Welcome to the third and final part of this chapter.
Thank you all for the 100s of comments and upvotes - maybe this post will take us above 1,000 for this topic!
Keep any feedback or questions coming in the replies below.
Before you read this note, please start with Part I and then Part II so it hangs together and makes sense.
Part III

• Squeezes and other risks
• Market positioning
• Bet correlation
• Crap trades, timeouts and monthly limits

Squeezes and other risks

We are going to cover three common risks that traders face: events; squeezes, asymmetric bets.

Events

Economic releases can cause large short-term volatility. The most famous is Non Farm Payrolls, which is the most widely watched measure of US employment levels and affects the price of many instruments.On an NFP announcement currencies like EURUSD might jump (or drop) 100 pips no problem.
This is fine and there are trading strategies that one may employ around this but the key thing is to be aware of these releases.You can find economic calendars all over the internet - including on this site - and you need only check if there are any major releases each day or week.
For example, if you are trading off some intraday chart and scalping a few pips here and there it would be highly sensible to go into a known data release flat as it is pure coin-toss and not the reason for your trading. It only takes five minutes each day to plan for the day ahead so do not get caught out by this. Many retail traders get stopped out on such events when price volatility is at its peak.

Squeezes

Short squeezes bring a lot of danger and perhaps some opportunity.
The story of VW and Porsche is the best short squeeze ever. Throughout these articles we've used FX examples wherever possible but in this one instance the concept (which is also highly relevant in FX) is best illustrated with an historical lesson from a different asset class.
A short squeeze is when a participant ends up in a short position they are forced to cover. Especially when the rest of the market knows that this participant can be bullied into stopping out at terrible levels, provided the market can briefly drive the price into their pain zone.
​
There's a reason for the car, don't worry
Hedge funds had been shorting VW stock. However the amount of VW stock available to buy in the open market was actually quite limited. The local government owned a chunk and Porsche itself had bought and locked away around 30%. Neither of these would sell to the hedge-funds so a good amount of the stock was un-buyable at any price.
If you sell or short a stock you must be prepared to buy it back to go flat at some point.
To cut a long story short, Porsche bought a lot of call options on VW stock. These options gave them the right to purchase VW stock from banks at slightly above market price.
Eventually the banks who had sold these options realised there was no VW stock to go out and buy since the German government wouldn’t sell its allocation and Porsche wouldn’t either. If Porsche called in the options the banks were in trouble.
Porsche called in the options which forced the shorts to buy stock - at whatever price they could get it.
The price squeezed higher as those that were short got massively squeezed and stopped out. For one brief moment in 2008, VW was the world’s most valuable company. Shorts were burned hard.
​
Incredible event
Porsche apparently made $11.5 billion on the trade. The BBC described Porsche as “a hedge fund with a carmaker attached.”
If this all seems exotic then know that the same thing happens in FX all the time. If everyone in the market is talking about a key level in EURUSD being 1.2050 then you can bet the market will try to push through 1.2050 just to take out any short stops at that level. Whether it then rallies higher or fails and trades back lower is a different matter entirely.
This brings us on to the matter of crowded trades. We will look at positioning in more detail in the next section. Crowded trades are dangerous for PNL. If everyone believes EURUSD is going down and has already sold EURUSD then you run the risk of a short squeeze.
For additional selling to take place you need a very good reason for people to add to their position whereas a move in the other direction could force mass buying to cover their shorts.
A trading mentor when I worked at the investment bank once advised me:
Always think about which move would cause the maximum people the maximum pain. That move is precisely what you should be watching out for at all times.

Asymmetric losses

Also known as picking up pennies in front of a steamroller. This risk has caught out many a retail trader. Sometimes it is referred to as a "negative skew" strategy.
Ideally what you are looking for is asymmetric risk trade set-ups: that is where the downside is clearly defined and smaller than the upside. What you want to avoid is the opposite.
A famous example of this going wrong was the Swiss National Bank de-peg in 2012.
The Swiss National Bank had said they would defend the price of EURCHF so that it did not go below 1.2. Many people believed it could never go below 1.2 due to this. Many retail traders therefore opted for a strategy that some describe as ‘picking up pennies in front of a steam-roller’.
They would would buy EURCHF above the peg level and hope for a tiny rally of several pips before selling them back and keep doing this repeatedly. Often they were highly leveraged at 100:1 so that they could amplify the profit of the tiny 5-10 pip rally.
Then this happened.
​
Something that changed FX markets forever
The SNB suddenly did the unthinkable. They stopped defending the price. CHF jumped and so EURCHF (the number of CHF per 1 EUR) dropped to new lows very fast. Clearly, this trade had horrific risk : reward asymmetry: you risked 30% to make 0.05%.
Other strategies like naively selling options have the same result. You win a small amount of money each day and then spectacularly blow up at some point down the line.

