Both advanced scans and technical alerts use a special language (which we refer to as “scan syntax”) to communicate with the Scan Engine. Though the language takes some time to master, it provides a great deal of flexibility that will prove invaluable in creating scans that cater to your needs. In this article, we'll equip you with the basic building blocks you need to write effective scans.
A scan is a series of tests that are performed by the Scan Engine to see which securities meet all your technical requirements. Securities that meet your first requirement are passed along to see if they meet your second requirement, etc. until the last of the tests has been conducted. Those securities that passed all the tests are thus determined to meet all your technical requirements and are subsequently included in your scan results.
Scans are divided into scan clauses, each of which gives the Scan Engine instructions about how to test for a particular technical requirement.
An individual clause has two expressions, which are compared to each other using a comparison operator. The clause is enclosed in square brackets. Below is a very simple example of a clause:
[close > 5]
The two expressions in this clause are “close” (the closing value of the stock) and “5” (a static value). The comparison operator is the “greater than” symbol (“>”). What that means is that this clause is testing to see if the closing value of each stock is more than $5.
If the statement indicated by the clause is true for a stock, said stock passes the test. In this case, all stocks with a closing value of more than $5 would be passed through to the next test, while all those with a closing value less than or equal to $5 would be excluded from the scan results.
Learn More: Scan Syntax Reference
We've already seen the “greater than” comparison operator, but there are several other operators that can be used in your scan clauses.
Some are usually used for text comparisons:
Some are usually used for number comparisons:
Note that “=” and “is” can be used interchangeably in most clauses. The same is true for “!= ” and “is not” operators. The scans “[type=stock]” and “[type is stock]” will be interpreted the exact same way by the Scan Engine.
Use caution when comparing numeric values with the equals operator - you may not get the results you expect from your scan. Check out our Troubleshooting Equality Scans article to learn more.
The Crosses Above operator (“x”) is frequently used when testing for a technical signal. Essentially, it means that the expression on the left has just crossed above the expression on the right; yesterday it was below, but today it is above. Here's a very simple example:
[close x 5]
This clause is testing for stocks whose closing value crossed above $5 on the day it is run. To put it simply, this means that yesterday the stock's value was below $5, while today it is above $5.
Note: We do not have a “Crosses Below” operator, as you can easily accomplish the same thing by using the “Crosses Above” operator with the positions of the two expressions switched. Building on our example above, if you wanted to test for the closing price crossing below $5, you would use the following clause:
[5 x close]
If $5 has just crossed above the close, then that means that the close has just crossed below $5.
We've seen one example of an expression (close), but there are so many more expressions you can use in your scans. Browsing through the Scan Components dropdown menus in the Advanced Scan Workbench will help you get a good idea of the expressions that are available to you. Be sure to check out our Scan Syntax Reference for all the details on adding those expressions to your scans.
Expressions are grouped into these general categories in the Scan Components area:
Some expressions require configuration parameters. As an example, if you are using the RSI indicator, you'll need to tell the Scan Engine how many periods you want to use to calculate the RSI value. These parameters appear in parentheses immediately after the indicator name. For instance, the clause “[RSI(14) > 0]” states that you want tickers where the 14-period RSI value is greater than zero.
When there are multiple parameters for a single indicator, they are separated by commas (e.g. “[MACD Line(12,26,9) > 0]”).
Some parameters are optional and will use a default value if left out. Make sure the default value is what you expect it to be. For example, the simple moving average indicator has an optional parameter indicating what values it uses to calculate the SMA; the default value is the close. If you want it to use a different value, that must be specified.
// Wrong: compares volume to 5-day SMA of close [volume < SMA(5)] // Right: compares volume to 5-day SMA of volume [volume < SMA(5,volume)]
The Technical Indicators section of the Scan Syntax Reference explains what the parameters are for each indicator.
Remember, we only use parentheses to designate the parameters for indicators. Square brackets are used to show the start and end of scan clauses.
Thus far, we have only looked at clauses with very simple expressions. However, there are a number of ways in which you can modify your expressions to make them more specific.
By default, expressions refer to daily data, but it is relatively simple to change an expression to use weekly or monthly data instead. Simply add the word “weekly” or “monthly” to the front of the the expression. For example:
[weekly MACD Line(12,26,9) x 0] [monthly RSI(14) > 0] [daily close > weekly close]
Note that you can even mix and match date periods within a single clause, as in the third example above.
In addition to changing the date offset for the entire scan with the Starting Date Adjuster, you can also change the date offset for individual expressions by adding a date modifier at the beginning of the expression.
By default, today's data is used, but you can use any of the following modifiers:
today's yesterday's n days ago (e.g. 3 days ago) this week's last week's n weeks ago (e.g. 2 weeks ago) this month's last month's n months ago (e.g. 4 months ago)
These date modifiers refer to trading days, not calendar days, so a clause with a “5 days ago” modifier is looking at data from 5 trading days earlier (typically equivalent to one week ago).
Here are a few examples of clauses with date modifiers:
[4 days ago close > 38] [this week's high > 1 week ago high] [monthly low < last month's low]
Note that you can mix and match date modifiers within a single clause, as in the second and third examples above.
However, be cautious with mixing date modifiers that use different date periods. Using a date modifier in your clause may change the period used to calculate the indicators in that clause. For example, the following clause is actually comparing today's 14-day RSI value to last week's 14-week RSI value:
[today's RSI(14) > last week's RSI(14)]
To make sure you're comparing apples to apples, you want both modifiers to use the same date period (in this case, daily):
[today's RSI(14) > 5 days ago RSI(14)]
Standard arithmetic operators are available for use in any expression:
Here are a couple of the most commonly encountered uses of arithmetic operators:
[volume > yesterday's volume * 1.05] [close > [[high + low] / 2]]
The first example scans for stocks where today's volume is at least 5% higher than yesterday's volume. The second example scans for stocks where the closing price is in the upper half of the day's range.
