Session overlaps and where liquidity concentrates

The hours when two regions are working at once 

A session boundary is not a switch. London does not go dark at the moment New York arrives, and Tokyo does not empty the instant London logs on. Each session begins while the previous one is still staffed, still quoting and still taking orders from its own region's clients. For a stretch of hours at each of those seams the market is the sum of two regions rather than one, and that stretch is what an overlap is.

Why the sum matters comes back to how a price is made. Prices exist because institutions are willing to post them: a price at which each will buy, a price at which each will sell, and a size each is willing to do at those prices. The best price available to the market at any instant is not any one institution's price. It is the best of all of them taken together. Adding participants enlarges the pool of postings, so the best of that pool improves. Removing participants does the reverse, and does it just as fast. An overlap is nothing more exotic than the hours when the pool holds two regions' postings instead of one.

The two overlaps, and their hours in Gulf Standard Time 

Of the three sessions covered so far, two seams produce an overlap of any consequence. Tokyo's afternoon runs into London's morning, briefly. London's afternoon runs into New York's morning, for much longer. The second is the larger one by every measure that can be counted: the number of institutions active, the volume reported through them, and the number of scheduled economic releases that land inside it, because the euro area and the United States both publish their data in their own mornings, and their own mornings sit inside that window.

The conversion is worth doing once rather than memorising. Gulf Standard Time sits four hours ahead of UTC and holds that offset through the whole year, so all of the seasonal drift belongs to the other side of the calculation. The consequence for a reader in the Gulf is that the largest overlap of the day falls in the evening rather than the working afternoon, which the next lesson takes up properly.

Why the quoted market is usually at its narrowest here 

The difference between the best price to buy and the best price to sell is the spread, and in a market made by many institutions that difference is composite. It is assembled from the highest bid anyone is showing and the lowest offer anyone is showing, and those two need not come from the same institution. That is the entire reason more participants produce a narrower quoted market: the composite takes the best half of each posting and discards the rest.

Depth is the second half of the same picture and the half that a quoted price hides. Behind the best offer sits a finite quantity, and behind that a further quantity at the next price, and so on down the book. A market with many participants is usually deeper at every level and not only tighter at the top one, because each additional institution contributes size as well as a price. Narrowness and depth tend to move together for that reason, and they are both consequences of the same count of participants rather than two separate properties of a good hour.

The same window is also where the range is widest 

The institutions that narrow the quoted market are not there to quote. They are there because their clients are transacting, and the same hours that concentrate the postings concentrate the orders. Scheduled releases land inside those hours by design, since a statistical agency publishes in its own morning. The distance between the highest and lowest price reached over a period is the range, and the range of an overlap hour is typically among the widest of the day for exactly the reason the spread is among the narrowest. One count of participants produces both.

Slippage is where that meets execution. A market order is an instruction to deal at whatever is available when it arrives, not at the price displayed when it was sent, and the gap between those two moments is where a fast market does its work. Two separate things can happen inside it. Postings can be withdrawn or repriced before the order lands, which is what happens in the seconds around a release. Or the order can simply be larger than the quantity posted at the best price, in which case the remainder deals at the next prices down the book.

The two effects are therefore not in tension and do not need reconciling. Depth is what narrows the composite quote, and depth is what gets consumed when a large number of participants act on the same information at the same instant. An overlap supplies more of it and asks more of it in the same hours.

The hours when the book is thin 

The trough sits between the New York close and the Tokyo open, with a shallower version of it in the hours before London arrives. Quiet is easy to read as calm, and structurally it is closer to the opposite. Fewer institutions are posting, so the composite is assembled from a smaller pool and the difference between the best bid and the best offer widens. Less size sits at each level, so an order of a given size reaches further down the book and deals at a worse average than the same order would in a busy hour. A single institutional transaction that an overlap would absorb without a mark can move a thin market visibly.

The cost, margin and execution mechanics from the earlier modules all come due here at once, and none of them is suspended because the hour is quiet. The spread is a cost paid on entry whatever the clock says, and it is wider in these hours. Slippage is measured against the depth that exists at the moment of arrival, and there is less of it. An adverse move is calculated on the full contract value rather than on the margin posted against it in a thin hour exactly as in a busy one, so a loss is not limited to the amount deposited at any hour of the day.

Where practitioners disagree 

Whether an overlap is the cheaper window is genuinely contested, and the two sides are measuring different things. One tradition treats the quoted spread as the cost, which makes the overlap the narrowest hours of the day by definition. Another argues that the cost that lands on a statement is the realised one, including slippage and including the wider range that the same hours produce, and that an order dealt in the seconds around a release can give back on execution more than it saved on the quote. Both are describing something real. The disagreement survives because the two descriptions diverge with order size: an order small enough to sit inside the quantity at the best price experiences mostly the spread, while an order large enough to walk the book experiences mostly the depth.

The count of sessions is contested too, which matters more than it sounds. Some accounts add a Sydney or Wellington session and describe four, some split continental Europe from London, and the hours quoted for each differ between sources by an hour or more in either direction. None of these is wrong, because no exchange opens or closes the currency market and no bell marks a boundary. The session names are inherited from an era of trading floors and they describe where the people are, not a rule the market enforces. They are a useful map of participation and a poor map of anything else, and an instrument written on an exchange traded underlying ignores them entirely in favour of its own published calendar.