Last Updated on July 16, 2022
How to measure trading risks? Risk should be measured based on the quality of the trading decision. A good quality trading decision has a high risk-adjusted expected outcome.
What are trading risks? Risk is a measure of potential permanent losses.
I use the term permanent as we are not referring to short term dips due to volatility.
Risk is a subjective and multifaceted topic.
Some measure risk based on historical volatility or drawdowns, others use hedging and portfolio diversification. These measures might be useful from a modeling perspective but in practice, they have their own shortcomings.
Moreover, I would argue that risks can’t even be measured after the fact. More on this later.
Let’s talk about the 3 ways to measure risks:
- Risk measurement by quality of trading decisions
- Risk measurement based on past performance
- Risk measurement based on hedging and portfolio diversification
Ways to measure trading risks
Risk measurement by quality of trading decisions
Risk should be measured based on the quality of the trading decision. A quality decision is one that gives the highest risk-adjusted expected outcome.
Expected Outcome = Probability of Success x Reward – Probability of Success x Loss.
An example of a good quality trading decision is to buy Bitcoin Cash at 3X leverage because you know that some puppet-master in the background is going to pump it.
An example of a low quality trading decision is to buy Bitcoin Cash at 3X leverage because “everyone thinks cryptocurrency is easy money and you FOMO-ed (fear-of-missing-out).”
Both decisions are the same, one is high risk, the other is not.
The quality of a trading decision can’t be evaluated by short term outcomes.Lucas Liew (Me!)
The outcome of one trade doesn’t indicate the quality of a trading decision.
Measuring risk via the quality of a trading decision has its shortcomings. The problem with this method is that it is subjective.
It is subjective because a high risk-adjusted expect value calculations involve inputting the probabilities of wins and losses, along with their corresponding profit or loss value. All these values are subjective and is up to the competence of the trader.
Despite the downsides, I believe this is the best way to measure risk in trading.
This is why some people can tolerate a 30% drawdown in their own portfolio but get uneasy when someone who is managing their money has a 30% drawdown.
Risk measurement based on past performance
It is not accurate to measure risk based on past performance, but most people do that as it is convenient for modeling and they have no other real way to measure risk in a standardised manner.
Risk can be measured using drawdowns, volatility of returns or some other fancy measure.
What is a drawdown?
A drawdown is the biggest % drop in your portfolio.
Generally you want a small drawdown as compared to your returns. The smaller the drawdown-returns ratio, the more likely you are risking less to make more.
If possible, you want to measure the drawdown of account equity (includes unrealised profit and losses) and not your account balance (does not include unrealised profit and losses).
Some traders have huge unrealised losses in their trades and managed to turn them around in the last moment. Since we want to know how much losses they are risking, account equity paints a more realistic picture.
What is volatility of returns?
It refers to how volatile your performance is, as opposed to how volatile the market is.
Volatility of returns can be measured using standard deviation of your returns.
Low volatility of returns tends to indicate that you are consistently making good and profitable decisions. It also tends to go hand in hand with low drawdowns, which indicates that you risk little to make more.
Problem with measuring risks using drawdowns and volatility of returns
High drawdowns and volatility of returns are not necessarily risky moves if the trader understands them and knows how to manage them.
Warren Buffett lost 50% of this portfolio between 2008 and 2009. I don’t think many people will think that he should be written off because of that.
Some traders use a strategy called martingale betting. It entails doubling your bets when you lose a trade.
This strategy usually produces consistent realised returns until… it blows up your account at one shot.
Here’s a video explanation of why Martingales don’t work: Martingale and Myths
Another problem with measuring risks using volatility of returns, is that different trading styles lead to different styles of returns.
For example, an alpha-based strategy like price arbitrage will tend to have very low volatility of returns.
A beta-based strategy that tracks innovative technology stocks will tend to have higher volatility of returns.
This doesn’t necessarily mean the latter strategy is riskier.
Problem with historical data
Drawdowns and returns are historical. The past doesn’t predict the future perfectly and these figures change over time.
This is especially dangerous if the figures were estimated from a specific period of time, this creates a selection bias.
