Week 11: Sentiment and positioning
Can't ignore it, but don't overhype it
Hi.
Welcome to Fifty Trades in Fifty Weeks!
This is Trade 11: Sentiment and positioning
Thank you for signing up!
50in50 uses the case study method to go through one real-time trade in detail, about once per week. This Substack is targeted at traders with 0 to 5 years of experience, but I hope that pros will find it valuable too. For a full description of what this is (and who I am), see here.
Listen to this as a podcast on the web
… or Spotify
… or Apple.
Update after ten trades
Taking profit on the XPO puts and the NASDAQ short. Full explanation below.
Only open trade remains the ASHR long.
Full results so far in the grid
After 10 trades, the results are not great, but not horrendous. Started 0-4 and 3-for-4 more recently. There is a lack of big winners so far and therefore the P&L is negative. My edge is in FX and it remains to be seen if it translates to single names, equity indexes, and commodities. If it doesn’t, that’s OK, the point is to teach tactics and methodologies and give you tools. Then, you apply those tools to markets where you have an edge. I still think it’s useful to collect and display the data, though, for intellectual honesty and also so you can see how to collect and organize trade-by-trade data.
50in50 results to date
Sentiment and positioning
First of all, I want to preface this whole thing by saying: People tend to overweight positioning. It is very often totally irrelevant. But then, once in a while, positioning and sentiment are the only things that matter. As such, the bar for me to care about or factor in positioning is very high, but at the megaextremes, I give it high or even dominant importance in my framework.
While I believe sentiment and positioning are important variables to consider in every trade, and can sometimes be the main driver of a trade idea, I also believe they are overrated as indicators by many traders. Here are the key points on sentiment and positioning:
Sentiment and positioning matter sometimes, but not most of the time.
80-90% of the time, sentiment just follows price so if you trade against sentiment and positioning you are often just fighting the trend.
At inflection points, a tuned-in feel for positioning and sentiment can be a huge source of edge.
Positioning is packed with potential energy, like fuel for a future fire. Positioning unwinds are like fires. They need heat (a spark or catalyst), fuel (large positions), and oxygen (more bad news, or mark-to-market losses that trigger feedback loops that lead to more position unwinds).
You need large positions, a catalyst, and mark-to-market losses all at once and then you get a chain reaction that continues until it finally burns itself out (or is put out by a central bank). When I was in grade school, they taught us this process with the fire triangle.
Large positions are like piles of dry tinder that can potentially burst into flame at any moment, but probably won’t. Remember, just because you have a large stack of dry wood next to an open can of gasoline… That does not mean there will be a fire.
Here is my basic positioning and sentiment framework:
In other words, going against sentiment works at the turning points but not during the trend. If you can identify a palpable turn in sentiment via observation of behavior or via data analysis, you can position ahead of the crowd.
Sources of sentiment and positioning data
The most commonly-used positioning data comes from the CFTC. This data shows the positioning of non-commercial speculators, and while it is closely followed and can be interesting information, it has no predictive power on its own. If you see “Copper longs on CFTC hit all-time high” your instinct is probably to go short copper but the inconvenient truth is that a setup like that is actually modestly bullish on all time horizons.
The CFTC data is slow-moving and trend-following and directional, not contrarian. Do not use CFTC data by itself as a contrarian indicator.
A source of sentiment data that is contrarian is the Daily Sentiment Index, produced by Jake Bernstein since the 1980s. That data is the output from a survey of retail traders that simply asks: "Are you bullish or bearish on X?" It shows what percentage of survey respondents are bulls. When bulls get below 10 or above 90, that is a good indicator of a near-term extreme. DSI data is available only via subscription.
Options markets often contain important positioning information. For example, when stocks rise and VIX rises at the same time, you know that something unusual is going on because the correlation is almost always inverted between those two. Stocks up / VIX up is a sign that the market is so bullish it is adding to upside with calls, instead of hedging the downside with puts. You can also look at the put/call ratio for a read on equity sentiment. In currency markets, people look at the risk reversal to see if traders have a strong preference for puts or calls, and how that preference is changing.
Positioning surveys can be a useful source of information. They are timely and fast moving. People change their minds quickly, as opposed to trend-following systems which move much more slowly as they wait for price confirmation. Again, with surveys, most of the time they are directional, and the only time they are contrarian is when they hit an extreme. In the positioning surveys I have done when working at banks, I found that when sentiment made a record high or low for the past year, that could often be a good sign of a turn.
In the stock market, there are many sentiment surveys and indicators available. I like the AAII survey (bulls vs. bears). CNN’s Fear and Greed indicator is also a nice simple snapshot of equity sentiment. As you become expert in your market, you will learn the best sentiment and positioning indicators to follow.
Finally, magazine covers can be the ultimate contrarian indicator. When a financial theme hits the cover of a mainstream magazine, always take note. By the time journalists believe a theme is worthy of a cover story, that theme is (by definition) well-known. That means the theme is probably also crowded. For more on the Magazine Cover Indicator, see here. Other anecdotal sentiment indicators were discussed in 50in50: Trade 7.