Market positioning

We have talked about short squeezes. But how do you know what the market position is? And should you care?
Let’s start with the first. You should definitely care.
Let’s imagine the entire market is exceptionally long EURUSD and positioning reaches extreme levels. This makes EURUSD very vulnerable.
To keep the price going higher EURUSD needs to attract fresh buy orders. If everyone is already long and has no room to add, what can incentivise people to keep buying? The news flow might be good. They may believe EURUSD goes higher. But they have already bought and have their maximum position on.
On the flip side, if there’s an unexpected event and EURUSD gaps lower you will have the entire market trying to exit the position at the same time. Like a herd of cows running through a single doorway. Messy.
We are going to look at this in more detail in a later chapter, where we discuss ‘carry’ trades. For now this TRYJPY chart might provide some idea of what a rush to the exits of a crowded position looks like.
​
A carry trade position clear-out in action
Knowing if the market is currently at extreme levels of long or short can therefore be helpful.
The CFTC makes available a weekly report, which details the overall positions of speculative traders “Non Commercial Traders” in some of the major futures products. This includes futures tied to deliverable FX pairs such as EURUSD as well as products such as gold. The report is called “CFTC Commitments of Traders” ("COT").
This is a great benchmark. It is far more representative of the overall market than the proprietary ones offered by retail brokers as it covers a far larger cross-section of the institutional market.
Generally market participants will not pay a lot of attention to commercial hedgers, which are also detailed in the report. This data is worth tracking but these folks are simply hedging real-world transactions rather than speculating so their activity is far less revealing and far more noisy.
You can find the data online for free and download it directly here.
​
Raw format is kinda hard to work with
​
However, many websites will chart this for you free of charge and you may find it more convenient to look at it that way. Just google “CFTC positioning charts”.
​
But you can easily get visualisations
You can visually spot extreme positioning. It is extremely powerful.
Bear in mind the reports come out Friday afternoon US time and the report is a snapshot up to the prior Tuesday. That means it is a lagged report - by the time it is released it is a few days out of date. For longer term trades where you hold positions for weeks this is of course still pretty helpful information.
As well as the absolute level (is the speculative market net long or short) you can also use this to pick up on changes in positioning.
For example if bad news comes out how much does the net short increase? If good news comes out, the market may remain net short but how much did they buy back?
A lot of traders ask themselves “Does the market have this trade on?” The positioning data is a good method for answering this. It provides a good finger on the pulse of the wider market sentiment and activity.
For example you might say: “There was lots of noise about the good employment numbers in the US. However, there wasn’t actually a lot of position change on the back of it. Maybe everyone who wants to buy already has. What would happen now if bad news came out?”
In general traders will be wary of entering a crowded position because it will be hard to attract additional buyers or sellers and there could be an aggressive exit.
If you want to enter a trade that is showing extreme levels of positioning you must think carefully about this dynamic.
​

Bet correlation

Retail traders often drastically underestimate how correlated their bets are.
Through bitter experience, I have learned that a mistake in position correlation is the root of some of the most serious problems in trading. If you have eight highly correlated positions, then you are really trading one position that is eight times as large.
Bruce Kovner of hedge fund, Caxton Associates
For example, if you are trading a bunch of pairs against the USD you will end up with a simply huge USD exposure. A single USD-trigger can ruin all your bets. Your ideal scenario — and it isn’t always possible — would be to have a highly diversified portfolio of bets that do not move in tandem.
Look at this chart. Inverted USD index (DXY) is green. AUDUSD is orange. EURUSD is blue.
​
Chart from TradingView
So the whole thing is just one big USD trade! If you are long AUDUSD, long EURUSD, and short DXY you have three anti USD bets that are all likely to work or fail together.
The more diversified your portfolio of bets are, the more risk you can take on each.
There’s a really good video, explaining the benefits of diversification from Ray Dalio.
A systematic fund with access to an investable universe of 10,000 instruments has more opportunity to make a better risk-adjusted return than a trader who only focuses on three symbols. Diversification really is the closest thing to a free lunch in finance.
But let’s be pragmatic and realistic. Human retail traders don’t have capacity to run even one hundred bets at a time. More realistic would be an average of 2-3 trades on simultaneously. So what can be done?
For example:

• You might diversify across time horizons by having a mix of short-term and long-term trades.
• You might diversify across asset classes - trading some FX but also crypto and equities.
• You might diversify your trade generation approach so you are not relying on the same indicators or drivers on each trade.
• You might diversify your exposure to the market regime by having some trades that assume a trend will continue (momentum) and some that assume we will be range-bound (carry).

