Welcome, everyone, to this new video, where we’ll take a detailed look at what happens when we implement a short strategy on equity markets, such as stock indexes or individual stocks.
Why are we focusing on this topic today? Because it tends to cause some confusion.
First of all, we may be unsure whether to include a short component in our trading, since, as we’ll see shortly, indexes tend to rise over the long term. And this is something we’ve seen especially clearly in recent years.
Also, even when we do implement a short component in our strategy, it may not be properly optimized – or, more accurately, it may not be ideal for the characteristics and nature of this market.
Shorting the S&P 500 with an Unfiltered Mean-Reversion Strategy
For today’s video, I’ve chosen a very simple strategy with no filters applied. It’s a mean-reverting strategy that tends to buy on weakness…
For example, we can see a long trade here. Let me zoom in a little so you can see it better. We have an excellent long trade here, as well as several other excellent long trades during this phase, when the market was falling.
…And this strategy sells on strength. Here we can see several short trades during this particular phase.
I also want to point out that the period I’ve plotted here on the 5-minute chart is this same phase we saw in 2026. As you can see, the market made a major move higher and, more importantly, it never really gave us a break, in a sense. There was never any real pullback. It just kept rising, session after session.
So, as I mentioned earlier, I haven’t applied any filters to this strategy. It only has a stop loss, a take profit, and this mean-reversion trigger.
Long and Short Strategies on the S&P 500: Comparing Performance
At this point, let’s look at its performance, because I think this already gives us an interesting insight into how to use a short strategy and what we can learn from what happened here.
Looking at the overall performance first, this is the equity curve starting in 2020. So I’ve used a relatively short historical period, because the purpose of this video is mainly to analyze a specific case, rather than assess the strategy’s long-term profitability.
And as you can see, mean reversion is working on this instrument, at least with this particular strategy.
But what happens when we analyze the two components separately? Let’s look at the long side, for example. As you can see, it performed extremely well.
We can see that this strategy didn’t lose money in 2022, for example, even though 2022 was an especially difficult year for equity markets. I also want to point out that it was a bad year for mean-reverting strategies in general. Most mean-reverting strategies didn’t work in 2022.
This strategy, however, held up reasonably well and performed very, very well during the most recent period.
Now let’s see what happens when we analyze the same strategy on the short side. Here you can see the short side. Clearly, its performance is weaker.
We can see that the strategy still performed well in 2022, or at least generated some profits. But during the most recent period, when the market rose fairly sharply, this short strategy consistently lost money.
And since I had this strategy running live during that exact period, I started wondering whether having a short component actually made sense. The market was rising so strongly that perhaps it was simply acting as a drag on performance.
And that was indeed the case, if we kept the strategy configured this way. Let’s take another look at its characteristics. Here we can see that it generated 630 trades starting in 2020, with 223 on the long side and 407 on the short side.
So there is clearly a major imbalance. There are many, many more short trades than long trades. This makes sense when we consider the strategy’s logic and the fact that no filters were applied.
For example, if we simply look at this chart, the strategy enters long when the market falls. As you can see, downward moves tend to be much less frequent than upward moves, so the strategy has far more opportunities to enter short.
This is the first thing that should make us stop and think. Does it make sense to use this kind of strategy on the short side, when it trades much more frequently than it does on the long side?
Short Strategies on Stock Indexes: When Do They Really Make Sense?
In other words, what am I really trying to capture with a strategy that trades short? To answer that question, we can look at this chart. This is simply the performance of the S&P 500, starting in 1995 in this case. But the exact starting date doesn’t really matter. Even if we went further back in time, the overall performance would still be quite positive. This is something that tends to rise over the long term.
Of course, there are difficult periods, such as the dot-com crash, the 2008 financial crisis, and the declines we see here in 2020 and 2022, and so on.
But fundamentally, the edge we can exploit – and I would say the easiest edge to exploit in finance, because it’s part of the very nature of the markets – is buying something with an upward drift, such as stocks… not every stock, I want to emphasize… or stock indexes.
Based on this information, the first thing we understand is that it certainly makes sense to have long-only strategies. But it can also make sense to have strategies that trade short only under specific conditions.
Market Regime Filters for Short Strategies: Moving Averages and the VIX
Now let’s try to understand how we can identify those conditions. There are generally two methods. The first is what we might call a regime filter, meaning that we try to identify the market regime we’re currently in. The second, as we’ll see in a moment, is the Unger Method.
First, let’s understand what a regime filter is and what I mean by that term. A regime filter is simply a kind of indicator that tells us which market regime we’re currently in.
For example, we could determine whether the market is in a bullish or bearish phase, or whether we’re in a period of high or low volatility.
To give you a few examples, I’ve added a classic 200-period moving average here, along with the fear index, the VIX. Either of these could easily be used as a regime filter. With the 200-period moving average, for example, I could set up a strategy that trades short only when the market is in a bearish regime.
Let’s think about the 2008 crash, for example. I would have enabled short trades only during this particular phase, because I’m not interested in making money on the short side while the market is rising. I’m interested in having protection when the market falls.
Just for context, if I had manually applied this filter to the strategy we looked at earlier, it wouldn’t have traded short during this rally in 2026, because the market was clearly in a bullish phase.
