[dc]S[/dc]ome of the most persistent myths about trading revolve around the idea of low-risk trades. You will often hear traders say that they took a trade because the risk was “only a few pennies” or they could use a “very tight stop” on the trade. Other times, you will hear traders say that they were justified taking a trade because it offered a good enough “risk / reward” profile. In all of these cases, traders are ignoring some important mathematical realities.
One of the challenges of trading well is learning to think about probabilities. We do not care about the outcome of any one event, or any one trade; the only thing that matters is what happens over a large sample of trades. Rather than focus on the fact that you can buy with a .02 stop, the question you need to ask is “what happens if I do the same trade over and over?” We naturally tend to focus on the outcome of one specific event, but the ability to think in large sample sizes is one of the keys to building intuition about probabilities. This is not a natural way to think; it must be actively cultivated and developed.
Expected value (or expectancy) is a tool that can help. Perhaps the best way to think of it is that it answers what the average result of doing something over a large number of trials would be. Mathematically:
Expected value = payoff if the event happens * probability of the event happening
Both parts of this equation, the payoff and the probability, are essential. One without the other is meaningless. For instance, the payoff from getting hit by an asteroid is very bad (certain annihilation), but the probability of that event happening is vanishingly small. Therefore, most of us don’t spend a lot of time thinking about this. What about not picking up the bread your spouse asked you to get on your way home? This is a scenario that carries a very high probability of a bad outcome, but the consequences are not dire: the overall “expected value” of that scenario isn’t that bad.
We can also use this concept to find the “fair value” of an asset in many situations. What should you be willing to pay for a ticket in a raffle where 1,000 tickets are sold for a $10,000 pot? Your chances of winning are 1/1000, so multiply that by the $10,000 payout to get the expected value of the ticket. Each ticket a fair value of $10; if you are able to buy a ticket for less than that, you are playing a positive expectancy game.
You now know that the “real risk” (expected value) of a stop is determined by this formula:
real risk = loss if stop is hit * probability of stop being hit.
It is important to train yourself to evaluate the risk in any trade according to this formula, rather than focusing on the just the size of the possible loss. It is meaningless to evaluate the reward / risk profile of a trade without also considering the associated probabilities. A tighter stop will always have a higher chance (probability) of being hit compared to a wider stop–a very tight stop might be a near-certain loss. Even though the size of each individual loss is very small, over a large sample size they add up to a very significant risk. (In trading, it is possible to bleed to death from a thousand paper cuts. I know hundreds of traders who have done just that.) Limiting your risk means having a stop and respecting that stop with perfect discipline. Good risk management means using the proper stop, not necessarily a very tight stop.
For some of you, you have seen this math before, but take a moment to consider its significance anew. For others of you, this will be new. You are lucky, because, if you can build this concept into your thinking, you can avoid many of the mistakes beginning traders struggle with. If you are going to trade well, you need to internalize this concept. Good gamblers do it. Good traders do it, and you must too if you’re going to make it in this business.
I’ll dig a little deeper into this concept in a future blog post. (I also have neglected the issue of position sizing in this post, but will address that in the near future as well.)
Even if you have statistics for a set-up, it is hard to grasp the probability for a trade working. My way to put a SL is not fixed. It is used to falsify my hypothesis. And if I get stopped out my SL pays me twice: It protects my account and it provides me with some valuable piece of information. It tells me my hypothesis has not worked and I have to frame the context a different way.
Cheers,
Markus
hi Markus,
Thanks for your comment. I agree very much with your idea that the stop must go at the point the trade is no longer invalid, and I also think there can be multiple ways to do that. However, I would argue that I DO know the probabilities for my trades working, at least over a large sample size. I am consistent in where I place my stops and targets, and I know the probability and risk/reward in my setups… and I’ve seen that it is stable over time and in different markets.
Also, with most of my trades I don’t get a second chance. I’m usually dealing with markets at inflection points, so if I’m stopped out of a trade I usually don’t get a chance to try the trade again in a different way.
I think it comes down to knowing your setups, your patterns and always being consistent in how you trade them. I guess I’d encourage you to rethink “Even if you have statistics for a set-up, it is hard to grasp the probability for a trade working.” What if your statistics included the setup, the stop and the target as a unit? Just something to consider maybe.
Thanks for your response, Adam.
Just for clarification: If one defines a set-up with strict rules and trades them for a long time one has great database and you can say what the probability of the trade was (sic!). The probability may change in the future because it is dynamic.
In my case I do not trade set-ups per se but have my main focus lies on the context and I plan for different scenarios. Simple example: If a support line breaks and I expect a certain level to be tested I try to ge in on a pullback but only before the level has been tested and only if I see buying stalling on the tape.
I guess that you have also context rules for your set-ups.
Cheers,
Markus
If your goal is consistency, remove degrees of freedom.
Thanks, good advice.
I appreciate very much what you do for the community.
Cheers,
Markus
Selecting an appropriate stop has been a challenge for me. I have had the habit of justifying a trade with a tight stop, often stopping out more than once in the same setup, and offer my experience to support Adam’s observation of bleeding to death from a thousand cuts.
I’ve also used tight stops to justify a position to satisfy a fear of missing, which hasn’t worked for me on sum.
Thanks for your comment. I’ll do a post on where to place stops (though there’s obviously not one correct answer) in the near future.
I believe that a stop is the price of knowing whether one’s hypothesis about the movement of the market is correct. Depending upon timeframe of chart being traded, and the type of order used for entry – stop sizes may vary dramatically.
For example: A limit order entry may require a wider stop but provide a greater excursion than a Stop Market order being used to enter with the momentum of the market. Tighter stops used in context with rapid market movement can allow larger size with similar risk as the exteced market movement either moves rapidly in your favour or you get out. (Right in or right out)
Having said this – I take and agree with your point that it is possible to bleed your account to death with the higher transaction costs and frequency of stops being hit inevitable with tighter stops.
Any automated strategy that I have programmed which shows gains over time ultimately requires a larger stop than I like to use on discretionary entries. Perhaps there’s something to learn in that.
Thanks for you blog ideas.
>Any automated strategy that I have programmed which shows gains over time ultimately requires a larger stop than I like to use on discretionary entries. Perhaps there’s something to learn in that.
I think you hit on a great truth there. My experience has been the same.
Pingback: Monday links: deficits and rates | Abnormal Returns