And so on. Basically you want to scan your portfolio of trades and make sure you are not putting all your eggs in one basket. If some trades underperform others will perform - assuming the bets are not correlated - and that way you can ensure your overall portfolio takes less risk per unit of return.
The key thing is to start thinking about a portfolio of bets and what each new trade offers to your existing portfolio of risk. Will it diversify or amplify a current exposure?

Crap trades, timeouts and monthly limits

One common mistake is to get bored and restless and put on crap trades. This just means trades in which you have low conviction.
It is perfectly fine not to trade. If you feel like you do not understand the market at a particular point, simply choose not to trade.
Flat is a position.
Do not waste your bullets on rubbish trades. Only enter a trade when you have carefully considered it from all angles and feel good about the risk. This will make it far easier to hold onto the trade if it moves against you at any point. You actually believe in it.
Equally, you need to set monthly limits. A standard limit might be a 10% account balance stop per month. At that point you close all your positions immediately and stop trading till next month.
​
Be strict with yourself and walk away
Let’s assume you started the year with $100k and made 5% in January so enter Feb with $105k balance. Your stop is therefore 10% of $105k or $10.5k . If your account balance dips to $94.5k ($105k-$10.5k) then you stop yourself out and don’t resume trading till March the first.
Having monthly calendar breaks is nice for another reason. Say you made a load of money in January. You don’t want to start February feeling you are up 5% or it is too tempting to avoid trading all month and protect the existing win. Each month and each year should feel like a clean slate and an independent period.
Everyone has trading slumps. It is perfectly normal. It will definitely happen to you at some stage. The trick is to take a break and refocus. Conserve your capital by not trading a lot whilst you are on a losing streak. This period will be much harder for you emotionally and you’ll end up making suboptimal decisions. An enforced break will help you see the bigger picture.
Put in place a process before you start trading and then it’ll be easy to follow and will feel much less emotional. Remember: the market doesn’t care if you win or lose, it is nothing personal.
When your head has cooled and you feel calm you return the next month and begin the task of building back your account balance.

That's a wrap on risk management

Thanks for taking time to read this three-part chapter on risk management. I hope you enjoyed it. Do comment in the replies if you have any questions or feedback.
Remember: the most important part of trading is not making money. It is not losing money. Always start with that principle. I hope these three notes have provided some food for thought on how you might approach risk management and are of practical use to you when trading. Avoiding mistakes is not a sexy tagline but it is an effective and reliable way to improve results.
Next up I will be writing about an exciting topic I think many traders should look at rather differently: news trading. Please follow on here to receive notifications and the broad outline is below.
News Trading Part I

• Introduction
• Why use the economic calendar
• Reading the economic calendar
• Knowing what's priced in
• Surveys
• Interest rates
• First order thinking vs second order thinking

News Trading Part II

• Preparing for quantitative and qualitative releases
• Data surprise index
• Using recent events to predict future reactions
• Buy the rumour, sell the fact
• The mysterious 'position trim' effect
• Reversals
• Some key FX releases

***
​
Disclaimer:This content is not investment advice and you should not place any reliance on it. The views expressed are the author's own and should not be attributed to any other person, including their employer.

Thanks for all the upvotes and comments on the previous pieces:

• Risk management parts 1,2,3
• News trading part 1

From the first half of the news trading note we learned some ways to estimate what is priced in by the market. We learned that we are trading any gap in market expectations rather than the result itself. A good result when the market expected a fantastic result is disappointing! We also looked at second order thinking. After all that, I hope the reaction of prices to events is starting to make more sense to you.
​
Before you understand the core concepts of pricing in and second order thinking, price reactions to events can seem mystifying at times
We'll add one thought-provoking quote. Keynes (that rare economist who also managed institutional money) offered this analogy. He compared selecting investments to a beauty contest in which newspaper readers would write in with their votes and win a prize if their votes most closely matched the six most popularly selected women across all readers:
It is not a case of choosing those (faces) which, to the best of one’s judgment, are really the prettiest, nor even those which average opinions genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be.
Trading is no different. You are trying to anticipate how other traders will react to news and how that will move prices. Perhaps you disagree with their reaction. Still, if you can anticipate what it will be you would be sensible to act upon it. Don't forget: meanwhile they are also trying to anticipate what you and everyone else will do.
​
Part II