How to use sentiment and positioning data
"You can’t escape the madness of crowds by dogmatically rejecting them… The most contrarian thing of all is not to oppose the crowd but to think for yourself.”
- Peter Thiel, Zero to One
Sentiment and positioning data can be the difference maker at turning points. The number one takeaway on sentiment and positioning, whether you are using surveys, options markets, or other sources of information, is that only the super-mega-extremes are times when you should consider fading. Otherwise, positioning and sentiment are mostly just trend indicators and they move up and down with the price.
Positioning and sentiment are part of the picture, NOT THE WHOLE PICTURE.
“Sentiment” is not always synonymous with “positioning”
While they tend to mostly move in lockstep, it is important to understand the difference between sentiment and positioning. Many traders and analysts use the two terms interchangeably but they are not always the same thing.
Sentiment is how people view the market. If you take a poll of traders, you will get a sentiment reading. An example of this sort of reading is the Daily Sentiment Index (DSI) published by Jake Bernstein, which is a daily survey of retail futures traders.
Positioning is a metric that measures ACTUAL POSITIONS. An example of this type of data is the Commitment of Traders report, which comes out every Friday and summarizes futures positions held by commercial and non-commercial traders.
You can have a market that is extremely bearish, but not heavily positioned and this is different from a market that is extremely bearish and extremely short. The difference is that the skew of future moves is completely different in the two situations. The reason is simple: A market that is bearish but not yet positioned is primed to go down as people take new positions to reflect their bearish view. On the other hand, a market that is bearish but heavily positioned short is more at risk of a short squeeze or reversal because there could be nobody left to sell.
Why would a market see extreme sentiment but not extreme positioning? There are many possible reasons. For example:
1. The theme is new. It takes a while for positions in one direction to become extreme. Sentiment moves faster than positioning. Many traders will build into a position as their conviction rises and as new information substantiates their view. As a general rule, sentiment leads positioning because it is faster and easier for a trader to get bearish than it is for her to put on a sizeable bearish trade.
2. Major event risk upcoming. This is one of the most common and highly profitable setups. Let’s say the market is extremely bearish EURUSD but there is an ECB meeting. Nobody wants to go into the ECB meeting short EURUSD because there is gap risk or event risk if they are hawkish.
The moment the ECB statement comes out, the EUR might drop aggressively as traders put on the trade they wanted to have all along. Traders were just waiting for the event to get out of the way and did not really care about the outcome of the event. This can lead to surprising moves where you get, for example, a somewhat hawkish ECB meeting (which should be bullish EURUSD) and EURUSD drops aggressively because traders care less about the ECB and more about the fact that the event risk is out of the way and now it’s safer to get short EURUSD.
If you can identify moments where the market has a very strong view but does not have the position yet because they are waiting for an upcoming event to pass, you can often profit by simply putting on the position faster than others once the event is out of the way.
3. Time of year. A trader’s risk appetite varies depending on her profitability and the time of year. For example, as we get close to the end of the year, traders like to reduce risk and book profits to start the new year with a clean slate. Therefore, you might see a situation at the end of the year where the market is extremely bullish USDJPY but flat because of an unwillingness to commit. In this case, the turn of the year could trigger substantial USDJPY buying even though nothing macro has changed between December 31 and January 1.
4. Positioning clear outs. From time to time, crowded positions will get rinsed by sharp counter moves. The market might hit a point where traders cannot take any more pain and the majority cut their losses. While they still hold the macro view, they cannot afford to hold the position any longer and so they cut it. This is called a stop loss run.
.
A good example of this is the price action in interest rate and FX markets in Fall 2014. Throughout 2014, the market was heavily positioned for higher US interest rates and a stronger dollar. In mid-October, however, the market started to lose faith in the interest rate trade as global and US inflation fell.
This reached a tipping point on October 15th as a weak US Retail Sales release was the final straw and the market stopped out of most of its dollar and interest rate positions all at once. In the weeks following this positioning wipeout, the market came to the conclusion that while US rates were not set to rise, the dollar should continue to benefit as the ECB and BOJ embarked on quantitative easing (QE) programs.
This led to a situation a few weeks later where the market was very bullish USD (sentiment) but flat (positioning) as they simply could not bring themselves to get long USD again as they were scarred by the pain of October 15th. Once new catalysts pointed to further USD gains, the market went from extremely bullish and lightly positioned back to extremely bullish and extremely long USD and the dollar shot higher.
While sentiment and positioning move together, it is important to remember that they are not always the same thing. Most of the time sentiment and positioning will match fairly closely but it is the times when they do not match that can be most important.
Understand Positioning
In FX, there are many proprietary positioning indicators published by the major banks and there is the Commitment of Traders (COT) report from the Chicago Mercantile Exchange (CME).
There are many potential sampling issues with the COT data but it is still generally regarded as the best aggregate measure of FX positioning available. Have a look at how positioning and price move together:
Daily EURUSD (black) vs. COT NON-COMMERCIAL NET EUR POSITION (red)
A few observations here that apply broadly to positioning:
Positions and price normally trend in the same direction.