If you were to measure a trader’s drawdown in the last 5 years (it is 2019) now, you might find that he has low drawdowns with high returns. But the last 5 years were decent years for stocks. You do not know how he will perform during crisis years.
This is why traders who traded well through both financial crises and good times are so sought after. For these traders, their historical drawdowns and return volatility are a more accurate measure of their risk management skill.
Risk measurement based on hedging and portfolio diversification
To trade is to take a risk. However, we do not want to take indiscriminate risks. We want targeted risk.
Hedging removes unwanted risks.
Let’s say we believe Stock X has great technology. We want to make a long bet on its tech.
But Stock X is more than just its tech, it includes operations, management, marketing, stock market influence and other factors. If we long Stock X, we will essentially be taking a bet on those other factors too.
The solution? Hedge away the unwanted factors (i.e. hedge away unwanted risk)
We find a stock that is similar to Stock X in every way except its tech. Let’s call this Stock Y.
And we short Stock Y.
The final trade is to long Stock X and short Stock Y. To put this in math form:
Stock X – Stock Y = (A+B+C) – (A+B-D) = C – D
Where A, B, C, D are factors and C and D represent the tech of Stocks X and Y respectively.
Hedging improves our risk-adjusted expected outcome.
We can use it loosely as a risk management tool by checking if the trader has hedged away unwanted risks at a cost-effective price.
The idea behind portfolio diversification is to produce the highest volatility-adjusted expected outcome.
Portfolio managers create a portfolio of assets to get a combination with historically high returns, low volatility and low correlation between assets.
Similarly, traders create a portfolio of different trading strategies to get a combination with historically high returns, low volatility and low correlation between assets.
The idea is that short term opposing moves between different assets/strategies cancel out leaving a long term upward trajectory.
2 issues here.
First, the managers use volatility to measure risks. In this case, we are referring to the volatility of the asset movement.
Similar to before, high volatility assets are not necessarily risky if the trader understands them and knows how to manage them.
Second, all figures (returns, volatility and correlation) are historical.
As mentioned, the past doesn’t predict the future perfectly.
For instance, if the correlation figures were taken from the last 10 years where the US stock market experienced a bull run with low interest rates, it will likely not hold in other market conditions.
Risks are not clear even in hindsight
Risks are not even clear in hindsight. What happened in the past is one of many possible outcomes.
For instance, we know that Lehman Brothers (a big US bank) went bankrupt in 2008.
Let’s assume that Trader X made a big short bet on Lehman Brothers before the crash and made a killing.
However, this doesn’t mean that he was right or that a collapse was the likely outcome.
Lehman Brothers could have been bailed out (I think they were actually close to one). If a few men in suits had decided differently in 2008 and bailed out Lehman Brothers, Trader X would have lost a lot.
If the chance of bankruptcy was low but it happened, betting big on a short trade was not the right trade. It would have been a risky trade, regardless of whether the bankruptcy happened or not.
Even in hindsight we aren’t sure if our probability and outcome estimates are accurate.
That said, hindsight is a sample point and we can learn from it. However, we need to view the past objectively with its many possibilities.
Humans are biased towards events that happened. We need to place emphasis on the many possible events that didn’t happen as well as the one that did.
What is the biggest risk in trading
The biggest risk in trading is not what you don’t know. It is believing strongly in something that just isn’t true.
A foolish trader is not aware that they are things he or she doesn’t know.
A prudent one knows that there are things he or she doesn’t know. He is aware of the known unknowns and the possibilities of unknown unknowns.
What are the different forms of risks? Here are many different forms of risk: the risk of losing money, the risk of falling short, the risk of missing opportunities, FOMO risk, credit risk, illiquidity risk, concentration risk, leverage risk, funding risk, manager risk, over-diversification risk, risk associated with volatility, basis risk, model risk, black swan risk, career risk, headline risk, event risk, fundamental risk, valuation risk, correlation risk, interest rate risk, purchasing power risk, and upside risk. I’m sure I’ve omitted some. Many times these risks are overlapping.
The above is a quote from Howard Marks of Oaktree Capital.
Many of the ideas here are inspired from the legendary investor Howard Marks from Oaktree Capital. You can read his take on risk here: “What Risk Really Means”