• Preparing for quantitative and qualitative releases
• Data surprise index
• Using recent events to predict future reactions
• Buy the rumour, sell the fact
• The trimming position effect
• Reversals
• Some key FX releases

​

Preparing for quantitative and qualitative releases

The majority of releases are quantitative. All that means is there’s some number. Like unemployment figures or GDP.
Historic results provide interesting context. We are looking below the Australian unemployment rate which is released monthly. If you plot it out a few years back you can spot a clear trend, which got massively reversed. Knowing this trend gives you additional information when the figure is released. In the same way prices can trend so do economic data.
​
A great resource that's totally free to use
This makes sense: if for example things are getting steadily better in the economy you’d expect to see unemployment steadily going down.
Knowing the trend and how much noise there is in the data gives you an informational edge over lazy traders.
For example, when we see the spike above 6% on the above you’d instantly know it was crazy and a huge trading opportunity since a) the fluctuations month on month are normally tiny and b) it is a huge reversal of the long-term trend.
Would all the other AUDUSD traders know and react proportionately? If not and yet they still trade, their laziness may be an opportunity for more informed traders to make some money.
Tradingeconomics.com offers really high quality analysis. You can see all the major indicators for each country. Clicking them brings up their history as well as an explanation of what they show.
For example, here’s German Consumer Confidence.
​
Helpful context
There are also qualitative events. Normally these are speeches by Central Bankers.
There are whole blogs dedicated to closely reading such texts and looking for subtle changes in direction or opinion on the economy. Stuff like how often does the phrase "in a good place" come up when the Chair of the Fed speaks. It is pretty dry stuff. Yet these are leading indicators of how each member may vote to set interest rates. Ed Yardeni is the go-to guy on central banks.

Data surprise index

The other thing you might look at is something investment banks produce for their customers. A data surprise index. I am not sure if these are available in retail land - there's no reason they shouldn't be but the economic calendars online are very basic.
You’ll remember we talked about data not being good or bad of itself but good or bad relative to what was expected. These indices measure this difference.
If results are consistently better than analysts expect then you’ll see a positive number. If they are consistently worse than analysts expect a negative number. You can see they tend to swing from positive to negative.
​
Mean reversion at its best! Data surprise indices measure how much better or worse data came in vs forecast
There are many theories for this but in general people consider that analysts herd around the consensus. They are scared to be outliers and look ‘wrong’ or ‘stupid’ so they instead place estimates close to the pack of their peers.
When economic conditions change they may therefore be slow to update. When they are wrong consistently - say too bearish - they eventually flip the other way and become too bullish.
These charts can be interesting to give you an idea of how the recent data releases have been versus market expectations. You may try to spot the turning points in macroeconomic data that drive long term currency prices and trends.
​

Using recent events to predict future reactions

The market reaction function is the most important thing on an economic calendar in many ways. It means: what will happen to the price if the data is better or worse than the market expects?
That seems easy to answer but it is not.
Consider the example of consumer confidence we had earlier.

• Many times the market will shrug and ignore it.
• But when the economic recovery is predicated on a strong consumer it may move markets a lot.

Or consider the S&P index of US stocks (Wall Street).

• If you get good economic data that beats analyst estimates surely it should go up? Well, sometimes that is certainly the case.
• But good economic data might result in the US Central Bank raising interest rates. Raising interest rates will generally make the stock market go down!

So better than expected data could make the S&P go up (“the economy is great”) or down (“the Fed is more likely to raise rates”). It depends. The market can interpret the same data totally differently at different times.
One clue is to look at what happened to the price of risk assets at the last event.
For example, let’s say we looked at unemployment and it came in a lot worse than forecast last month. What happened to the S&P back then?
​
2% drop last time on a 'worse than expected' number ... so it it is 'better than expected' best guess is we rally 2% higher
So this tells us that - at least for our most recent event - the S&P moved 2% lower on a far worse than expected number. This gives us some guidance as to what it might do next time and the direction. Bad number = lower S&P. For a huge surprise 2% is the size of move we’d expect.
Again - this is a real limitation of online calendars. They should show next to the historic results (expected/actual) the reaction of various instruments.

Buy the rumour, sell the fact

A final example of an unpredictable reaction relates to the old rule of ‘Buy the rumour, sell the fact.’ This captures the tendency for markets to anticipate events and then reverse when they occur.
​
Buy the rumour, sell the fact
In short: people take profit and close their positions when what they expected to happen is confirmed.
So we have to decide which driver is most important to the market at any point in time. You obviously cannot ask every participant. The best way to do it is to look at what happened recently. Look at the price action during recent releases and you will get a feel for how much the market moves and in which direction.

