How to turn a volatility reading into a position size

Two different things wear the same word

Realised volatility is a measurement of what has already happened. Take the sequence of past returns, measure how widely they were dispersed, and you have a number describing how turbulent the recent past was. It is backward looking and factual.

Implied volatility is extracted from option prices and describes what the market is currently charging for exposure to future movement. If traders expect a turbulent period, options become more expensive, and the implied figure rises. It is forward looking and reflects an expectation rather than a fact.

Both are usually quoted as annualised figures, which is a convention that confuses almost everyone when they first meet it. An annualised volatility number does not mean gold is expected to move that much over the next year in some simple sense. It is a standardised unit, chosen so that different instruments and different periods can be compared, and it has to be rescaled before it means anything about a single day. Doing that rescaling is the step that turns an abstract number into something you can use when placing an order.

How a volatility index is constructed

A gold volatility index works by reading option prices rather than price history. Options across a range of strikes and a target horizon are observed, and the figure is derived by asking what single volatility assumption would be consistent with what those options cost. The result is a market implied expectation of movement magnitude over a standard forward window.

Three features follow from the construction and all three are commonly missed. The index measures expected magnitude only and contains no directional content whatsoever, because an option buyer paying for protection against a fall and one paying for participation in a rally both push the same number up. It is anchored to a fixed forward horizon, so it is rolled continuously and is not a statement about any particular date. And it is derived from a specific options market, in practice options on a large listed gold fund rather than on spot metal, so it inherits the quirks of that market including its trading hours.

That last point means the index can be stale precisely when you most want it, since it stops updating when the market it is built from is closed.

Turning an annual figure into a daily expectation

Volatility does not scale with time, it scales with the square root of time. This single piece of arithmetic is the bridge between a quoted index and a stop distance.

To convert an annualised figure into a one day expectation, divide it by the square root of the number of trading sessions in a year. To extend a daily figure out to several sessions, multiply by the square root of the number of sessions rather than by the number itself. The consequence is important: expected range grows, but it grows progressively more slowly, so holding a position for four sessions does not expose you to four times the daily range. It exposes you to roughly double.

This is why a cone rather than a straight wedge is the right mental picture for how far price may wander from here. It also explains why short horizon traders are not automatically safer than longer horizon ones. The range shrinks with the square root while costs and spreads do not shrink at all, which is a large part of why frequent trading is harder than it looks.

Realised volatility and its practical cousin

Most traders never touch an implied volatility series and do not need to, because a simpler measurement of recent realised movement serves the same purpose at the desk. An average of recent session ranges gives you a direct answer to the only question that matters for sizing: how far does this market currently move in a session.

The mechanics of using that for stop placement are covered in range based stops on gold, and the principle is worth restating here. A stop should be placed where the idea is wrong, and then the position should be sized so that distance is affordable. Placing the stop where the loss feels tolerable, which is what most people do, means the stop sits inside the market's normal noise and gets taken out by movement that proves nothing.

The practical difference between a range measure and an implied index is their timing. The range measure tells you what has been happening and adapts with a lag. The implied figure tells you what the options market currently expects, which can shift before realised movement does, particularly ahead of a scheduled event.

The gap between expected and actual movement

Implied volatility tends to sit above subsequently realised volatility, and the mechanism is insurance rather than inefficiency. Option sellers take on open ended risk and will not do it at a fair price, so they charge a margin. Buyers accept the margin because they want protection or convexity. The resulting premium is a feature of a functioning market, not an anomaly waiting to be harvested.

What makes the gap worth watching is how it behaves. When expected volatility is well above what gold has actually been doing, the market is paying up for protection against something it anticipates, which usually means a scheduled event or a visible risk. When expected volatility sits at or below recent realised movement, options are cheap relative to how the market has actually been behaving, which often happens after a sharp move when the event causing it has passed.

Neither reading is a signal. A high implied figure is not a prediction that a big move will happen, and a low one is not a promise of calm. Both describe pricing, and pricing is frequently wrong in both directions.

Sizing from volatility instead of from hope

This is where the whole subject earns its place. The procedure is mechanical and it removes a decision you should not be making emotionally.

• Decide what fraction of the account a single loss may cost. This is fixed in advance and does not change with how confident you feel.
• Place the stop where the idea would be proven wrong by structure, measured in the market's current range units rather than in money.
• Divide the permitted loss by that distance to get the position size.

The consequence is that size falls automatically when volatility rises and rises when it falls, so the amount you can lose on a trade stays constant while the market changes around you. That is the opposite of what happens when size is fixed by habit, where an unchanged lot size means your real risk doubles in a turbulent week without you deciding anything. The broader framework sits in risk management on gold, and volatility is the input that makes it adaptive rather than static.

Regimes, and the moves a volatility average cannot contain

Volatility arrives in regimes rather than evenly. Quiet periods cluster, turbulent periods cluster, and transitions between them are abrupt. A thirty day expectation is an average over a window, so it describes a regime while that regime lasts and is badly wrong across a transition.

Scheduled events are the clearest example of what an average cannot capture. A single release can produce a move larger than several ordinary sessions combined, within minutes, with spreads widening at exactly the moment the move happens. No average of recent ranges contains that, because the shape of the distribution is different. Stops sized for a normal session are not stops in those conditions, they are suggestions.

The honest conclusion is that volatility measures tell you how to size for ordinary conditions and tell you nothing about extraordinary ones. Recognising which you are in is a judgement, and the case for standing aside rather than sizing down is made in when to sit out. Sometimes the right position size is zero, and no formula will suggest it for you.

Where a volatility reading misleads

Four limitations are worth carrying around.

It is directionless by construction, so any attempt to read a high reading as bearish or a low reading as bullish is importing a belief the number does not contain. It says nothing about path, so two weeks with identical volatility can be a smooth trend and a violent range, which require completely different trade management. It is derived from a specific options market with specific trading hours, so it goes stale outside those hours and can be unresponsive exactly when gold is moving on overnight news. And in thin conditions the option prices feeding it are themselves wide and unreliable, which makes the derived figure noisier than its smooth appearance suggests.

Used within those limits it remains one of the most honest numbers available to a trader, because it is explicit about measuring magnitude and makes no claim about direction. Let it set your size and your stop distance, check how the current session compares with recent range on the live chart, and keep direction as a separate question answered by structure.