Musings

// Brief observations from the digital garden

Platform 6 at KSR Bengaluru had the color of old steel and turmeric light. Porters stood near the edge with red shirts darkened at the collar, scanning the indicator board like fast readers of thin liquidity. Families clustered around painted numbers on the concrete, each group holding position for a coach that had not yet arrived. When the train entered, the air shifted first — diesel, warm iron, and the dry friction of brakes — then the crowd repriced distance all at once.

For a minute the platform became a visible book. Reserved passengers were patient depth with printed certainty; unreserved travelers moved like urgent flow, searching for any open door before the spread closed. I stood with my backpack and a folded draft, watching hands pass tiffin boxes through windows before departure. The whistle was a hard timestamp. After it, every late decision paid impact.


The first heavy monsoon burst arrived over Manipal in uneven sheets, as if the sky were sending orders in bursts instead of a steady stream. Tin roofs amplified each strike into sharp white noise; between bursts there were brief, suspicious silences. By evening the gutters were full and the neem leaves held a brighter green, almost over-marked to market. I walked back from the department with damp cuffs and a notebook wrapped in yesterday’s newspaper.

Rain in this city rarely mean-reverts on schedule. Quiet hours cluster, then sudden intensity, then another spell of quiet that feels provisional. Volatility behaves similarly in the data — not random enough to be comforting, not persistent enough to be simple. At night I dried my pages under the fan and revised a proof line by line. Outside, thunder moved farther away but never fully left the book.


Before sunrise, the temple on our street opens without announcement: one tube light, one broom across stone, one priest arranging flowers with the economy of a practiced trader. Then the first bell rings. The sound is bright and metallic, expanding through lanes where shutters are still half down. Dogs pause. A fruit seller aligns guavas into small pyramids. People who were walking without urgency begin to move with purpose.

I think of opening auctions when the bells overlap — dispersed intentions pulled into a single print. For a few minutes, private estimates become public rhythm. The city accepts a reference price for the morning: incense, camphor, diesel from the first bus, and the wet granite smell of washed steps. By the second bell I am at my desk, reading overnight notes and updating priors. The day starts less with a decision than with a clearing.


I thought the paper was finished last week. The PDF had clean section breaks, references aligned, and a concluding paragraph that sounded final enough to fool me. Then, on the bus home to Bengaluru, I reread Section 3 under a flickering yellow light and felt a quiet mismatch: the theorem held, but the story around it overclaimed. Completion had been a cosmetic print, not true price discovery.

Rewriting is a form of adverse-selection control. You pull stale orders, widen claims, wait for better information, then quote again with less bravado. I spent Sunday at the dining table with annotated pages spread beside a steel tumbler of water gone warm. The edit was mostly subtraction: one adjective removed, one lemma moved earlier, one figure caption made honest. By evening the argument felt smaller and more durable. Not finished, but now tradable.


Elegant models are good company in the wrong way. They listen, they do not interrupt, and if you choose assumptions carefully they reward you with smooth surfaces and interior consistency. Late at night in my room, with traffic reduced to a distant ribbon, I can spend hours inside a few equations and feel briefly protected from contingency. The symbols hold their posture even when I do not.

But elegance can thin social liquidity. Friends call and I say I will reply after one more calibration; family asks what I am working on and I offer a compressed answer that clears nothing. The loneliness is not dramatic, just narrow spread over long horizons. Some evenings I close the notebook and sit on the balcony until the neighboring pressure cooker whistles. Real life arrives as unmodeled flow: noisy, costly, and necessary for identification.


I returned to Bengaluru with a thesis in my bag and a softer sense of time. Hong Kong felt like a clean execution venue — everything precise, the bars high, the spreads tight. Here the air is warmer and the cadence is human. I am a Fellow in Residence at MAHE now, which feels like a quiet threshold between student and researcher. The work on belief management and optimal execution is complete in the administrative sense, but the questions still breathe.

At a cafe near a traffic circle, I watched the morning queue dynamics unfold. Autos cut in like market orders, buses hold the line like patient depth, and the slow merge is a Bayesian update in public. Rain tapped the window in small pulses. The barista dragged chalk across a menu board — a small, precise sound that reminded me of orders arriving one at a time. I am learning that completion is only a timestamp. Still updating.


I have been listening to Sri Venkateswara Suprabhatam in the mornings, a small ritual after writing about attention arbitrage. I do not understand every line, but repetition is a kind of scaffolding. The chant arrives a fraction sooner each day, like a price discovery loop that remembers the last print.

Microstructure taught me that liquidity forms where attention rests. When I split my focus, the book thins. I am practicing a simple constraint: one screen, one cup of coffee, one sound. The discipline is not heroic; it is procedural. Some mornings I fail and the spread widens anyway.


The Almgren-Chriss curve still feels like a life strategy: fast at the start, slow in the middle, fast at the end. The urgency profile is not just a trading heuristic; it is a rhythm for work. Begin with a burst, then sink into the slow liquidity of craft, then push through the closing auction of a deadline.

I am writing a novel titled Rain Hits Glass, and the pacing behaves the same way. A story needs momentum early, patience in the middle, and a clear exit. The middle is where most of the signal hides. Not all value shows up in the arrival price.


I cycled through Cubbon Park at dawn. The chain made a steady, metallic whisper, and the air smelled like wet soil and eucalyptus. The city wakes like an order book: first the walkers, then the vendors, then the blur of scooters. Each layer adds depth. You can feel the day set its price.

Vendors setting up carts are liquidity providers arriving — laying out inventory, quoting with their posture. The old man with the tea urn is a patient bid. A sudden dog chase is a market order. I returned home with clearer thoughts and a quiet ache in my legs, still rolling.


I read Kyle (1985) for the tenth time and still find a new corner of it. The model is small, almost austere, but its implications are fractal. Every reread changes the story. The informed trader is quiet, but the market listens harder than it admits.

Information wants to be traded, but slowly. Good papers age like whiskey; they get smoother without losing the burn. I underlined a line I have underlined before, and the chalk dust on my fingers felt like a reminder that even clean theories leave residue.


A Starbucks near the station has become my trading floor. The ambient noise is market noise — orders called out, milk steaming, cups clinking. Too quiet and I overfit; too loud and I miss the signal. There is a narrow band where focus finds depth.

I watch people execute their own life strategies. The student sprints in and out like an aggressive market order. The retired couple parks like a patient limit order at a good price. Everyone has an arrival price, and no one writes it down.


In Munnar, the Shola forests sit like sky islands — fragments of green floating in mist. The ecosystem is broken into pockets, but the pockets still speak to one another through water, birds, and hidden roots. It feels like fragmented liquidity: visible pools separated by distance, connected by unseen pathways.

The wind moves in small bursts, and the leaves answer in quiet waves. There is structure inside the fragmentation, a geometry you do not see until you are patient. Markets and forests share that patience. I am still learning to stay long enough to notice.


The gap between model and reality keeps widening and narrowing like a reluctant spread. Bayesian market makers are elegant on paper; in code they are temperamental. The belief updating rates that look clean in notation behave like wet clay once I implement them.

I keep a log of failed experiments and the line count keeps rising. It is humbling to build something that does not quite work, and to keep building anyway. I am learning to fail forward, slowly.