Tick size and minimum increments

Prices do not move continuously. They move in fixed steps, and the size of the step is set by the venue rather than emerging from the market. That step is the tick size, and it is the reason a price lands on one permitted value or the next and never between them. The permitted values form a grid, and every quote, every resting order and every trade has to sit on it.

The grid is not a technical limitation. It is a deliberate design choice with a well understood trade off, and exchanges revise it periodically for exactly that reason. A finer grid lets a quote express a smaller difference and lets participants compete on price in smaller increments. A coarser grid forces participants onto fewer price levels, which concentrates the quantity resting at each of them.

The trade off a tick size makes 

A tick that is too coarse imposes a floor on the spread. Where the smallest permitted increment is wide, the best bid and best ask cannot be closer than one increment apart however competitive the market is, so participants pay a spread that reflects the grid rather than the risk of quoting. Instruments in this state show a spread that is almost always exactly one tick.

A tick that is too fine has the opposite problem. Competing participants can improve on each other by an economically meaningless amount, so quantity spreads itself across a large number of price levels and there is very little resting at any one of them. The top of book becomes narrow and shallow at once, which looks liquid on a quote board and behaves poorly for anything but the smallest orders.

Exchanges therefore set tick sizes with the aim of landing between the two, and revise them when an instrument's price moves far enough that the grid no longer suits it. Some venues apply a tiered schedule in which the tick size increases with the price of the instrument, so a share trading in single digits and one trading in the hundreds are not held to the same grid.

Three grids, not one 

Tick size is the most discussed of the three constraints an order has to satisfy, and it is not the one that most often causes a rejection. The other two are the volume step and the minimum volume, and together the three define the full set of orders a venue will accept.

• Tick size, the smallest permitted change in the price. It constrains where an order may rest and what prices may print.
• Volume step, the smallest permitted change in the quantity. An order has to be a whole multiple of it, which sets the resolution of any position sizing calculation.
• Minimum and maximum volume, the smallest and largest quantity the instrument accepts in one order. A quantity below the minimum is rejected, and one above the maximum has to be submitted as several orders.

The third grid, the one nobody names, is the product of the first two: the smallest change in the money value of a position is the tick size multiplied by the contract size multiplied by the volume step. That figure is the resolution of every risk calculation performed on the instrument, and it is why a calculation that produces a precise target has to resolve onto a coarser reality.

The last row is the practical content. A sizing calculation produces a real number and the instrument accepts only points on a grid, so the calculation's output has to be resolved to one of them. On an instrument with a large contract size and a coarse volume step, the gap between adjacent permitted sizes is itself a meaningful amount of exposure.

Orders that do not sit on the grid 

An order submitted at a price the tick grid does not permit has to be handled somehow, and venues differ. Some reject it outright with a message naming the constraint. Some round it to the nearest permitted level, which may be in either direction. Some round it in the conservative direction, away from the market for a limit order and towards it for a stop.

None of those behaviours is wrong and the difference matters, because a rounded stop level is not the level that was submitted and a rounded limit price is not the price that was calculated. The behaviour is documented per platform, and it is one of the few places where reading the documentation changes what an order actually does rather than merely explaining it.

Where the grid becomes visible 

In a depth display, the grid is the reason price levels are evenly spaced: the levels are the grid, and an empty level is a permitted price with nothing resting at it. In a chart at a fine timeframe, the grid is the reason candles on some instruments appear to move in visible steps rather than smoothly, which is not a rendering artefact but the actual set of prices that traded.

It also explains a class of apparent execution anomalies. An order that seems to have filled a fraction away from an obvious level frequently filled at the nearest permitted level to it, and a stop that appears not to have triggered at a level often had a trigger sitting one tick beyond where the chart implied. Both are the grid, and both are checkable against the instrument's published tick size.