The arithmetic operators are processed from left to right, but items enclosed between square brackets will be evaluated as a single unit. For example:
[close > low + [0.90*[high - low]]]
The brackets around “high - low” ensure that this subtraction will be done before the resulting value is multiplied by 0.90. Similarly, the brackets around “0.90*[high-low]” ensure that all of that math will be completed before that result is added to the low.
Remember, we only use parentheses to designate the parameters for indicators. Square brackets are used to show the Scan Engine in what order to process arithmetic operators.
There are several standard arithmetic functions that can be applied to your scan expressions:
For more details on using these functions, see the Technical Functions section of our Scan Syntax Reference.
It's pretty rare for a scan to only have one clause. As a general rule, you are testing for multiple technical conditions and signals. The logical operators “AND” and “OR” are used to string multiple clauses together into a scan.
If you use the “AND” operator to add multiple clauses to your scan, that means that all the clauses need to be true: stocks must pass the first test and the second test. Generally, you want your scan to pass all the tests (otherwise, why include a test in your scan?), so “AND” is the default operator used to connect clauses.
As an example, let's look at the following scan:
[type = stock] AND [country = US] AND [close >10]
Let's go through this scan clause by clause. First, each ticker symbol needs to pass the first test (Is it a stock?). Only securities that are stocks will pass and go on to the next test. Since the second clause is connected with an “AND” operator, we know that the stock also must pass the second test (Is US the country?). Only US stocks will pass both the first and second tests and move on to the third test. Again, since the clause is connected with an “AND” operator, we know that the stock must also pass the third test (Is today's closing price above $10?). Only US stocks with a close above $10 will pass all three tests and be included in the scan results.
While you will generally want ticker symbols to pass all the tests in your scan, there are some scenarios where multiple similar clauses could meet a requirement you're looking for. If any one of those clauses is true, then you want to return that ticker symbol in your scan results. For example, let's say you want to return all stocks that either trade on the Nasdaq OR on the NYSE. It doesn't have to trade on both exchanges; as long as it trades on one of them, you want the stock to be part of your search results.
In that case, we use the “OR” operator to define the list of clauses that could fulfill this requirement. In our example above, it might look like this:
[type = stock] AND [[exchange = Nasdaq] OR [exchange = NYSE]]
In this case, we're looking for stocks that trade on either the Nasdaq or the NYSE exchange.
Notice how, in addition to the square brackets around each clause, we also have an additional set of brackets that encompasses the entire list of “OR” clauses we are testing for. These brackets are very important, as they make it clear to the scan engine which clauses are included in your “OR” list and which ones aren't. Without them, you may get unexpected results. For more on why this is so important, check out the OR Clause section of our Troubleshooting Scans article in the Support Center.
With these extra brackets, we're basically telling the scan engine to run the first test (Is it a stock?) and, if it passes that test, to move on to the second test (“Does it trade on the Nasdaq or NYSE?). Those extra brackets make it clear that the second and third clauses are all part of a single multi-part test.
Note: We recommend putting each “test” on a separate line in your scan. It makes no difference to the Scan Engine, but it definitely makes it easier for humans to read.
For more examples of the many ways you can use OR clauses in your scans, check out our Writing Scans with OR Clauses article in the Support Center.
OR clauses can provide a lot of flexibility and power to your scans, as long as they are used with caution. The important takeaway is to always put an additional set of square brackets enclosing all of the OR clauses that should be part of that test.
You can optionally add a line to the very end of your scan that controls the order in which your scan results are displayed. That line must start with “RANK BY,” followed by the expression that you want your results sorted by. For example:
RANK BY [close]
This example would sort the scan results in descending order of current closing price.
There are a few rules for Rank By clauses:
Note: The value of the Rank By expression is also added to the right side of the scan results table. This allows you to examine the value of any expression in your scan directly by making that expression the Rank By expression as well.
It is possible to have the Scan Engine ignore portions of the scan text by adding either the '#' character or two forward slashes ('//') in your scan text. The Scan Engine will ignore all the text from that point to the end of the current line.
Comments are very useful in two ways. The first use is to document a particularly complex scan which contains a lot of clauses. With comments, you can add useful notes for yourself without actually affecting the Scan Engine at all. For example:
// My ETF scan. // July 20, 2014 [type = stock] # only get ETFs > 100 AND [close > 100] // Let's get the expensive ones! AND [group is ETF]
When the Scan Engine processes this, it only sees:
[type = stock] AND [close > 100] AND [group is ETF]
As you can see, the Scan Engine ignores all the comment text.
The other use for comments is to selectively add or remove clauses and view the results. Here is an example of toggling between two clauses within a scan:
[type = stock] AND [open > 25] // AND [group is ENERGYSECTOR] AND [group is FINANCIALSECTOR]
In this example, we can quickly switch between scanning the Financial or Energy sectors by commenting out one of the 'group is' clauses. Being able to quickly add or remove clauses using comments is very helpful for debugging scans. For more information on this usage, please see our Troubleshooting Scans article in the Support Center.
Scanning is one of the most difficult parts of our site to master, but, with a little trial and error, you'll soon be writing complex scans. When working from the Advanced Scan Workbench, remember to click the “Check Syntax” button as you make changes in order to ensure you uncover any problems with your scan syntax early. If you're not getting errors, but also not getting the results you expect, be sure to check out our Troubleshooting Scans article in the Support Center for more troubleshooting tips.