Trimming or taking off positions

One thing to note is that events sometimes give smart participants information about positioning. This is because many traders take off or reduce positions ahead of big news events for risk management purposes.
Imagine we see GBPUSD rises in the hour before GDP release. That probably indicates the market is short and has taken off / flattened its positions.
​
The price action before an event can tell you about speculative positioning
If GDP is merely in line with expectations those same people are likely to add back their positions. They avoided a potential banana skin. This is why sometimes the market moves on an event that seemingly was bang on consensus.
But you have learned something. The speculative market is short and may prove vulnerable to a squeeze.

Two kinds of reversals

Fairly often you’ll see the market move in one direction on a release then turn around and go the other way.
These are known as reversals. Traders will often ‘fade’ a move, meaning bet against it and expect it to reverse.

Logical reversals

Sometimes this happens when the data looks good at first glance but the details don’t support it.
For example, say the headline is very bullish on German manufacturing numbers but then a minute later it becomes clear the company who releases the data has changed methodology or believes the number is driven by a one-off event. Or maybe the headline number is positive but buried in the detail there is a very negative revision to previous numbers.
Fading the initial spike is one way to trade news. Try looking at what the price action is one minute after the event and thirty minutes afterwards on historic releases.

Crazy reversals

​
Some reversals don't make sense
Sometimes a reversal happens for seemingly no fundamental reason. Say you get clearly positive news that is better than anyone expects. There are no caveats to the positive number. Yet the price briefly spikes up and then falls hard. What on earth?
This is a pure supply and demand thing. Even on bullish news the market cannot sustain a rally. The market is telling you it wants to sell this asset. Try not to get in its way.

Some key releases

As we have already discussed, different releases are important at different times. However, we’ll look at some consistently important ones in this final section.

Interest rates decisions

These can sometimes be unscheduled. However, normally the decisions are announced monthly. The exact process varies for each central bank. Typically there’s a headline decision e.g. maintain 0.75% rate.
You may also see “minutes” of the meeting in which the decision was reached and a vote tally e.g. 7 for maintain, 2 for lower rates. These are always top-tier data releases and have capacity to move the currency a lot.
A hawkish central bank (higher rates) will tend to move a currency higher whilst a dovish central bank (lower rates) will tend to move a currency lower.
A central banker speaking is always a big event

Non farm payrolls

These are released once per month. This is another top-tier release that will move all USD pairs as well as equities.
There are three numbers:

• The headline number of jobs created (bigger is better)
• The unemployment rate (smaller is better)
• Average hourly earnings (depends)

Bear in mind these headline numbers are often off by around 75,000. If a report comes in +/- 25,000 of the forecast, that is probably a non event.
In general a positive response should move the USD higher but check recent price action.
Other countries each have their own unemployment data releases but this is the single most important release.

Surveys

There are various types of surveys: consumer confidence; house price expectations; purchasing managers index etc.
Each one basically asks a group of people if they expect to make more purchases or activity in their area of expertise to rise. There are so many we won’t go into each one here.
A really useful tool is the tradingeconomics.com economic indicators for each country. You can see all the major indicators and an explanation of each plus the historic results.

GDP

Gross Domestic Product is another big release. It is a measure of how much a country’s economy is growing.
In general the market focuses more on ‘advance’ GDP forecasts more than ‘final’ numbers, which are often released at the same time.
This is because the final figures are accurate but by the time they come around the market has already seen all the inputs. The advance figure tends to be less accurate but incorporates new information that the market may not have known before the release.
In general a strong GDP number is good for the domestic currency.

Inflation

Countries tend to release measures of inflation (increase in prices) each month. These releases are important mainly because they may influence the future decisions of the central bank, when setting the interest rate.
See the FX fundamentals section for more details.

Industrial data

Things like factory orders or or inventory levels. These can provide a leading indicator of the strength of the economy.
These numbers can be extremely volatile. This is because a one-off large order can drive the numbers well outside usual levels.
Pay careful attention to previous releases so you have a sense of how noisy each release is and what kind of moves might be expected.

Comments

Often there is really good stuff in the comments/replies. Check out 'squitstoomuch' for some excellent observations on why some news sources are noisy but early (think: Twitter, ZeroHedge). The Softbank story is a good recent example: was in ZeroHedge a day before the FT but the market moved on the FT. Also an interesting comment on mistakes, which definitely happen on breaking news, and can cause massive reversals.
​