variables, it is incredibly difficult to untrain it. I find myself trapped in the very noise I created, constantly trying to make sense of a beautifully chaotic sport. The irony isn't lost on me: I built a massive database to find clarity, only to realize that clarity was there at the very beginning, sitting in a simple, elegant formula.

Sometimes, the best data system is the simplest one.