Many of my posts are marathon length, but this one will be short and to the point.  I want to teach you a useful heuristic about risk.  The issue is, for any given trade you put on, how much risk should you be willing to take.  The risk you take on a given position is related to two things: the size of the position you take (in shares, contracts, etc.) and the amount you’re willing to let the position move against you before exiting via a stop order.  Generally the amount you’ll let the position move against you is determined by your trading strategy, so position size is the variable you want to adjust to keep your risk under control.  The rule of thumb has three parts:

• If a strategy has not been proven profitable via the trading system design process, set your max loss at 0% (ie. don’t bet).
• If the strategy is profitable, size your positions such that your maximum loss in the absence of slippage is no more than 2% of your risk capital.  In a near-worst-case slippage scenario, it should be no worse than 5%.
• If you have multiple correlated positions, take that into account.  For example, if you’re long equities and long oil (about 0.8 correlated on the 1-day time frame – nearly the same thing) you should set your positions so you only lose 2% if both positions move against you.

This rule generally only works well for relatively short term trading.  Longer term trading you almost always have to take bigger risk to get decent returns, which is part of why I advocate shorter time frames.

The effect of the 2% rule is that you can have a large number of losing bets and still stay in the game.  If you trade long enough, even with well tested systems, you will eventually have a prolonged losing streak due purely to bad luck.  It also helps control psychological issues associated with large draw-downs in risk capital (“I’ve lost more than half my money – what am I going to do now?”).

It can be hard to implement this rule in some derivative instruments where the contract size is very large.  If you can’t stick to a 2% max loss without placing unreasonably close stop orders, that’s a sign you’re too under-capitalized to trade that contract.

It turns out you can get a slightly better answer to the position sizing question in complicated cases by using a high octane calculus-heavy variant on the Kelly Criterion.  But the 2% rule is good enough for most situations and well worth memorizing.