A $100m multi-asset book · Sai Sumedh Garimella

What the book held

Single-name semiconductor and storage equity in Micron, AMD and Seagate. WTI crude. Nasdaq index futures as the beta overlay. Mexican peso as the emerging market FX expression. CDX as the credit leg.

Five very different instruments, which is the point. A book that is long semis, long crude and short credit protection is not diversified. It is one macro view expressed three times, and it finds that out on a single afternoon.

The mechanics that mattered

Curve steepeners. Constructing the trade so the view sits on the slope, not accidentally on the level.
DV01 neutral sizing. Two legs balanced in notional are usually not balanced in risk. Sizing on duration exposure is the only version that means anything.
CDS index selection. Choosing the index whose constituents carry the risk you are trying to express, rather than the most liquid one.
Kelly sizing. Useful as a ceiling rather than a target. Full Kelly on a discretionary view with an unstable edge estimate is a route to being right and still finishing behind.

Most of the difference between a good book and a bad one over five weeks is not the calls. It is whether position sizes reflect what you actually know.

Alongside this

The same year, IMC Prosperity 4: second globally in manual trading with team ROARINGLIONS, and top 6.5% overall of 18,803 teams. And the Natixis Students Investment Challenge with the Green Folio fund, on a 9.4% return against a 1.8% benchmark, a Sharpe of 2.04 and beta of 0.35.

The Natixis numbers are the ones worth pointing at. A beta of 0.35 alongside that return means the result did not come from taking more market risk than the benchmark.

A trading simulation is not a track record. It removes the two hardest constraints in the real thing, which are capacity and the cost of being wrong with someone else's money.

← Back to work