How the firm forms a view, commits capital to it, and takes the risk off.

Markets and instruments

The firm trades exchange-listed options and futures. Our core activity is in equity index derivatives, where depth across strikes and expiries makes it possible to hold structured volatility exposure without depending on a single line of the surface. Alongside that, we trade interest rate futures and options, and we take positions in a small number of commodity complexes where the term structure carries information we can price.

We do not warehouse bilateral exposure that cannot be independently valued or closed. Everything the firm trades clears through a central counterparty, which imposes a discipline on the business that we regard as an advantage rather than a constraint. Margin is a real cost of carrying a position, and treating it as one tends to produce more honest sizing than a framework in which leverage is cheap and largely invisible.

Capacity is deliberately limited. Several of the relationships we trade are only exploitable at a size the market will absorb without moving against us, and we would rather run a smaller book at its intended risk than a larger one on degraded terms.

Forming a view

We make no claim to know the direction of the underlying market, and our strategies are not built on that claim. What we do instead is price volatility carefully and take the other side of the gap when the market's quoted levels part company with our own valuation.

In practice this means fitting the implied volatility surface across strikes and maturities, examining how that surface has behaved through comparable conditions in the past, and identifying the points at which the premium being paid or received is inconsistent with the risk actually being transferred. Some of that opportunity sits in the spread between implied volatility and the volatility subsequently realized, which is persistent but far from constant. Some of it sits in the relative pricing of one part of the surface against another, where the market prices a correlation or a term relationship more confidently than the available evidence supports.

Directional exposure does appear in the book from time to time. When it does, it is a consequence of a relative value position we wanted for other reasons, and it is hedged or sized accordingly rather than treated as a source of return in its own right.

Position construction

Signal generation is systematic. Our models produce the candidate positions, the hedge ratios that neutralize the exposures we do not intend to hold, and the size at which a given trade is consistent with the limits already assigned to that strategy. Discretion enters in a narrow and defined way, in judgments about liquidity, about the execution path through a scheduled event, and about whether prevailing conditions still resemble those in which a model was validated.

Each position is built with its exit specified. Before a trade goes on, we know the level at which the thesis has been invalidated, the level at which the risk has grown beyond its allocation regardless of whether the thesis still holds, and the instruments through which the position would be unwound in a market where the primary line has stopped quoting. A position whose exit depends on conditions remaining normal is not a position this firm is willing to hold.

Risk framework

Risk management at Northbridge is a precondition of trading rather than a review conducted afterward. Every book operates against defined limits on net and gross exposure to the primary greeks, on concentration by underlying, by expiry and by strategy, and on loss under a specified set of stress scenarios. Those limits are set when a strategy is approved, and they are not adjusted to accommodate a position that has become uncomfortable.

Our stress testing is built around the moments when models are least reliable. We test gap moves rather than continuous ones, shocks to the volatility of volatility itself, and the breakdown of correlations a portfolio has been implicitly relying on, since the historical relationships that make a book look diversified are usually the first thing to fail under pressure. We test the liquidity assumption directly as well, by asking what an exit would cost if the instruments we intended to use were unavailable at the size we needed.

The firm carries tail protection as a standing expense. In quiet periods that protection reduces returns and looks unnecessary, and we accept the drag, because hedges bought once a move is already underway are priced by a market that has understood the problem. The alternative approach, in which convexity is acquired reactively, has ended a number of firms with better return histories than ours.

Risk oversight is independent of the traders whose positions it governs. A limit breach requires the position to be brought back within its allocation, rather than the limit to be reconsidered.

Research and review

We begin from the assumption that any edge decays. A relationship that pays today is being observed by other participants with comparable resources, and the return it offers will compress as capital finds it. Research at the firm is therefore continuous, and a meaningful part of it consists of establishing that something we currently trade has stopped working.

New strategies are tested out of sample and priced with the full cost of trading them, including expected slippage, financing, and the margin the position will consume through a stress event rather than under ordinary conditions. Performance is attributed to specific, identifiable sources, so that a profitable period can be distinguished from a fortunate one. Where returns cannot be traced to the mechanism that was supposed to produce them, we treat the strategy as unexplained rather than validated.

Strategies are retired when the source of return no longer compensates for the risk carried to obtain it. That decision is made on the evidence, not on reluctance to abandon work that took a long time to build.

Execution and operations

Our infrastructure is built in house. That includes the pricing and surface fitting libraries, the risk system that aggregates exposure across books in real time, and the execution layer that manages orders into the venues we trade on. Building rather than buying keeps the assumptions inside the models visible to the people relying on them, and it means the risk system reflects the way this firm thinks about exposure rather than the way a vendor has chosen to categorize it.

Post-trade processing, reconciliation, and margin management receive the same attention as the trading itself. Operational failure has closed more derivatives businesses than incorrect market views, and the firm is organized on that understanding.