A simple way to determine the price trend is to plot a single moving average (MA) on the chart.
If the price is above the moving average, then it indicates that it is in an UPTREND.
If the price is below the moving average, then it indicates that it is in a DOWNTREND.
A popular trading strategy involves using two moving averages: a “faster” moving average and a “slower” moving average.
In a previous post, I discussed the bullish crossover.
This is when the “faster” moving average crosses ABOVE the “slower” moving average.
In this post, it’s time to look at the bearish crossover.
This is when the “faster” moving average crosses BELOW the “slower” moving average.
In the chart above, I used a blue circle to highlight where the 10 SMA (orange line) crossed below the 50 SMA (purple line).
When a bearish crossover occurs, it is considered a SELL signal.
Because when a bearish crossover occurs, the price is expected to continue to fall!But does it? Does the price actually fall afterward?
Do bearish moving average crossovers work?
Can it predict future price action, where price trends downward?
We already know from our recent analysis, that bullish MA crossovers don’t have much (if any) predictive power.
Let’s see if the bearish MA crossover is a different story.
Just to be clear, the point of this analysis is to answer the question:
“When the 10 SMA crosses below the 50 SMA, does price continue to fall?”
The reason we’re interested in finding the answer is that we have an observation and if our observation can be supported with evidence, then we have something to try and build an actual trading system with.
An observation can be described simply as, “If X happens, then the price does Y.”
Our observation is this:
“If a bearish MA crossover occurs, the price will continue to fall, so we should sell (or go short).”
We need to know if a bearish crossover can predict future price action (continued downtrend).
Because if it doesn’t, then there’s no point trying to optimize further like adding different exit strategies (e.g. stop losses and profit targets).
We want to first prove the observation (or phenomenon) is real, and then we can spend time building an optimized trading system around it, complete with position sizing, stop losses, profit targets, etc.
But first things first.If we want to know, “If a bearish MA crossover occurs, then the price will fall.”, then a simple way to find out is when it happens, open a short position, and hold it.
We’re not sure how long the price will continue to fall, which is why we will hold it for different lengths of time.
Here’s what I found…
Since January 1, 2010, there were 45 instances of the bearish crossover.
This means a total of 45 sell signals occurred.
Let’s take a look at the data…
The x-axis represents the different holding times (how long you keep a position open).
The y-axis represents the returns (whether the price ended lower (or higher).
Unlike the results from the bullish crossover study, which were ALL negative, the bearish crossover study showed some positive average returns.
For example, going short and holding for 1 day resulted in a 0.06% return. This means when the trade was exited, on average, the price had fallen by 0.06%.
Going short and holding for 20 days resulted in a 0.24% return. This means when the trade was exited, on average, the price had fallen by 0.24%.
Holding for 90 days showed the worst result, with an average return of -1.09%.
The chart above shows what percentage of trades ended up with a gain (or loss) at the end of the holding period.
It answers the question:
“If I go long and hold the position for X number of days, what percentage of my trades ends up with a gain or a loss?”For example, if you held each trade for 30 days, 61.36% of trades ended up with a gain (or positive return), while 38.64% of the trades ended up with a loss (or negative return).
Most holding periods had a win rate of over 50%. It wasn’t until the trades were held longer for 60 or 90 days did the losing trades outnumber the winning trades.
Average Gains vs. Average Losses
The chart above shows the average gain per winning trade and the average loss per losing trade.
It answers the question:
“When my trade ends up a winner, what is the average gain? And when my trade ends up a loser, what is the average loss?”
For example, if you held a trade for 1 day and then exited, if you won, your average gain was 69 pips. And if you lost, your average loss was 52 pips.
If you held the trade for 144 days and then exited, if you won, your average gain was 460 pips. But if you lost, your average loss was 118 pips.
Except for holding periods of 10, 30, and 90 days, the average gains were higher than average losses.
If we were theoretically going to trade this as a system, since we know the win rate and the average gain and loss, we can determine the system’s expectancy.