Of course, this is just one idea. There are other possible regime filters. I’ve included another one here, which is the VIX. We know that the VIX is the fear index and that it is inversely correlated with equity markets. So when this index rises, it tells us that fear is increasing in the markets. That could potentially be a good time to stay in cash, meaning we avoid long positions, or possibly start trading on the short side.
I’m giving you a few examples here, but there’s plenty of room for creativity. If someone finds a particular approach, method, or indicator that can translate the concept of a regime filter into rules that a computer can use, that’s perfectly valid.
Since the VIX is an index that can theoretically range from zero to infinity, for example, we could trade only when the VIX is above a specific threshold. Let’s say the VIX is above 30. I would enable short trading only during that phase. Or perhaps I could require the VIX to be above its 200-period moving average. That could work as well.
The main concept I’m trying to convey is that there are several ways to determine whether we’re in a market regime where it may be appropriate to use short strategies.
But I also want to make it clear that this isn’t the only method. We could call this the traditional approach, and it’s the one we commonly see in the literature. First, I identify the regime I’m in, and only after identifying the filter – sorry, only after identifying the regime – do I choose a particular trading approach. But this isn’t usually what happens with the Unger Method.
How to Develop a Short Component with the Unger Method™
With the Unger Method, we do something that is almost the opposite, in a sense. Here, for example, I’ve loaded a strategy from our database. What we try to do first is find an engine, a type of trigger. In the strategy I showed you earlier, for example, that trigger could be an overbought or oversold condition. I identify a rule that tells me when to buy, sell, or enter short.
Once I’ve found that rule and have a relatively large statistical sample, I’ll have plenty of trades to analyze. I might find a strategy that performs extremely well on the long side, while the short side isn’t particularly profitable but at least doesn’t lose money. The short side may also work very well during periods of market weakness or panic. Only at that point do I apply specific filters that work well for that particular strategy.
So this is a very different concept from what we saw here in this workspace. In this case, I decide in advance when to use a short strategy. I create the setup first, such as the 200-period moving average, and only then do I enable short trading.
With the Unger Method, the exact opposite happens. First, I find a trigger that appears to work reasonably well on the short side too. Only afterward do I identify the best conditions for that short component to perform effectively.
What are the differences between these two approaches? The first difference is that, if we decide in advance which market regime to use, that regime might not be the optimal condition for that particular strategy.
If we consider this short strategy, for example, perhaps the strategy I’ve presented here – which sells when the market is overbought – might not be the best choice when the market is in a panic phase. Think about the 2008 crash, for example. There may be other conditions in which this strategy can work, such as when the market is showing strength. That isn’t what we saw in this particular case, but it could happen with other strategies.
Another advantage of the Unger Method, compared with using a regime filter that we define in advance, is that we start with a much larger statistical sample than the one we see here.
For example, if I only allow short trades when the market is below its 200-period moving average, the initial number of trades will clearly be relatively small compared with saying, “Okay, I’ll test short trades under all market conditions. Then, using a much larger statistical sample, I’ll decide under which conditions the short side should actually be enabled.”
Using Short Strategies on the Nasdaq as a Portfolio Hedge
At this point, let’s also take a look at the strategy I’ve loaded here, which trades the Nasdaq. As I mentioned earlier, it’s part of our database, and once again I’ve loaded the most recent years.
As you can see, it’s a very good strategy overall, but it has one particularly interesting characteristic compared with the strategy we saw at the beginning of the video.
If I look at the long side, which is essentially the component that performs very, very well, you can see that the overall equity curve is almost identical to the long-side equity curve. If I then look at the short side, the performance is a little more uneven, if you’ll allow me to use that term. However, even in this case, when I plot the strategy starting in 2020, I can see that it provides a degree of protection during the periods when I need it.
For example, we can see what happened here in 2022, when the market fell. As expected, this strategy made more money during that period than it did during phases when the market was rising, such as the period we’ve seen over the last year.
Once again, let’s look at the Total Trade Analysis. What’s interesting is that the strategy generated 1,457 trades: 1,000 on the long side and only 433 on the short side. So, as you can see, the imbalance here favors the long side. There are far more long trades than short trades.
And this, in my opinion, is the key concept behind using a strategy on the short side. Fundamentally, there’s nothing wrong with using only long strategies on instruments like these, where the upward trend is a long-term tendency.
But if I also wanted to include a short component, I would first identify the phases in which that component works. And since the short side is specifically intended to provide protection, it isn’t what I rely on to benefit from the historical price series. It isn’t what allows me to exploit the edge that actually exists in a market that rises over the long term. So, by applying filters, the short component should be limited relative to the core strategy.
We’ve looked at two possible methods. The first is to apply a regime filter from the outset. The second is to proceed later, after identifying the trigger, as we do with the Unger Method™. We could use patterns or other conditions to identify the best circumstances for enabling a short component.
So, should we eliminate short strategies entirely from our portfolio? As always, it depends partly on the individual trader’s characteristics. But in my opinion, having a short component that can provide protection during periods when the market may perform poorly is definitely not a bad idea.
With that said, let me know in the comments which method you use. Do you prefer to have only long strategies, or do you prefer to define a regime filter in advance?
That’s all for today, and I’ll see you in the next video.
