Glossary

Craig is the author and narrator of these entries. He works on an institutional trading desk — Bay Street, Canada's financial centre — and has done so for the better part of two decades. He covers small and mid-cap equities, which means he spends a lot of time thinking about liquidity, microstructure, and the gap between what a company is worth and what the market says it's worth on any given day.

He started writing partly out of frustration: the financial commentary he found online either talked down to readers or assumed they already knew everything. He wanted something in between. A letter to a reasonably smart friend who has never stared at a Bloomberg terminal at 7:45 in the morning.

He uses a pseudonym here. The views are his own. They do not represent his employer, his clients, or anyone's position, long or short.

A recurring presence in these entries, referred to only as "my desk head" or "the first desk head." An unnamed mentor figure from early in Craig's career whose observations Craig has a habit of quoting, often years after the fact.

What little we know: he ran a trading desk, he had a gift for reducing complicated ideas to single sentences, and he was right about things more often than the market had any right to require of a person.

An axe is a dealer's private directional bias — a desire to buy or sell a particular security that goes beyond simply providing a two-sided market. A dealer with an axe to sell has inventory they want to offload; a dealer with an axe to buy is looking to accumulate.

The term matters because dealers with axes are better counterparties for the corresponding side of a trade. If you want to sell and a dealer has an axe to buy, you're likely to get better execution — tighter spread, better price, larger size — than you would from a dealer who is simply warehousing risk neutrally.

Historically, knowing which dealer had axes on which names was considered proprietary market intelligence. The MiFID II reforms and similar regulations have changed the landscape considerably, but the underlying concept — that your counterparty's private inventory position affects the quality of the market they make — remains as true as ever.

In the simplest sense, flow is just order activity — the buy and sell orders moving through a market at any given time. But "reading flow" means something more than that. It means understanding who is behind the activity, why they're doing it, and what it signals about where a name is likely to go.

Institutional flow and retail flow look different. Algorithmic flow looks different still. A trader who has spent years in a market develops intuitions about these signatures that are genuinely difficult to articulate — something like the way a doctor can read an X-ray not by following an algorithm but by pattern recognition accumulated over time.

Flow can be "natural" (a fund rebalancing, an index reconstitution, a large seller working out of a position) or "informed" (someone who knows something moving before the news breaks). Telling the difference is one of the harder parts of the job.

The bid-ask spread is the difference between the highest price a buyer is willing to pay (the bid) and the lowest price a seller will accept (the offer or ask). When you buy at the offer, you've immediately "crossed the spread" — your mark-to-market is the bid, which is lower. You start every trade with a small loss equal to half the spread.

In highly liquid markets — large-cap equities, major currency pairs, on-the-run government bonds — this cost is negligible, often fractions of a cent. In illiquid markets, the spread can be wide enough to represent a meaningful hurdle that the trade must overcome just to break even.

Beyond its cost function, the spread is also an information source. Dealers widen spreads when they're uncertain, when they've seen informed flow, or when they're carrying inventory they'd rather not. Craig's desk head put it simply: "The spread is the market telling you how hard it thinks this trade is going to be."