Positioning often leads price at turning points. If you look at the lows in mid-2012, for example, you can see that positioning (the red line) led the move off the lows in spot (the black line).
Positioning can trend for a long time. The market took more than a year to get from the highs in positioning (and price) to the lows between May 2011 and June 2012.
Absolute levels of positioning mean less than the rate of change of positioning. In May 2010, you can see that short EUR positioning reached what was then an extreme level of more than 100,000 contracts. This then led to a very aggressive reversal from 1.18 to 1.48 in EURUSD. The next time positioning reached short 100,000 contracts (early 2012), many traders took that as an indication that positioning again could go no further. Positions then went from short 100,000 to more than 200,000 contracts and EURUSD kept going down from 1.28 to 1.22 before it bottomed.
.
When you hear someone say that positioning is extreme, don’t get too excited. Usually, all that means is that the market is in a strong trend. Ask more questions: Can positioning get more extreme? How long has it been extreme? Is the price continuing to go in favor of those who are positioned with the trend (i.e., do price and positioning confirm each other?) Are the variables and narrative that led to the big position still in play? Or are the fundamentals changing? What upcoming catalysts could lead to a turn in positioning?
Finally, as a market maker, I have often noticed that you can get a good read on positioning by the number of inquiries you get when the market moves. You can get the same vibe on Twitter.
If people are generally asking “what the??” about a move, or saying the move is “stupid” or “dumb” or “ridiculous” you can assume they are NOT on the winning side of the move. Similarly, if you are at a hedge fund and you get 11 different “BUY USDCAD!” recommendations in your inbox; you should probably hit a bid. The deeper your involvement in the market and the greater your experience level, the better your sense of positioning and sentiment.
A QUICK MARKETING BLURB
You should sign up for am/FX. That’s the daily global macro note I have been writing since 2004. In there, I explain all my live trade ideas (lotta FX, but also other asset classes) and give you all the global macro in an easily-digestible two or three-page note.
If you like what you are reading in 50in50 so far, you will love am/FX. It also includes a spicy nonsequitur every day to stimulate your brain jelly. Thanks.
END OF MARKETING BLURB
Trade 11: Cover shorts and go long S&P 500 futures
Late last week, the American Association of Individual Investors (AAII) published its weekly sentiment survey. Here are two charts. The first one shows BULLS dropped to a 30-year low. The second one shows BULLS minus BEARS also at an extreme.
How do we know if this matters? The simplest way is to backtest the setup and see what the forward performance looked like other times readings were this extreme. So I did that. Here is the performance of (long US equities when BULLS minus BEARS is below -25). The current reading is -33.
P&L of long SPX if (AAII bulls minus bears < -25), four-week holding period
The win rate is 66% vs. 62% in all other 4-week periods, the sample size is 38 non-overlapping 4-week periods, and the median return is +2.9% vs. +1.2% in all other 4-week periods. The effect is strong and hard to ignore. If you take out 2008, the effect is gigantic. Today (April 18) is also tax day, and that can often be a low point for stocks in periods of heavy retail involvement because retail selling to fund tax payments dries up.
While the FOMC is coming in a few weeks and the overall macro setup still looks extremely suspect to me, I’m going to close my eyes, plug my nose, and flip from short NASDAQ to long S&P futures.
Seasonals work very well here too as stocks tend to rally after tax day (today) and into May. Every April and every May since 2017 have been bullish S&P. Furthermore, there is nothing major going on this week, event-wise, and bear markets tend to need fuel to keep going. “Never short a dull market” as they say. The stop loss is 4279 and the take profit is 4579.
We also cover the XPO puts (the stock was getting pretty close to the take profit anyway) and cover the NASDAQ short. The NASDAQ short and XPO puts were the best trades so far (NASDAQ short at 14970, out at 13908 XPO puts in at $2.60, out at $6.10).
I’m using TYPE II sizing here for the S&P long, risking 2% of capital ($2,000). I have not been varying my sizing much as my conviction is about equal most weeks. That is probably a function of doing a weekly piece instead of just trading whenever I have conviction. In real life, I vary my position size more.
That’s it for today! If you liked this 50in50, please do me a quick favor and hit the like button.
50in50 goes long ESM2 at 4385 with a stop loss at 4279 and take profit at 4579.
Risking 106 to make 194. The full trader framework is here. As always, I will monitor the performance and offer detailed updates as we progress.
Thanks for reading.
Trade at your own risk. Be smart. Have fun. Call your mom.
My global macro daily is here
My crypto substack is here
And this is my Twitter
DISCLAIMER: Nothing in “50 Trades in 50 Weeks” is investment advice. Do your own research and consult your personal financial advisor. I’m putting out free thoughts for people who want to learn. This is an educational Substack. Trade your own view!
This is the part I keep coming back to.
COT is useful, but it’s easy to misuse. Crowded does not mean a trade has to reverse. It just means the easy asymmetry may be gone, and price has to prove more.
Context, not timing.
Thanks Brent for the trip into Sentiment & Positioning. The trader framework link works for me, but there is only 1% risk instead of 2% as in the article. Ups ;)