Risk-Reward

Apples and Unicorns; The 'Risk/Reward' Ratio ...is a LIE (unless you're an unusual kind of trader).

Reward to Risk Ratio: So you think the greater the risk, the greater the reward huh!?

A lot of people like to talk about the risk/reward ratio. A lot of those folks seem to be salesman-brokers; I have a somewhat cynical secret opinion of why that is, but I will instead advance the notion that they are simply assimilated into the hive when they become brokers, and this is part of the Industrythink that they are taught. Heaven forbid you think all brokers do this, by the way, but I do seem to hear about the risk/reward ratio a lot from them, directly or indirectly.

Oh, and before I go on, there is one category of folks who will get beet-faced at this article and send me email about what a moron I am, so let me soothe them first. Those of you who are employing a complete program of statistical risk management where you are actually measuring historical risks and rewards and applying it forward to new trades... and this is how you establish and liquidate your positions... I'm not talking about you. Go find another essay to read. You are the only folks for whom risk/reward means anything other than another round-turn sale today. Oops! Darn, my little cynical thought just slipped out after all.

Okay, let's get on with this. The "risk/reward ratio" is more or less defined as taking the expected risk on a trade and comparing it to the expected return. If the risk is sufficiently overshadowed by the return, you take the trade. Sounds pretty decent, no? You look at Wheat, decide that you should go long, but that if you're wrong that it'll cost you £250. If you're right, though, you should make £1,000. Profit potential is four times risk... great trade! Let's rock!

Stop. You're comparing apples and unicorns. Not apples and oranges; oranges exist in reality.

How do we define "risk"? Most people do it by where they set their stops; that's what I call "normal risk", because while you're not guaranteed to get your stop price, and occasionally you get a hideous exit from a stop, normally you get out near your price. People pre-define risk depending on what their strategy is, and it doesn't really matter. You can set your stop on the other side of a trend or support line, or moving average or chart pattern, or set your stop based strictly on a dollar value. Ultimately, you've got a loss goal in mind for the trade... at least, you'd BETTER. If you don't have a predefined loss point when you enter a position, you are begging for a frying pan in the face. So that's your apple; risk is easy, though when applied in this ratio it's secretly dangerous to your trading health, and I'll say more about that later.

On to a more fun subject: Reward. To calculate reward, you just... that is, you... hmmm. Okay, lots of people forecast and predict rewards, I'll give you that. They've got retracements or price targets or calculations that double or halve the size of past formations or moves, or they flip and invert the last series of price bars. You name it, folks use it to predict price levels. Do some people sometimes get those prices right? Sure. Do MOST people USUALLY get those reward forecasts correct? Get real; this is futures trading we're talking about. You may predict a price that you think the market will attain, but you don't really have any assurance that it'll get there.

Think about it. You are defining normal maximum risk for a trade by placing (or intending to place) a stop at a specific price. If you are wrong, you have defined the maximum that you are likely to lose on the trade, more or less. If you are going to compare apples to apples, you must compare your defined maximum risk to a reward that is also maximally defined. Most traders don't do that (and if you do, you're reading this like I told you not to). Most traders want to let their profits run, right? So that ratio is comparing apples to oranges, except that while you explicitly define risk based on an adverse move, the profit target cannot be so explicitly defined. It's a guesstimate, a guideline, a rule of thumb, not a line on a chart. It's a unicorn: something we're all familiar with, recognize, agree upon, and can't find in the real world to save our lives.

A lot goes on in a trade to change it once it begins. Maybe the market goes way up, but not quite to your target. You gonna let that entire big gain evaporate because it didn't hit some arbitrary price that you picked? I sure hope not. What if the market just goes sideways for three weeks? Are you going to roll the contract to another month because it's hit neither your profit nor your loss target? Maybe, but there's times to do that and times to look elsewhere for a trade. What if it breaks out above your profit target? Will you exit at the target price, even though there's no end in sight for the market? Sure, but only if you always trade that way, and only if you have some statistical information on how often you'll be right and wrong. The point is that if the market doesn't take out your loss stop, all kinds of new decisions can be made about the amount of profit, above or below your "reward" unicorn... sorry, target, but the one thing that NEVER changes for a trade is the maximum you'll allow yourself to risk.

The risk/reward ratio is a lie. It is a fun fantasy, but because one of its components has a horn, it does not exist. If a salesman calls you up (I won't even call him or her a broker) and tells you he has a dynamite trade in Wheat, as evidenced by the risk/reward ratio of 8 to 1 profit to risk, just say no. It is NOT real. Take that trade only if there is something else out there that's convincing.

Sit down, I'm not done yet.

Now let's talk about that dangerous part I mentioned. Let's say you really could forecast a profit target that was real. If I tell you I've got a trade for you that has a five to one profit to risk ratio, would you take it? Sure, sounds great, let's go. Well, you lose, but at least your risk was controlled, right? What if you had a £5,000 account, and you traded a T-Bond contract where the expected profit was £5,000? Ow! That's a 20% loss on your account in a single trade. A couple more of those and you're done. So, you DIDN'T control your risk because you were comparing your per-trade risk against a potential reward, not against what you were really risking which was the money in your account. You've got apples and unicorns in multiple dimensions, now.

You want to measure risk? Compare apples. Use what I call the risk/account ratio. Allow your signal generator, whatever that might be, to pick your entry points. You should have a good reason for getting into every trade, and "If the market goes up 3 points I make £3,000," is NOT a good enough reason by itself. Forget about reward. Once you find a good trade, see how much it will cost you if you're wrong. Use the same stop-setting guidelines as when you began this essay; your determination of risk is just as valid as it's ever been.

Don't compare the risk/reward ratio (i.e. don't take the expected risk on a trade and comparing it to the expected return) - use the risk/account ratio instead (i.e. the amount you're risking to the amount in your account).

But now, compare the amount you're risking to the amount in your account. If the risk would take too big of a chunk out of your account, walk away. DO NOT TRADE, even if the signal is a good one, because you will hurt yourself if you are wrong. You are comparing risk to account size; the number of dollars you have now vs. the number of dollars after you lose. Apples to apples. Risk/Account. What's the correct ratio? Lots of folks say 2%, so that if you have a £10,000 account you'd allow yourself to lose £200 on a trade. I agree with that, but I've found that 2-4% feels pretty safe to me, depending on the circumstances. Some people don't get the willies risking 10% of their account at a time, though I think that's excessive. It all depends on your personal tolerance for risk.

What if you're trapped by this rule, and you can't find anything to trade that gives you conservative risk management? My opinion about that is almost always that your account size is too small, not that your risk is too great. You're absolutely right that you can't take a £2,000 account and trade it with a full-service broker and hope to ever have a small risk/account ratio. The commission alone exceeds the ratio. Get a bigger account, and if you can't do that, then wait until you can. A trader who is right so often that the account size doesn't matter exists only in the land of the unicorns.