Expectancy is the average amount you expect to gain or lose per trade based on previous performance.
The chart above shows what can expect to win (or lose) per trade depending on how long you hold the trade.
It answers the question:
“What do I expect to earn (or lose) per trade in the long run?”
For example, if you held a trade for 20 days and then exited, you can expect to gain 31.12 pips per trade.
If you held a trade for 90 days and then exited, you can expect to lose 162.04 pips per trade. You definitely don’t want to hold trades this long.
As you can see, holding trades for 1, 3, 5, 20, or 30 days had positive expectancy.
A positive expectancy means that when you average out all the wins and losses, you make money.
A negative expectancy means that when you average out all the wins and losses, you lose money.
When the 10 SMA crosses below the 50 SMA, and you go short, it looks like price does continue to fall enough to make some money.
The data shows that after a bearish crossover, going short and holding it for 5 days or 20 days shows the most promise.
Holding it for 90 days is definitely NOT a good idea as it looks like the price seems to reverse trend direction by then.
Can we make it better?
Here’s an idea…
What if instead of taking every bearish crossover signal, we could “filter” out the losing signals which would improve the results?
In my previous post, I attempted to use the 200 SMA as a filter for bullish crossovers.
Let’s do the same thing here.
In the chart above, notice how the bearish MA crossover (10 SMA crosses below 50 SMA) occurs ABOVE the 200 SMA (blue line).
What if we skipped these trades and only traded when the bearish MA crossover occurred BELOW the 200 SMA?
This is how the filter works:
When the 10 SMA crosses below the 50 SMA, if the 10 SMA and 50 SMA are both BELOW the 200 SMA, then SELL.
If not, then stay out.
Basically, you only go short when the bearish crossover occurs BELOW (or underneath) the 200 SMA.
Why use the 200 SMA as a filter?
The 200 SMA is used to determine the long-term trend direction.
- If price and the faster-moving averages (like 10 SMA and 50 SMA) are ABOVE the 200 SMA, price is considered to be in an uptrend.
- If price and the faster-moving averages (like 10 SMA and 50 SMA) are BELOW the 200 SMA, price is considered to be in a downtrend.
By selling (“going short”) ONLY when both 10 and 50 SMAs are below the 200 SMA, we are “trading with the trend”.
We are (supposedly) following the path of least resistance. We are “going with the flow”. We are selling in a downtrend.
We don’t want to sell when the price is in an uptrend. Sounds ludicrous. We want to sell when the prevailing long-term trend is down.
Sounds great on paper. But let’s see if adding the 200 SMA as a filter made a difference.
After running the report, there were 29 instances when a bullish crossover occurred AND both MAs were above the 200 SMA.
Remember, without the filter, there were 45 instances of a bullish crossover.
So the 200 SMA filter removed 16 trades.
Did the removal of these trades help or hurt?
Average Returns (No Filter vs. Filter)
Uhhh, this so-called “filter” must be defective since it made average returns worse across the board.
Win Rate (No Filter vs. Filter)
The filter continues to disappoint. It decreased the percentage of winning trades.
Take a look at the 30-day holding period. Without the filter, 61.36% of trades were profitable, but with the filter, it dropped to 38.64%.
That drop is almost as bad as Golden State Warriors’ winning percentage this past season.
Average Gains vs. Average Losses (No Filter vs. Filter)
Except for the 1-day holding period, the filter did manage to reduce the average loss.
That said, it also reduced the average gain!
So while you lost less money per losing trade, you also made less money per winning trade. Bleh.
Expectancy (No Filter vs. Filter)
Except for the 20-holding period, the filter put every holding period into negative territory!
Check out how if you started to hold trades for 30 days or longer, how bad the results were.
The filter seems pretty worthless. 👎
When a bearish crossover (10 SMA crosses below 50 SMA) occurs, and you enter a short position ONLY if both the 10 and 50 SMA are below the 200 SMA, this does not improve the forecasting power of future price direction.
The data shows you’re better off simply trading the bearish crossover WITHOUT the filter.